2018-04-01 04:48:31 <danielkza[m]> Hello folks! I've recently needed Squid (the caching proxy) 4.0 to use with Alpine, while the repos are currently at version 3.5 which is the latest stable (Edge included). Should I submit a PR to update it even if 4.0 is still a release candidate, or should I wait until a stable release is cut?
2018-04-01 13:54:16 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  vcgovt: pablog[m] dnull rajivr ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:17 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  alitc: drebrez[m] eleksir jomat ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:17 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ilsqctflki: stwa _ikke_ ephemer0l_ ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:17 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  wwdwpbcrel: RyabyyDenis[m] Ganwell clandmeter1 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:21 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  shqmxfvnbh: lucy_bunny JStoker eleksir ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:21 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  yusesnbcep: RyabyyDenis[m] mmlb xj9[m]1 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:22 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  dpfogtrqk: mksully22 opendata andypost[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:54:22 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  zoorc: clandmeter letoram ata2001[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:27 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  pqnrg: mcrute hl MatrixTraveler[m ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:27 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  uvhdvokafh: Bureaucat hrmlgon-z fekepp ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:27 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  yxbalpofx: lostd xj9[m]1 arch3y ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:54:27 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  xdjoifb: letoram mcrute mepholic ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:32 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  pbasm: Bloo[m] ata2001[m] NIN101 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:32 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  jnhwrobxa: Pomidora rdutra pardis ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:32 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ffxyjj: DLange ollieparanoid[m] z3ntu ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:32 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  rzyfmzdl: ata2001[m] Pomidora rdutra ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:54:36 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  qlelovgvhh: ephemer0l_ mmlb lucy_bunny ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:37 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  xxualrf: RyabyyDenis[m] MH0815 wictory[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:54:37 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  crnrmhd: hl IAMB3NW hrmlgon-z ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:37 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  xvrny: Ganwell YoursTrul RyabyyDenis[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:41 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ncgkcqytz: ncopa jirutka fcolista ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:41 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  viajqurgux: tru_tru rajivr dnull ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:42 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  wrjclpwph: fcolista clemens3 hrmlgon-z ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:42 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  bycxxgau: MH0815 mepholic eleksir ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:46 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  fmvwb: BitL0G1c tmh1999[m] JStoker ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:46 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  gcixiqn: doppo fcolista hl ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:47 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ftizxsqi: NIN101 Ganwell tty` ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:47 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  emkdkmjs: drebrez[m] bernhardgruen[m] mepholic ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:52 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  srdmc: Michitux blahdodo wictory[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:52 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  qfneutkns: Bloo[m] andypost czart ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:52 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  fyhxtso: fcon Bloo[m] eleksir ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:52 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  aqkjh: jirutka hrmlgon-z coredumb ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:57 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  hweqzb: clemens3 mitchty z3ntu ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:57 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  uzsixxvdx: hl Bloo[m] dnull ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:57 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  oznru: rajivr mmlb terra ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:54:58 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  nbvrbtl: Ganwell Michitux andypost ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:02 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  oxnyff: Aerdan[m] z3ntu lostd ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:02 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  fypdhkv: MartijnBraam[m] hrmlgon-z fcolista ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:02 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ceokx: danielkza[m] mixi ncopa ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:03 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  whjuppowdl: czart jpo runelabs[m] ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:06 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  rlzpnxq: ncopa2 runelabs[m] dobson ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:07 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  sootrh: fcolista MartijnBraam[m] blahdodo ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:07 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  fmucohlyyo: opendata tmh1999[m] tdtrask[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:08 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  vucdft: tdtrask[m] bernhardgruen[m] RyabyyDenis[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:55:11 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  dutlxg: fcolista_web drebrez[m] clandmeter ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:11 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  lsbmtbab: fekepp Bloo[m] drebrez[m] ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:12 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  crbhr: fekepp ncopa2 ata2001[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:13 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  qremwwt: blahdodo tmh1999[m] bfritz_ ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:16 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ofmpwdobh: tru_tru czart Michitux ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:16 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  yxvvrkm: tboerger[m] bfritz_ tmh1999[m] ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:16 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ndavfcod: Ganwell bernhardgruen[m] RyabyyDenis[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:18 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  vzxitzi: dobson mmlb ephemer0l_ ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:21 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ovjngzzs: drebrez[m] wictory[m] clandmeter1 ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:21 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  uohwhk: runelabs[m] lucy_bunny xj9[m]1 ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:22 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  eiloj: andypost rajivr fcolista_web ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:23 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  cblgsyw: stwa fekepp Ganwell ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:26 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  qtsasbb: mmlb _ikke_ arch3y ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:26 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ygujxzqor: Ganwell fcon jnettlet[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:27 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  evwvtyf: wictory[m] pablog[m] BitL0G1c ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:28 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  fjffls: Bloo[m] bernhardgruen[m] fcolista ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:31 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  uedave: ollieparanoid[m] tmh1999[m] clandmeter ▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:55:31 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  kfhdld: kaniini Aerdan[m] hl ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:32 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  grjagwln: NIN101 xj9[m]1 PureTryOut[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:33 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  rplhn: MartijnBraam[m] Pomidora jnettlet[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:55:36 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  sxmrx: blahdodo mitchty JStoker ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:36 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  assnxxo: jnettlet[m] clemens3 mitchty ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:55:37 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  pacattwufe: notchisa hrmlgon-z Aerdan[m] ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:38 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  euzduxm: jomat ephemer0l_ MH0815 ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:41 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  lhfhvqee: MatrixTraveler[m rdutra czart ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:42 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  dmuvb: tmh1999[m] Pomidora mcrute ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:42 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  axqvsk: mepholic jnettlet[m] mitchty ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:43 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  dxyjn: pardis tmh1999[m] MatrixTraveler[m ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:55:46 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  pqxrlc: mcrute fekepp BitL0G1c ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:47 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  tqftjgeq: Aerdan[m] MatrixTraveler[m tdtrask[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:47 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ptyppxhqyx: detha dobson clemens3 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:48 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  oecoglsk: fcolista_web wictory[m] coredumb ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:52 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  jqjtzg: kaniini MartijnBraam[m] YoursTrul ▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:55:52 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  asofdowgmw: detha bernhardgruen[m] lostd ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:52 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  kvqaf: fcon xentec mksully22 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:53 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  bjlkj: pablog[m] tdtrask[m] ollieparanoid[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:56 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  fysbecih: drebrez[m] tmh1999[m] jpo ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:56 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  pgxpih: DLange jnettlet[m] dsabogal ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:55:57 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  seokvro: MatrixTraveler[m runelabs[m] dnull ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:55:58 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  syxjfbs: jirutka rdutra MartijnBraam[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄�
2018-04-01 13:56:02 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  epftokx: doppo pablog[m] lucy_bunny ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:02 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  putiwzi: mepholic mitchty pardis ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:02 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  smeoiqvgal: DLange dobson tboerger[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 13:56:03 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  nvywukml: mepholic drebrez[m] clemens3 ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:08 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  bjort: fcon terra YoursTrul ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:08 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  fytizjt: andypost[m] dnull mitchty ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:08 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  zbmmr: hl ncopa2 detha ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:10 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  himdx: ed-packet terra ncopa ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:13 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ahdzkbc: xentec fekepp clandmeter1 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:13 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  gwgykzby: ephemer0l_ xentec andypost ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:13 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  zpapgtbuec: Bloo[m] jnettlet[m] mksully22 ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:14 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  olizfd: carlosdavidepto[ jpo detha ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:17 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  tqihzwcfw: tmh1999[m] mcrute coredumb ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:17 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  pwzdm: RyabyyDenis[m] ollieparanoid[m] fekepp ▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:17 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  iapbhvt: bernhardgruen[m] Bureaucat dsabogal ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:19 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  drequfcls: carlosdavidepto[ mepholic andypost[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄�
2018-04-01 13:56:23 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  tbwvicpzc: wictory[m] rdutra jnettlet[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:23 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  vclozcnwfa: andypost[m] fcolista_web clandmeter1 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:23 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  dubqkmpme: tmh1999[m] Pomidora jirutka ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄�
2018-04-01 13:56:25 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  pygaaczgrd: ollieparanoid[m] runelabs[m] dsabogal ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄�
2018-04-01 13:56:26 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  kbwvdor: arch3y MatrixTraveler[m wictory[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:27 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  sctvgwtznz: dobson YoursTrul MatrixTraveler[m ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:27 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ltmwrzw: MartijnBraam[m] fcolista RyabyyDenis[m] ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:29 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  utpigh: JStoker dobson notchisa ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:32 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  otupwsw: Pomidora MatrixTraveler[m doppo ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:32 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  bzzmnagyaz: clandmeter1 MatrixTraveler[m dsabogal ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:32 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  cjqupevcb: blahdodo mepholic jpo ▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:33 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  gfggng: jirutka MatrixTraveler[m jomat ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:36 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  mpbvgj: fcolista[m] ollieparanoid[m] doppo ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:36 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  lfavcpv: mepholic andypost opendata ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:37 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  pypbblgi: notchisa letoram ephemer0l_ ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:38 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  bzybwhvnkv: _ikke_ mepholic eleksir ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:41 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  nxdaoh: kaniini clemens3 MH0815 ▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:42 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  luynu: clandmeter ncopa2 Ganwell ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:42 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  ewzhxozjp: detha mepholic ncopa2 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:43 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  isdbn: mitchty z3ntu BitL0G1c ▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:46 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  fbqcwdx: xj9[m]1 opendata Bureaucat ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:46 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  mebretlzd: tdtrask[m] clandmeter mksully22 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄

2018-04-01 13:56:47 <ifbizo875> ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ HAPPY APRIL FLOODS DAY BROUGHT TO YOU BY iяс.sцреяиетs.ояg сни sцреявоwl  jneoxokrga: mcrute tmh1999[m] jomat ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
2018-04-01 14:15:05 <tboerger[m]> really funny... not.
2018-04-01 14:24:03 <rdutra> :-/
2018-04-01 14:37:34 <TBB> same old asshole spammer again
2018-04-01 18:19:24 <Truxx> there is an issue here:  https://dev.alpinelinux.org/irclogs/%23alpine-devel-2016-01.log
2018-04-01 18:19:42 <_ikke_> Truxx: ?
2018-04-01 18:19:49 <Truxx> tried modprobe ppp and says that ppp is not in modules.dep
2018-04-01 18:20:49 <Truxx> Trying to set up ppp, but it just won't work with ppp and usb-modeswitch - is there anything else needed?
2018-04-01 18:20:53 <Truxx> I mean packages?
2018-04-01 18:21:29 <Truxx> ppp should be in /etc/modules  - right?
2018-04-01 18:21:55 <mps> modprobe ppp_async
2018-04-01 18:22:40 <clandmeter> ppp_generic
2018-04-01 18:23:29 <Truxx> Thank you for the input. So is it ppp_generic needed in /etc/modules?
2018-04-01 18:23:35 <mps> in my /etc/modules is only ppp_async and it works
2018-04-01 18:23:58 <clandmeter> it probably pulls in generic
2018-04-01 18:24:04 <Truxx> Ok, thank you @mps, I'll try that. Did you only use ppp and usb-modeswitch?
2018-04-01 18:24:13 <mps> right
2018-04-01 18:24:36 <Truxx> Ok, that good news, thx. At least it's not a question of missing packages...
2018-04-01 18:24:46 <mps> both
2018-04-01 18:25:03 <Truxx> Yes, I installed both.
2018-04-01 18:27:35 <Truxx> Just a last question: Do you also have an usbserial line in modules in order to switch the usb from storage mode?
2018-04-01 18:28:16 <mps> of course
2018-04-01 18:29:05 <Truxx> Ok, thank you. At least I got that part right and the id changes as it should...
2018-04-01 18:29:11 <mps> actually no, it is puled by ppp_async probably
2018-04-01 18:29:37 <mps> or hot-pluged when device is inserted
2018-04-01 18:29:50 <Truxx> Godgood... I messed up that already then...
2018-04-01 18:30:51 <Truxx> Anyway, good to know, thx. I'm gonna test it. CU
2018-04-01 18:50:59 <danielkza[m]> Since I might have been missed due to the spam: Hello folks! I've recently needed Squid (the caching proxy) 4.0 to use with Alpine, while the repos are currently at version 3.5 which is the latest stable (Edge included). Should I submit a PR to update it even if 4.0 is still a release candidate, or should I wait until a stable release is cut?
2018-04-01 18:53:38 <sroracle1> I would just submit a PR and they will merge it when they see it as appropriate
2018-04-01 18:54:03 <sroracle1> Tag as WIP until the stable release is out
2018-04-01 18:54:23 <danielkza[m]> [[sroracle]]: I will do that, thank you!
2018-04-01 20:03:38 <Truxx> It's me again... Just a remark: It looks like here: https://wiki.alpinelinux.org/wiki/Using_HSDPA_modem  and here:  https://wiki.alpinelinux.org/wiki/Using_serial_modem  ppp is a bit confusing
2018-04-01 20:04:05 <Truxx> Meanwhile I tried in /etc/modules ppp_generic and ppp_async and booth seem to work fine
2018-04-01 20:05:43 <Truxx> Also: There is no need for an usbserial-line in /etc/modules but: in /etc/network/interfaces I've got pre-up usb-modeswitch -c /etc/usb-modeswitch.conf
2018-04-01 20:07:15 <Truxx> In that .conf file there is data for the Vendor Id etc. - so it's recognized as a gsm modem. Still: There is no ppp device for me.
2018-04-01 20:07:30 <Truxx> Maybe someone could point in the right direction?
2018-04-01 20:11:27 <Truxx> I have this in dmesg: usbserial: USB Serial support registered for GSM modem (1-port)
2018-04-01 20:12:07 <Truxx> And after some lines in between: usb 1-1.1: GSM modem (1-port) converter now attached to ttyUSB4 - I suppose that look sok
2018-04-01 20:19:12 <Truxx> I have in /etc/ppp/peers and /etc/chatscripts the supposedly correct scripts for the modem (found in a tuto) - it still does not work. Any ideas why not?
2018-04-01 20:52:33 <Truxx> I forgot to mention: It does not even ask for the pin to unlock the device while on other distros it works fine after unlocking
2018-04-01 21:09:12 <skarnet> andypost: please don't do non-maintainer-upgrades
2018-04-01 21:09:44 <skarnet> https://github.com/alpinelinux/aports/pull/3862/commits/9271ff13faa58d71d58a3f420e7b8bc5e6a03073 is a very bad idea, have you even looked at what you're attempting to mergE?
2018-04-01 21:16:46 <andypost> skarnet, sure, I did tested it
2018-04-01 21:17:17 <andypost> skarnet, and I filed PR to get maintainer's review
2018-04-01 21:18:06 <andypost> skarnet, execline is a dependency for other libs to build them
2018-04-01 21:19:53 <skarnet> well, this is the maintainer's review: don't merge that, and cancel your PR immediately
2018-04-01 21:20:26 <andypost> skarnet, np, can you elaborate why?
2018-04-01 21:21:26 <skarnet> and in the future, please pay attention to what you're trying to merge. Even though I do appreciate the intent of upgrading to the latest version as soon as possible, this is in no way a required dependency to anything else (everything depending on execline will work with 2.3.0.4 perfectly)
2018-04-01 21:22:13 <skarnet> yes, I can elaborate why, but the fact that it's not entirely obvious to you why is proof that you did not even LOOK at the software
2018-04-01 21:22:35 <skarnet> you tested it? nothing shocked you when you tested it?
2018-04-01 21:22:50 <andypost> I'm just using s6
2018-04-01 21:23:04 <skarnet> there's no regression, indeed
2018-04-01 21:23:14 <skarnet> but that's not the problem
2018-04-01 21:24:26 <andypost> skarnet, s6 require newer skalibs https://skarnet.org/software/s6/
2018-04-01 21:24:58 <skarnet> and?
2018-04-01 21:28:23 <andypost> hm, s6-linux-utils probably
2018-04-01 21:28:30 <skarnet> no
2018-04-01 21:28:46 <andypost> looks I missread something
2018-04-01 21:29:01 <skarnet> my point is that yes, the new s6 requires the new skalibs, and the required abumps have already been sent
2018-04-01 21:29:05 <skarnet> by me, on the ML
2018-04-01 21:29:12 <skarnet> and no, it does not require the new execline
2018-04-01 21:29:31 <andypost> skarnet, so I can remove this commit
2018-04-01 21:30:18 <skarnet> now, it's half an hour too early for this, but if you really need it spelled out, here's what you attempted to merge: https://skarnet.org/cgi-bin/archive.cgi?1:mss:1156:nddjgnadpajgmgomcian
2018-04-01 21:30:41 <skarnet> and a very cursory glance would have shown you this: https://github.com/skarnet/execline/commit/9b0f07047aa0b9cde7e09e57528e0d314c1a6435
2018-04-01 21:32:40 <andypost> skarnet, is that really a 1Apr joke?
2018-04-01 21:32:47 <skarnet> Did I stutter?
2018-04-01 21:33:09 <andypost> 
ACTION upset(

2018-04-01 21:33:22 <skarnet> not as much as I am
2018-04-01 21:34:03 <skarnet> try to make a joke, people just don't look, don't even realize there's a joke, and merge the thing
2018-04-01 21:34:09 <skarnet> this is why we can't have nice things.
2018-04-01 21:38:49 <andypost> skarnet, it's hard to work at 1Apr(
2018-04-01 21:39:28 <skarnet> you're telling me
2018-04-01 21:40:28 <skarnet> if you're going to upgrade software to a new version, and you're not the author, reading the release notes is the ABSOLUTE MINIMUM you should be doing
2018-04-01 21:41:20 <skarnet> so, I'm sorry, but it's 100% on you
2018-04-01 21:41:48 <andypost> agree, but I read them only for s6 - is what I actually use
2018-04-01 21:42:57 <skarnet> you did not. The upgrade notes for s6 do not say you need execline-2.4.0.0.
2018-04-01 21:44:47 <andypost> skarnet, http://skarnet.org/software/s6/upgrade.html
2018-04-01 21:45:47 <andypost> skarnet, execline I did add just as other deps
2018-04-01 21:46:47 <skarnet> jesus, you don't know when to stop, do you? Where in that page is "execline dependency bumped to 2.4.0.0" written?
2018-04-01 21:46:54 <skarnet> so let me sum things up for you
2018-04-01 21:46:58 <skarnet> YOU FUCKED UP
2018-04-01 21:47:02 <skarnet> NOW SHUT UP AND MOVE ON
2018-04-01 21:47:07 <skarnet> and stop pissing me off.
2018-04-01 21:47:21 <andypost> skarnet, s6 require newer skalibs that's it
2018-04-01 21:47:26 <skarnet> yes.
2018-04-01 21:47:52 <skarnet> the bump for skalibs has already been sent. It's safe.
2018-04-01 21:48:08 <andypost> and also it require execline which I just bumped to latest version
2018-04-01 21:48:32 <skarnet> I give up.
2018-04-01 21:48:56 <andypost> skarnet, and knowing that you will review this anyway I just bumped othe s6 relates
2018-04-01 21:49:13 <andypost> what's wrong?
2018-04-01 21:49:27 <skarnet> You.
2018-04-01 21:50:25 <andypost> I can't get you...
2018-04-01 21:51:15 <andypost> PR updated and waiting for review
2018-04-02 10:05:10 <Aerdan[m]> 
ACTION reads log, facepalms.

2018-04-02 16:13:31 <deedubs> my .patch files hate me
2018-04-02 16:13:39 <deedubs> I can't figure out the magic sauce to get them to apply
2018-04-02 16:14:31 <deedubs> + is_remote disable-race-tests-on-alpine.patch
2018-04-02 16:14:31 <deedubs> + return 1
2018-04-02 17:33:00 <deedubs> also is there a good  Golang example of a package?
2018-04-02 17:40:10 <deedubs> I had used the Docker pkg but apparently that was a mistake :D
2018-04-02 19:04:22 <ncopa> libressl 2.7.2 is stable!
2018-04-02 19:17:31 <TBB> nice, I've been waiting for 2.7 series to become stable
2018-04-02 19:28:33 <ncopa> this is going to be fun...
2018-04-02 19:29:02 <ncopa> openssldh_link.c:78:1: error: static declaration of 'DH_get0_key' follows non-static declaration
2018-04-02 19:29:03 <ncopa>  DH_get0_key(const DH *dh, const BIGNUM **pub_key, const BIGNUM **priv_key) {
2018-04-02 19:29:04 <ncopa>  ^~~~~~~~~~~
2018-04-02 19:29:21 <ncopa> 
ACTION sent a long message: ncopa_2018-04-02_19:29:20.txt <https://matrix.org/_matrix/media/v1/download/matrix.org/DiacptTctWHMWwltqrxbBNHI>

2018-04-02 19:32:18 <ncopa> now we have to remove the || defined(LIBRESSL_VERSION_NUMBER)
2018-04-02 19:33:24 <ncopa> i wonder how many of those we can remove now...
2018-04-02 19:33:25 <ncopa> oh well
2018-04-02 19:35:10 <ncopa> the correct fix is to do configure check for DH_get0_key
2018-04-02 20:30:04 <azarus> yeah, the new libressl implements quite a bit of then openssl 1.1 API, wonder how that'll play out
2018-04-02 21:00:02 <ncopa> it will make messy code even more messy
2018-04-02 21:03:01 <azarus> oh joy ;(
2018-04-03 07:07:46 <azarus> I wonder if the file '/lib/firmware/iwlwifi-7260-7.ucode' (in package linux-firmware-other) would be better categorized in linux-firmware-intel
2018-04-03 07:08:07 <azarus> Would make sense to me to put iwlwifi firmware in the linux-firmware-intel package
2018-04-03 07:08:58 <azarus> (oh whoops, I don't just mean /lib/firmware/iwlwifi-7260-7.ucode, I mean /lib/firmware/iwlwifi-7260-*.ucode)
2018-04-03 09:54:59 <wener_> why github pr for aports get so long to review ?
2018-04-03 09:57:10 <wener_> dose every pr needs to reviewed by ikke ?
2018-04-03 10:03:12 <ncopa> wener_: we do as good as we can. I am currently working on libressl 2.7 upgrade
2018-04-03 10:04:29 <ncopa> wener_: i have also noticed that PRs which needs lots of changes tends to get pushed back in the queue
2018-04-03 10:06:45 <wener_> I pushed my first pr tow weeks ago, has bean reviewed, then I fixed the pr as required, as a contributor I do hope alpine getting better, but there is nothing more I can do.
2018-04-03 10:07:02 <ncopa> which is it?
2018-04-03 10:07:27 <wener_> https://github.com/alpinelinux/aports/pulls/wenerme
2018-04-03 10:08:02 <ncopa> i think we need a bot that can indicate that we got feedback and that it is our turn to respond
2018-04-03 10:08:39 <wener_> dose nmeum a bot ?
2018-04-03 10:10:16 <ncopa> the problem is that we struggle with keeping up
2018-04-03 10:10:46 <ncopa> i asked you to squash the commits so it becomes a single commit instead of 2
2018-04-03 10:10:52 <nmeum> wener_: ?
2018-04-03 10:11:25 <wener_> Sorry, I thought you are the bot to add label for every pr ~
2018-04-03 10:12:12 <nmeum> no, i am not a bot. I mostly add those manually but usually don't get a around to reviewing every pull request i added a label to
2018-04-03 10:12:34 <ncopa> i never met nmeum in real life so I cannot confirm that he is not a bot :)
2018-04-03 10:12:45 <nmeum> :)
2018-04-03 10:13:13 <ncopa> and if he is a bot, he is a pretty intelligent and nice bot :)
2018-04-03 10:13:39 <ncopa> wener_: you can squash the commits with `git rebase -i` and then git push --force
2018-04-03 10:14:39 <wener_> thx, not very familiar with the pr process. I should reread the CONTRIBUTING.md
2018-04-03 10:15:21 <wener_> so, maybe nmeum is a bot in real life ?
2018-04-03 10:46:56 <ncopa> wener_: i added some more comments to grpc and tested to build it with libressl
2018-04-03 10:47:12 <ncopa> it fails to build with libressl 2.6 but i think it will build with libressl 2.7
2018-04-03 10:47:31 <ncopa> src/core/lib/security/credentials/jwt/jwt_verifier.cc:526:47: error: 'RSA_set0_key' was not declared in this scope
2018-04-03 10:47:32 <ncopa>    if (!RSA_set0_key(rsa, tmp_n, tmp_e, nullptr)) {
2018-04-03 10:49:07 <ncopa> which means that i should merge grpc with my libressl upgrade
2018-04-03 10:53:46 <wener> when 2.7 will be ready ?
2018-04-03 10:59:19 <ncopa> wener_: i have actually merged grpc, and fixed it in my libressl rebuild queue
2018-04-03 10:59:19 <ncopa> so it will be pushed together with the libressl upgrade
2018-04-03 11:05:08 <wener_> ncopa: if you pushed grpc, please also merge grpc-java
2018-04-03 11:08:54 <wener_> if you got time, please also check my other prs, thx
2018-04-03 11:28:15 <bernhardgruen> Hello, there are currently some packages that depend on sourceforge.net Downloads. It seems many of those downloads have changed. This will create problems while rebuilding for 3.8. How to handle that?
2018-04-03 11:30:50 <jirutka> ncopa: please do NOT merge grpc abuild without very thorough check what does this google sh*t contains; their ruby binding for grpc is still the worst Ruby gem I’ve ever seen
2018-04-03 11:31:49 <jirutka> bundled CA certs, bundled binaries with their own openssl fork and few other libs, broken build etc.
2018-04-03 11:33:14 <jirutka> I have abuild for ruby-grpc (unfortunately I need it for GitLab), but I intentionally didn’t push it to Alpine aports
2018-04-03 11:35:36 <jirutka> ncopa: kaniini: please respond to https://github.com/alpinelinux/aports/pull/3848
2018-04-03 11:56:24 <ncopa> wener_: there you ahve the response why it takes time to get those PRs merged ^^^
2018-04-03 12:25:21 <jirutka> clandmeter: few days ago there were interesting question re PXE boot, could you please look at it? https://dpaste.de/LmgV/raw
2018-04-03 12:28:12 <clandmeter> jirutka, if i read correctly, he wants ipv6 in our initramfs?
2018-04-03 12:28:20 <jirutka> clandmeter: yes
2018-04-03 12:28:33 <jirutka> clandmeter: he has IPv6-only network with SLAAC
2018-04-03 12:29:43 <clandmeter> i dont have any experience with slaac, maybe somebody could help patch it in.
2018-04-03 12:29:50 <jirutka> clandmeter: if he will not figure out how to do that with Alpine, he will have to use some CentOS, so it’s a matter of honor to make it work! :)
2018-04-03 12:30:21 <clandmeter> honor doesnt pay my bills ;-)
2018-04-03 12:30:52 <jirutka> clandmeter: I have some experience with IPv6 SLAAC, but zero experience with PXE boot
2018-04-03 12:31:11 <ncopa> sounds like you need each other :)
2018-04-03 12:31:17 <jirutka> yeah
2018-04-03 12:31:18 <clandmeter> i dont think the pxe part is the problem.
2018-04-03 12:31:42 <ncopa> i need both of you ;)
2018-04-03 12:31:44 <clandmeter> what happens is ipxe will setup networking
2018-04-03 12:31:45 <jirutka> XD
2018-04-03 12:32:05 <jirutka> but how iPXE set ups networking?
2018-04-03 12:32:13 <clandmeter> and initramfs will take over and do dhcp or static assignement.
2018-04-03 12:32:35 <clandmeter> this part doesnt support it probably.
2018-04-03 12:33:01 <clandmeter> Benw2:	I will be chainloading with iPXE which supports ipv6 slaac
2018-04-03 12:33:16 <clandmeter> so we dont care about ipxe
2018-04-03 12:34:10 <clandmeter> i guess you need to provide something to generate the slaac ipv6 address?
2018-04-03 12:34:11 <jirutka> IIRC SLAAC is fullay managed by Linux itself, just need to accept RA (net.ipv6.conf.<iface>.accept_ra)
2018-04-03 12:35:00 <clandmeter> my ipv6 is limited, but we are going to implement it here soon so its good for me to know more about the topic.
2018-04-03 12:35:49 <clandmeter> yes, thats what a quick read tells me, there is no negotiation for address, its calculated.
2018-04-03 12:35:50 <jirutka> oh and also net.ipv6.conf.<iface>.autoconf must be enabled
2018-04-03 12:40:18 <korylprince> Hello, all. Just wondering what sort of schedule the builders run on? (i.e. what's the time frame for accepted patches to be built?)
2018-04-03 12:40:36 <clandmeter> jirutka, we probably have to check how others implement ipv6 in initramfs.
2018-04-03 12:40:43 <jirutka> clandmeter: yes
2018-04-03 12:40:53 <clandmeter> https://www.kernel.org/doc/Documentation/filesystems/nfs/nfsroot.txt doesnt mention anything.
2018-04-03 12:42:20 <jirutka> clandmeter: I’ve contacted pavlix to help us
2018-04-03 12:42:44 <pavlix> Hi. I can't promise much regarding iPXE but I'm interested.
2018-04-03 12:42:58 <clandmeter> hi pavlix
2018-04-03 12:43:46 <jirutka> pavlix: relevant links: our iPXE support: https://github.com/clandmeter/alpine-netboot, request from user: https://dpaste.de/LmgV/raw
2018-04-03 12:44:23 <clandmeter> repo is located at https://github.com/alpinelinux/alpine-netboot
2018-04-03 12:44:38 <jirutka> pavlix: and our tool for creating initramfs https://github.com/alpinelinux/mkinitfs
2018-04-03 12:45:15 <clandmeter> but the alpine-netboot is just scripts to generate images.
2018-04-03 12:45:19 <jirutka> clandmeter: aha, sry; could you please remove your repo and fork it from alpinelinux, so there will be explicit connection between them?
2018-04-03 12:45:23 <clandmeter> we also have alpine-ipxe
2018-04-03 12:45:31 <clandmeter> it generates the ipxe loaders
2018-04-03 12:45:45 <pavlix> Are you in active contact with Benw2?
2018-04-03 12:46:22 <clandmeter> we need to add support for ipv6 in initramfs
2018-04-03 12:46:31 <clandmeter> which is located here: https://github.com/alpinelinux/mkinitfs/blob/master/initramfs-init.in#L146
2018-04-03 12:47:18 <pavlix> clandmeter: Are you using a custom scripted solution for that?
2018-04-03 12:47:37 <jirutka> pavlix: unfortunately I forgot to ask Benw2 for contact and he’s not on IRC now, so not
2018-04-03 12:47:43 <pavlix> clandmeter: Is dhcpcd an option for you?
2018-04-03 12:48:03 <jirutka> pavlix: Benw2 asked for SLAAC, i.e. just RA
2018-04-03 12:48:03 <clandmeter> slaac doesnt use dhcp right?
2018-04-03 12:48:38 <jirutka> pavlix: am I right that there’s no daemon needed for SLAAC, it’s fully managed by Linux itself?
2018-04-03 12:48:43 <jirutka> pavlix: on client side
2018-04-03 12:49:02 <fabled> correct. SLAAC is handled automatically by kernel and does not require any servers on network.
2018-04-03 12:49:23 <clandmeter> i think the router needs to advertise the subnet and the address is generated on the host. but i could be wrong :)
2018-04-03 12:50:04 <fabled> you always get SLAAC address with link-local address. if router advertises prefixes, you get one additional address per each advertised address.
2018-04-03 12:50:38 <clandmeter> SLAAC is a method in which the host or router interface is assigned a 64-bit prefix, and then the last 64 bits of its address are derived by the host or router with help of EUI-64 process which is described in next few lines. SLAAC uses NDP protocol to work.
2018-04-03 12:50:42 <fabled> on normal working setups you get two SLAAC addresses: one link-local, and one in the router advertised subnet
2018-04-03 12:51:44 <pavlix> jirutka: dhcpcd does exactly that.
2018-04-03 12:52:19 <pavlix> jirutka: In userspace, so it can help you get over all sorts of issues intrinsic to the kernel autoconf.
2018-04-03 12:52:38 <pavlix> clandmeter: Correct.
2018-04-03 12:53:27 <clandmeter> busybox cannot provide this?
2018-04-03 12:53:28 <pavlix> fabled: You can use the kernel autoconf or you can do all that in the user space which generally plays nicer.
2018-04-03 12:53:46 <pavlix> clandmeter: Busybox isn't very IPv6 friendly AFAIK.
2018-04-03 12:56:20 <fabled> pavlix, yes, it's doable in userspace too. but kernel does that automatically unless configured otherwise.
2018-04-03 12:56:40 <pavlix> fabled: Kernel doesn't configure DNS though.
2018-04-03 12:57:49 <pavlix> fabled: And it also doesn't cover the DHCP cases that weren't requested in this case but might well be required due to other environments where router discovery is not considered sufficient, like Windows or Cisco.
2018-04-03 12:58:22 <pavlix> fabled: Therefore you can't claim full autoconf support with just SLAAC and neither RA based RDNSS setting nor DHCPv6.
2018-04-03 12:58:45 <fabled> yes, i was talking specifically about SLAAC. not autoconf.
2018-04-03 12:59:02 <pavlix> fabled: Yeah but will they provide DNS manually?
2018-04-03 12:59:20 <pavlix> fabled: Are we going to use 1.1.1.1 or something like that?
2018-04-03 12:59:57 <azarus> the mt76 driver is enabled on linux-vanilla, right?
2018-04-03 13:00:36 <jirutka> pavlix: is DNS actually needed for iPXE boot?
2018-04-03 13:01:18 <pavlix> jirutka: Good point. As long as you just use TFTP by IP address or something along that lines, you should be good.
2018-04-03 13:01:58 <pavlix> jirutka: You might just need to wait for the autoconf to finish before querying the server.
2018-04-03 13:02:41 <pavlix> (Which is also usually better handled by a process that already handles waiting for IPv4 DHCP.)
2018-04-03 13:03:22 <trafex> Hi, I'm not sure where to ask this question, but the latest update on php7-mongodb (rebuild against php7.2) breaks the extension: [03-Apr-2018 12:52:57 UTC] PHP Warning:  PHP Startup: Unable to load dynamic library '/usr/lib/php7/modules/mongodb.so' - Error relocating /usr/lib/php7/modules/mongodb.so: zend_vspprintf: symbol not found in Unknown on line 0. Is there something I can do to resolve this?
2018-04-03 13:06:05 <pavlix> trafex: Are both updated to the lastest version?
2018-04-03 13:06:40 <pavlix> trafex: Better the whole system.
2018-04-03 13:06:49 <clandmeter> apk -U upgrade -a
2018-04-03 13:07:08 <trafex> pavlix: yes, it's a Docker image. So I rebuild the whole system everytime
2018-04-03 13:08:31 <trafex> I tried every combination of repositories. It only seems to work when all php7-* packages are from the same edge/testing repo. But then other stuff breaks because not all php7- packages are the same minor version (7.2)
2018-04-03 13:09:55 <trafex> It would be nice if I can use the previous version of php7-mongodb, but I guess that isn't possible? Adding the version to the package like this: 'php7-mongodb=1.4.2-r1' gives a error about a non-existing package
2018-04-03 13:10:08 <mps> pavlix: jirutka: I noticed that iso images autoconfigure IPv6 if the ipv6 module is loaded
2018-04-03 13:12:30 <pavlix> mps: As long as accept_ra (and a couple of other sysctls) are set, the addresses get configured at some point of time but it's not like the traditional configuration where you have a userspace tool that returns when the configuration is finished.
2018-04-03 13:13:18 <pavlix> And the bad thing about router discovery is that the configuration is actually never to be considered finished but that's another story about someone fixing things that weren't broken.
2018-04-03 13:13:24 <trafex> ah, seems there is a working combination; getting the php7-* packages from 'edge/main' (seems to include all 7.2 versions). And getting php7-mongodb package from 'edge/testing'
2018-04-03 13:13:34 <mps> I know, but the network is working
2018-04-03 13:14:20 <pavlix> mps: Yeah, sometimes. :)
2018-04-03 13:14:31 <clandmeter>  if you load ipv6 module the interface is up automatically?
2018-04-03 13:15:05 <mps> clandmeter: no, by 'ip link set eth0 up'
2018-04-03 13:15:44 <pavlix> clandmeter: I didn't play with systems with ipv6 module not included, tbh, anyway whether the interface is up or not isn't related to protocol version as mps suggested.
2018-04-03 13:19:02 <jirutka> pavlix: clandmeter: we just need to add `iface eth0 inet6 auto` in /etc/network/interfaces, don’t we? but I’m not sure if busybox also enables net.ipv6.conf.<iface>.autoconf and net.ipv6.conf.<iface>.accept_ra
2018-04-03 13:19:29 <jirutka> clandmeter: actually, is /etc/network/interfaces used in initramfs?
2018-04-03 13:19:52 <clandmeter> jirutka, look at the uri i pasted
2018-04-03 13:20:05 <clandmeter> it will show you the config steps initramfs does
2018-04-03 13:20:16 <jirutka> clandmeter: which URI?
2018-04-03 13:20:29 <clandmeter> https://github.com/alpinelinux/mkinitfs/blob/master/initramfs-init.in#L146
2018-04-03 13:20:36 <jirutka> aha, this one
2018-04-03 13:20:44 <pavlix> jirutka: How do you expect to solve the ordering issue? Do you expect to retry tftp until it succeeds or do you expect to wait explicitly for successful IPv6? Also do you expect to support IPv4+IPv6 together and therefore waiting for either of them to succeed?
2018-04-03 13:21:09 <mps> I think when the ipv6 module is loaded it enables autoconf and accept_ra
2018-04-03 13:21:19 <pavlix> jirutka: In my opinion the easiest way to achieve all of that is to leave kernel autoconf alone and use dhcpcd in its more or less default configuration for everything.
2018-04-03 13:21:34 <clandmeter> pavlix, ipxe does its own networking.
2018-04-03 13:21:40 <jirutka> clandmeter: aha, so it’s handled directly, that’s good; however, why do you use "ifconfig" instead of "ip"?
2018-04-03 13:21:41 <pavlix> mps: Yeah, should be the default.
2018-04-03 13:22:03 <pavlix> clandmeter: So what is the method that passes the configuration to initramfs system?
2018-04-03 13:22:37 <mps> pavlix: it is default long time ago, IIRC
2018-04-03 13:22:43 <clandmeter> for ipv4 its dhcp
2018-04-03 13:22:48 <pavlix> mps: Yeah.
2018-04-03 13:23:17 <pavlix> clandmeter: Then it means for IPv4 you have to reconfigure. I expect you have to reconfigure for IPv6 just as well.
2018-04-03 13:23:29 <clandmeter> right
2018-04-03 13:24:47 <pavlix> clandmeter: There might be some additional issues but for now the problem is still the same, you want to get it configured and you want to know when it's done... i.e. you basically want to do DHCP but using RA which was unfortunately explicitly defined differet from DHCP but still can be used that way with a bit of care.
2018-04-03 13:25:33 <pavlix> Not to say the whole idea of configuring IPv4 and IPv6 separately is not that great.
2018-04-03 13:34:06 <clandmeter> pavlix, i would like to know how other distros handle it in initramfs.
2018-04-03 13:35:07 <pavlix> clandmeter: I would say they either ignore IPv6 in initramfs or they use some ready to use tool like dhcpcd, Wicked or NetworkManager.
2018-04-03 13:35:42 <pavlix> clandmeter: But wasn't the question about iPXE? Or does iPXE handle it correctly already?
2018-04-03 13:35:54 <clandmeter> ipxe is ok already
2018-04-03 13:36:01 <clandmeter> thats what im saying all the time already :)
2018-04-03 13:36:36 <pavlix> clandmeter: Then what exactly do you need IPv6 in initramfs for? NFS? iSCSI?
2018-04-03 13:37:05 <clandmeter> for netboot we need to fetch modules and apk need to fetch packages.
2018-04-03 13:37:14 <pavlix> clandmeter: HTTP?
2018-04-03 13:37:19 <clandmeter> yes
2018-04-03 13:37:23 <clandmeter> or https
2018-04-03 13:37:51 <clandmeter> but we also support NFS
2018-04-03 13:37:56 <pavlix> clandmeter: From known IP or via hostname? In the case of HTTPS I guess the latter applies.
2018-04-03 13:38:18 <clandmeter> we would need dns yes.
2018-04-03 13:39:06 <pavlix> clandmeter: But then you need something that will configure it and ideally will support both IPv4 and IPv6 and will be capable of waiting for “first successful”.
2018-04-03 13:39:22 <pavlix> clandmeter: But then again I suggest dhcpcd and everything in userspace.
2018-04-03 13:39:32 <pavlix> As the most lightweight solution.
2018-04-03 13:40:18 <jirutka> pavlix: with dhcpcd you mean ISC DHCP client?
2018-04-03 13:40:41 <pavlix> jirutka: Nope. With dhcpcd I mean https://roy.marples.name/projects/dhcpcd
2018-04-03 13:40:47 <clandmeter> https://pkgs.alpinelinux.org/package/edge/main/x86_64/dhcpcd
2018-04-03 13:41:04 <pavlix> jirutka: It's a complete L3 configuration tool that simply gets you up and running.
2018-04-03 13:41:23 <pavlix> jirutka: Which requires just a little more than DHCPv4, DHCPv6 and router discovery.
2018-04-03 13:41:25 <jirutka> aha, why I thought that dhcpcd is ISC’s :/
2018-04-03 13:42:23 <pavlix> clandmeter: You can always switch to something more sophisticated if necessary.
2018-04-03 13:42:36 <_ikke_> jirutka: what is wrong with it?
2018-04-03 13:42:46 <jirutka> _ikke_: nothing
2018-04-03 13:43:08 <pavlix> _ikke_: What's wrong with ISC DHCP? Pretty much everything? ;)
2018-04-03 13:43:21 <jirutka> _ikke_: ah, you mean what’s wrong with ISC DHCP?
2018-04-03 13:43:25 <clandmeter> pavlix, i would like to prevent adding deps to initramfs if possible.
2018-04-03 13:43:33 <pavlix> clandmeter: dhcpcd is a good dep
2018-04-03 13:43:52 <pavlix> clandmeter: It is standalone and solves real problems without cluttering the image.
2018-04-03 13:43:55 <skarnet> pavlix: yes, pretty much everything
2018-04-03 13:43:55 <clandmeter> tell that to ncopa ;-)
2018-04-03 13:44:02 <jirutka> _ikke_: I just wanted to be sure, b/c I remember that it’s names of various DHCP solutions are a bit confusing
2018-04-03 13:44:28 <jirutka> clandmeter: he’s right, it’s really quite small binary without external deps https://pkgs.alpinelinux.org/package/edge/main/x86_64/dhcpcd
2018-04-03 13:44:37 <skarnet> doesn't mean it's good
2018-04-03 13:44:54 <jirutka> skarnet: yes, ofc… is there s6-dhcpc? :P
2018-04-03 13:45:05 <skarnet> busybox udhcpc works well enough for me
2018-04-03 13:45:24 <skarnet> oh, that doesn't look like the ISC dhcpcd
2018-04-03 13:45:25 <pavlix> skarnet: Did you test it in IPv6 only environment?
2018-04-03 13:45:28 <skarnet> but one from rsmarples
2018-04-03 13:45:42 <skarnet> who gave us openrc, so it's likely a double-edged sword
2018-04-03 13:45:52 <pavlix> skarnet: ISC one is called dhclient
2018-04-03 13:45:55 <jirutka> skarnet: well, it’s quite limited; I use it too, but I had problems in network with static routes etc. also IIRC does not support IPv6
2018-04-03 13:46:03 <skarnet> ah, good point
2018-04-03 13:46:13 <skarnet> but there was another dhcpcd for some time
2018-04-03 13:46:19 <skarnet> that's the one I had problems with
2018-04-03 13:46:50 <skarnet> dhcpv6 is ugh
2018-04-03 13:47:16 <terra> clandmeter: Had any chance for check out these? http://lists.alpinelinux.org/alpine-aports/5294.html http://lists.alpinelinux.org/alpine-aports/5293.html
2018-04-03 13:47:17 <skarnet> if you need it, then yeah, use whatever implements it and isn't too ugly
2018-04-03 13:47:23 <jirutka> pavlix: a bit OT question, I see that our package is outdated (6.1.5), the latest is 7.0.1, is it okay to upgrade? I didn’t find full changelog yet
2018-04-03 13:47:34 <pavlix> skarnet: I'm not saying dhcpcd is perfect or anything like that. But it's small and capable of solving all common IPv4 and IPv6 use cases. And if issues arise, there's still the posibility to fix the code and rsmarples is quite responsive.
2018-04-03 13:47:50 <skarnet> I agree
2018-04-03 13:48:01 <skarnet> if you need dhcpv6 chances are it's not a bad solution
2018-04-03 13:48:42 <pavlix> skarnet: Even if you just need to solve SLAAC+RDNSS use case along with good old IPv4.
2018-04-03 13:49:35 <jirutka> skarnet: what’s your problem with DHCPv6? I personally prefer it to RA, which seems quite… I don’t know… unconsidered?
2018-04-03 13:50:45 <skarnet> honestly ipv6 was made for SLAAC
2018-04-03 13:51:12 <skarnet> needing a dhcp with ipv6 means we're not gaining much over ipv4
2018-04-03 13:51:44 <jirutka> skarnet: and what is so great about SLAAC? what benefits does it offer for sysadmins? I quite can’t see any
2018-04-03 13:52:02 <skarnet> it's all managed by the kernel and admins have _nothing_ to do
2018-04-03 13:52:19 <skarnet> so that's quite a substantial benefit, I'd say
2018-04-03 13:52:21 <jirutka> skarnet: and what the hell did they think about when releasing SLAAC without support to set up DNS?!
2018-04-03 13:52:34 <skarnet> wat
2018-04-03 13:52:35 <jirutka> skarnet: ha ha, I mean on the router side, not client
2018-04-03 13:52:54 <jirutka> skarnet: and what’s so great about moving this to kernel instead of userspace?!
2018-04-03 13:53:08 <skarnet> you asked about sysadmin benefits
2018-04-03 13:53:18 <skarnet> having nothing to do is a nice benefit
2018-04-03 13:53:35 <skarnet> you have one router for N clients, it makes sense to move complexity over to the router
2018-04-03 13:54:03 <skarnet> I haven't looked at radvd's code very deeply, but at least it builds fine with musl :P
2018-04-03 13:54:45 <terra> I don't see ncopa here as used to.. is there something?
2018-04-03 13:55:06 <skarnet> Easter week-end, then work? :P
2018-04-03 13:55:23 <pavlix> skarnet: In my opinion (cc jirutka) the idea of doing DHCP and similar stuff in kernel is just wrong, at least in the world of initramfs.
2018-04-03 13:55:40 <jirutka> with DHCP I have full control what IP address, routes, DNS and any other parameters the clients gets, I can anytime change it and let client reconfigure etc. I don’t remember that this is possible with RA
2018-04-03 13:55:56 <skarnet> because that's not the point
2018-04-03 13:56:18 <pavlix> jirutka: And especially you have full control of the *whole configuration package* and you know exactly when you received all necessary configuration and when you dispose of it.
2018-04-03 13:56:20 <jirutka> pavlix: IMO it’s completely wrong approach at general (I’m fan of microkernels)
2018-04-03 13:56:32 <jirutka> skarnet: that ^
2018-04-03 13:56:50 <pavlix> jirutka: Anyway I have other things to do... going offline for now. You know how to reach me, do you? :)
2018-04-03 13:57:07 <skarnet> don't get me wrong, I like microkernels in theory too
2018-04-03 13:57:22 <skarnet> I just thought you were asking practical questions :P
2018-04-03 13:57:46 <jirutka> pavlix: yes I know :) I’ll call you before we start killing each other XD
2018-04-03 13:57:57 <pavlix> 
ACTION loves practical networking where you (*) receive the confiuguration, (*) apply the confiuguration and (*) use the configuration / all of that at exact known points in time.

2018-04-03 13:57:58 <skarnet> that's a good idea
2018-04-03 13:58:11 <skarnet> because if you kill each other first, then calling him will be a little harder
2018-04-03 13:58:22 <jirutka> skarnet: we are asking practical questions; why do you think microkernel approach is not practical? ;)
2018-04-03 13:58:27 <jirutka> skarnet: XD
2018-04-03 13:58:36 <jirutka> pavlix: and thank you for your help!
2018-04-03 13:59:07 <skarnet> jirutka: do we have Alpine running on a microkernel OS today? Mach does not count
2018-04-03 14:00:58 <jirutka> skarnet: unfortunately we don’t have, but I’d love to have
2018-04-03 14:01:45 <terra> libressl is 2.7.2 and we are on 2.6.4
2018-04-03 14:01:50 <skarnet> so let's start with writing a microkernel OS, then we'll wonder how to configure ipv6 on it :P
2018-04-03 14:01:53 <jirutka> terra: we know
2018-04-03 14:02:08 <jirutka> skarnet: we don’t have to, there are already some microkernels out there
2018-04-03 14:02:10 <Shiz> skarnet: shoot
2018-04-03 14:02:13 <Shiz> i was gonna say MkLinux
2018-04-03 14:02:14 <Shiz> :p
2018-04-03 14:02:26 <jirutka> Shiz: hi! :)
2018-04-03 14:02:31 <jirutka> Shiz: glad to see you!
2018-04-03 14:02:38 <Shiz> hi
2018-04-03 14:02:59 <jirutka> Shiz: could you please push LLVM 6 farther? :)
2018-04-03 14:05:41 <Shiz> I'll look at it
2018-04-03 15:01:34 <jirutka> are OpenRC devs f*** kidding me? this is NOT a changelog! https://github.com/OpenRC/openrc/blob/master/ChangeLog
2018-04-03 15:01:53 <jirutka> ah, uff, actual changelog is in https://github.com/OpenRC/openrc/blob/master/NEWS.md
2018-04-03 15:04:44 <ncopa> :D
2018-04-03 15:44:21 <jirutka> ncopa: what populates default services in runlevels?
2018-04-03 15:45:11 <jirutka> I thought that it’s handled by scripts for generating images, but I found it only in genapkovl-dhcp.sh and genapkovl-xen.sh
2018-04-03 15:51:29 <jirutka> ncopa: https://github.com/alpinelinux/aports/pull/3886
2018-04-03 15:58:33 <[[sroracle]]> nmeum: please see https://github.com/alpinelinux/aports/pull/3758 when you get a chance
2018-04-03 16:04:51 <jirutka> [[sroracle]]: do you use tlp yourself?
2018-04-03 16:05:42 <[[sroracle]]> jirutka: yes
2018-04-03 16:06:27 <jirutka> [[sroracle]]: and does it work correctly?
2018-04-03 16:06:44 <deedubs> is there a policy / best practice for embedding iptables rules / monit in packages?
2018-04-03 16:07:45 <jirutka> deedubs: yes, do not embed iptables rules in packages
2018-04-03 16:08:02 <[[sroracle]]> jirutka: As far as I can tell. The release that’s currently in the repos is broken; there’s a lot of write errors that disappear in the new version.
2018-04-03 16:09:15 <jirutka> [[sroracle]]: I’m thinking if we should move it to the community repo
2018-04-03 16:10:34 <[[sroracle]]> jirutka: up to you. I should add the caveat that I’m testing it from Adélie but afaik that shouldn’t make a difference
2018-04-03 16:11:00 <jirutka> [[sroracle]]: hm, it may; what shell do you have as /bin/sh?
2018-04-03 16:11:18 <[[sroracle]]> It does need a kernel module for additional functionality but that’s optional
2018-04-03 16:11:47 <[[sroracle]]> bash
2018-04-03 16:13:06 <jirutka> [[sroracle]]: could you please test it with Busybox ash or dash?
2018-04-03 16:13:43 <[[sroracle]]> I can test it with dash, sure
2018-04-03 16:26:30 <[[sroracle]]> jirutka: seems to work fine. Debian also has TLP packaged and they use dash as /bin/sh, and I don't see any patches for TLP there
2018-04-03 16:27:45 <jirutka> [[sroracle]]: okay, thanks! I’m moving it to the community repo
2018-04-03 16:33:25 <[[sroracle]]> jirutka: thanks
2018-04-03 17:05:03 <bernhardgruen> I was on the way to package an additional module for zabbix-agent to monitor docker hosts. At start it did not work with musl but I have an (hackish) approach to fix that. Does someone have a better idea than that? https://github.com/monitoringartist/zabbix-docker-monitoring/issues/77
2018-04-03 18:33:32 <_ikke_> ncopa: trying to (test) upgrading git to  2.17.0 (released today), and noticed that the perl.mak rule is gone (has been restructured). Do you know what the reason was that was done? (make -j1 prefix=/usr DESTDIR="$pkgdir" perl/perl.mak)
2018-04-03 18:34:31 <_ikke_> https://github.com/git/git/commit/20d2a30f8ffe14deb09b4e6fc7b70ead53b315f0
2018-04-03 18:45:32 <jirutka> _ikke_: I’ve encountered the same so postponed it :)
2018-04-03 18:46:19 <_ikke_> hehe
2018-04-03 18:52:18 <pavlix> jirutka: So what did I miss? :)
2018-04-03 19:06:49 <ncopa> perllibdir needs to be set too
2018-04-03 19:07:08 <ncopa> i can upgrade git with libressl rebuild. it needs to be rebuilt anyway
2018-04-03 19:08:03 <_ikke_> Yeah, I was just looking at perllibdir
2018-04-03 19:09:44 <_ikke_> ncopa: alright
2018-04-03 19:21:22 <ncopa> _ikke_: in case you cannot wait: http://tpaste.us/VRkv
2018-04-03 19:21:37 <_ikke_> thanks
2018-04-03 19:22:01 <ncopa> its ok if you push it early. i can alwas rebase the rebuild
2018-04-03 19:22:07 <ncopa> will probably take a couple of more days...
2018-04-03 19:24:59 <_ikke_> I have no push access anyway :-)
2018-04-03 19:25:27 <_ikke_> I don't have a hurry for it
2018-04-03 19:26:24 <ncopa> ok i leave it in my queue here then
2018-04-03 19:26:36 <ncopa> jirutka: i hope you dont get too annoyed that i'll push 200 commits with many only pkgrel bumps
2018-04-03 19:44:32 <jirutka> ncopa: it isn’t too hard to squash it ;)
2018-04-03 19:45:28 <ncopa> eh...
2018-04-03 19:46:09 <ncopa> its a happy mix of commits with patches
2018-04-03 19:46:14 <ncopa> updates
2018-04-03 19:46:15 <ncopa> etc
2018-04-03 19:46:16 <jirutka> aha
2018-04-03 19:46:29 <jirutka> so you already fixed all affected packages?
2018-04-03 19:46:32 <ncopa> not sure its worth spend another day sort it out
2018-04-03 19:54:23 <duncan^> 
ACTION beeps beep beep beep https://holeybeep.ninja/#

2018-04-03 19:55:21 <duncan^> (is alpine beep setuid?)
2018-04-03 19:55:32 <jirutka> duncan^: "Please run this command to find out: curl https://holeybeep.ninja/am_i_vulnerable.sh | sudo bash" … they’re joking, right?
2018-04-03 19:55:44 <duncan^> Not sure
2018-04-03 19:55:52 <duncan^> 
ACTION beeps for good measure

2018-04-03 19:56:03 <duncan^> 
ACTION beep beep beep

2018-04-03 19:56:12 <jirutka> duncan^: beep is symlink to busybox
2018-04-03 19:56:36 <duncan^> jirutka: No - there is https://git.alpinelinux.org/cgit/aports/tree/main/beep/APKBUILD
2018-04-03 19:56:39 <duncan^> beep
2018-04-03 19:56:42 <duncan^> whoops
2018-04-03 19:57:07 <jirutka> has anyone actually installed this, like… ever? XD
2018-04-03 19:57:09 <duncan^> I wonder if busybox beep is similarly vulnerable
2018-04-03 19:57:16 <duncan^> well it exists
2018-04-03 19:57:20 <duncan^> IDK if anyone uses it
2018-04-03 19:57:26 <duncan^> but they might like it to be updated
2018-04-03 19:57:36 <duncan^> 
ACTION beeps for their safety

2018-04-03 19:57:46 <jirutka> beep pkg does not have option="setuid", so it does not set setuid
2018-04-03 19:57:56 <duncan^> oh well
2018-04-03 19:57:59 <jirutka> btw why the heck would someone setuid beep?
2018-04-03 19:58:19 <duncan^> so that they can beep all day long without being root?
2018-04-03 19:58:21 <jirutka> anyway, I’m gonna make fun of users of major distros on Twitter XD
2018-04-03 19:58:31 <duncan^> fedora does not have beep
2018-04-03 19:58:42 <jirutka> ok, so users of Debian/Ubuntu XD
2018-04-03 19:58:53 <duncan^> Beep for their souls
2018-04-03 19:58:58 <_ikke_> zsh: command not found: beep
2018-04-03 19:58:59 <jirutka> does beep actually need setuid to work? or just usual stupidity…?
2018-04-03 20:00:30 <_ikke_> So the only thing that script does is load a module and run beep?
2018-04-03 20:00:39 <_ikke_> so they just assume that beep is vulnerable in general
2018-04-03 20:01:02 <_ikke_> Sounds like a joke to me
2018-04-03 20:01:30 <_ikke_> I think it's an april fools joke
2018-04-03 20:02:09 <duncan^> it's not
2018-04-03 20:02:18 <duncan^> it's a real vulnerability in beep
2018-04-03 20:02:28 <duncan^> >;)~
2018-04-03 20:02:41 <duncan^> well if it is, it's late
2018-04-03 20:02:51 <_ikke_> Why do they link to a youtube video about sheep
2018-04-03 20:02:53 <_ikke_> ?
2018-04-03 20:02:56 <_ikke_> https://www.youtube.com/watch?v=CZlfbep2LdU
2018-04-03 20:03:13 <jirutka> hm, this is quite suspicious… https://github.com/holeybeep/holeybeep/commits/master
2018-04-03 20:03:27 <duncan^> well the site does say it's not affiliated with the beep demon
2018-04-03 20:03:35 <duncan^> sorry the researcher who found the vuln in beep
2018-04-03 20:04:02 <_ikke_> If it's a real vulnerability, that site doesn't do justice to it
2018-04-03 20:04:10 <duncan^> The site is a joke site yes
2018-04-03 20:04:17 <duncan^> https://www.debian.org/security/2018/dsa-4163
2018-04-03 20:04:38 <_ikke_> local privilege escalation
2018-04-03 20:05:47 <_ikke_> I don't think alpines beep is setuid: https://git.alpinelinux.org/cgit/aports/tree/main/beep/APKBUILD
2018-04-03 20:05:53 <jirutka> https://nvd.nist.gov/vuln/detail/CVE-2018-0492
2018-04-03 20:07:04 <duncan^> I mean
2018-04-03 20:07:08 <duncan^> it could be an April fool
2018-04-03 20:26:29 <hs3dUBwdmCjy> jirutka for some reason beep is a predependency of mcabber, so all mcabber users have beep installed it seems.
2018-04-03 20:31:03 <jirutka> hs3dUBwdmCjy: huh, we should remove this dependency from mcabber then
2018-04-03 20:34:08 <jirutka> sigh, I really hate this hype-driven culture where something that is not changing rapidly each year is dead… this future of personal computing looks horrible https://www.theverge.com/2018/4/3/17191986/apple-intel-cpu-processor-design-competition
2018-04-03 21:11:06 <Truxx> Could someone maybe point in the right direction? There is the modemmanager package, but it did not work (error message "no modemmanager im bus").
2018-04-03 21:11:31 <Truxx> When I tried to run modemmanager, I received: "Could not acquire the 'org.freedesktop.ModemManager1"
2018-04-03 21:12:29 <terra> Truxx: is dbus service  running?
2018-04-03 21:12:46 <Truxx> Yes, dbus was added, it is running
2018-04-03 21:13:17 <Truxx> rc-update add dbus - right after install
2018-04-03 21:13:40 <terra> so you did start dbus manually
2018-04-03 21:14:03 <Truxx> I found some threads with this error message, but nothing alpine specific or helpful
2018-04-03 21:14:18 <Truxx> @terra I did NOT start dbus manually
2018-04-03 21:14:49 <Truxx> Added to start it automatically and rebooted since then several times
2018-04-03 21:15:15 <Truxx> (Only the first time I started it manually)
2018-04-03 21:15:31 <Truxx> (rc-service dbus start)
2018-04-03 21:15:54 <terra> dbus-launch modemmanager ?
2018-04-03 21:17:10 <Truxx> Id
2018-04-03 21:17:18 <Truxx> I did not try that yet...
2018-04-03 21:17:49 <mps> regarding beep, 'beep' on terminal
2018-04-03 21:17:52 <mps> Could not open /dev/tty0 or /dev/vc/0 for writing
2018-04-03 21:18:57 <terra> mps: are you in "tty" group ?
2018-04-03 21:19:18 <mps> no
2018-04-03 21:23:01 <terra> then add yourself into "tty" group then relogin after logoff. Or use a login shell.
2018-04-03 21:24:36 <mps> terra: I don't need that, just pointed to previous discussion about beep vulnerability
2018-04-03 21:25:24 <terra> mps: sorry for noise then.
2018-04-03 21:27:14 <mps> terra: never mind, irc is noisy anyway :)
2018-04-03 21:27:42 <Truxx> @terra Thank you for the suggestion. Meanwhile I found this: https://ask.fedoraproject.org/en/question/41271/how-to-start-a-mobile-broadband-connection-in-fedora-20/  Will check how it works out.
2018-04-03 21:58:53 <Truxx> @terra Tried dbus-launch -> it's the same when trying only ModemManager, the same error message
2018-04-03 22:18:38 <Truxx> Isn't it possible that it's because of systemd?
2018-04-03 22:19:30 <Truxx> /etc/dbus-1/system.d - I guess this is where ModemManager looks for that file
2018-04-03 22:22:04 <Truxx> I have in that folder org.freedesktop.UPower.conf or ConsoleKit, but no org.freedesktop.ModemManager1.
2018-04-03 22:53:49 <Truxx> Someone figured out a solution by using "sudo mv /usr/share/dbus-1/system-services/org.freedesktop.ModemManager1.service /usr/share/dbus-1/services/"
2018-04-03 22:54:11 <Truxx> However that .service file is completely missing for me
2018-04-03 22:54:41 <skarnet> Truxx: is your problem Alpine-related, or systemd-related? It can't be both.
2018-04-03 22:55:13 <Truxx> Alpine-related I guess as I don't use systemd here and only installed modemmanager from the repo
2018-04-03 22:55:36 <skarnet> then .service files will be entirely irrelevant
2018-04-03 22:55:46 <skarnet> since they're systemd service definitions
2018-04-03 22:57:13 <Truxx> Thank you. At least I've got a direction then. What am I supposed to do with "Could not acquire the 'org.freedesktop.ModemManager1' service name"?
2018-04-03 22:57:39 <Truxx> This is why ModemManager is not in bus, i.e. it does not start.
2018-04-03 22:58:22 <Truxx> I was guessing it is looking for that .service file in /usr/share/dbus-1/services
2018-04-03 22:59:47 <skarnet> I can't help you much with it, sorry. I tend to stay as far away from D-Bus as I can.
2018-04-03 23:00:20 <skarnet> but I suppose modemmanager is trying to talk to D-Bus and somehow fails.
2018-04-03 23:00:42 <skarnet> you probably need a dbus-daemon running before launching modemmanager, or something like that.
2018-04-03 23:00:49 <Truxx> It does not start
2018-04-03 23:00:56 <skarnet> what doesn't?
2018-04-03 23:00:58 <Truxx> dbus is running
2018-04-03 23:01:03 <Truxx> modemmanager does not start
2018-04-03 23:01:16 <skarnet> well of course it's not.
2018-04-03 23:01:23 <Truxx> becasue of that error message
2018-04-03 23:02:04 <skarnet> but if dbus-daemon is running, then it's something more complex, and it's D-Bus land, and I won't try and debug anything in it.
2018-04-03 23:02:19 <skarnet> Sorry about that. I hope you'll be able to solve it.
2018-04-03 23:02:46 <Truxx> I see... Thank you for the hints. Maybe here is a dbus channel on irc...
2018-04-03 23:03:46 <skarnet> there certainly is. It's even possible that 2 or 3 people on this channel know what they're talking about. XD
2018-04-03 23:03:58 <skarnet> this = the d-bus channel, I mean.
2018-04-03 23:05:50 <Truxx> Yes, there is dbus. I'm gonna give it a try...
2018-04-03 23:07:13 <skarnet> gl.
2018-04-04 09:25:30 <ncopa> morning
2018-04-04 09:25:32 <ncopa> re beep
2018-04-04 09:25:50 <ncopa> did you guys look at the am_i_vulnerable.sh script?
2018-04-04 09:26:09 <ncopa> it modprobes pcsprk and run beep
2018-04-04 09:26:24 <ncopa> it means that the vulnerability is that you run curl $script | sudo bash
2018-04-04 09:27:47 <ncopa> the am_i_vulnerable.sh script does not reveal any real privilege escalation
2018-04-04 09:29:04 <ncopa> fwiw, i have use been in production, on headless machines
2018-04-04 09:29:07 <ncopa> it was ran at the end of the boot to indicate that boot was successful
2018-04-04 09:30:28 <ncopa> that said, if beep is a suid root app, then its very possible to be a real sec issue
2018-04-04 09:31:50 <mps> I installed mcabber and it pulled beep, but beep is not suid by default install
2018-04-04 09:32:26 <ncopa> yes, i dont think we suid root our beep
2018-04-04 09:32:52 <ncopa> it may still be a bug in busybox beep
2018-04-04 09:33:32 <mps> so, that means it is not 'end of the word' problem, although beep should be fixed
2018-04-04 09:33:44 <ncopa> exactly
2018-04-04 09:33:48 <mps> s/word/world/
2018-04-04 10:31:18 <_ikke_> ncopa: they say it was this version that had a privilege escalation: http://www.johnath.com/beep/
2018-04-04 10:36:06 <_ikke_> https://github.com/johnath/beep/issues/11
2018-04-04 11:43:09 <jirutka_> problem with cross-compiling using abuild: https://dpaste.de/Faak/raw -- any clue? (I will post more details later)
2018-04-04 11:46:29 <duncan^> Well the website is a joke
2018-04-04 11:46:36 <duncan^> it is called holeybeep.ninja ffs
2018-04-04 11:47:03 <duncan^> But there is a real CVE that didn't turn up on April 1 so I assume i't sreal
2018-04-04 11:47:18 <duncan^> or not...
2018-04-04 11:54:02 <ncopa> not sure what cargo is trying to do. is it trying to execute aarch64-alpine-linux-musl/bin/as when it gets the error?
2018-04-04 11:54:54 <ncopa> jirutka: is bootstrap-rust built with static libgit?
2018-04-04 11:54:57 <ncopa> can i upgrade libgit2?
2018-04-04 11:55:14 <ncopa> some of the tests are failing with libgit2-0.26
2018-04-04 11:55:24 <ncopa> it looks like par to f the test is trying to connect to an external git site
2018-04-04 11:55:40 <ncopa> where they seem to have introduced authentication or changed password or something
2018-04-04 11:55:43 <ncopa> so test fails
2018-04-04 11:56:40 <jirutka_> ncopa: rustc does not link libgit, cargo does, our bootstrapping cargo is static binary
2018-04-04 11:58:41 <ncopa> ok, i remember it used to be dynamically linked to libgit2
2018-04-04 11:58:46 <ncopa> but that is fixed then
2018-04-04 11:58:46 <ncopa> thanks
2018-04-04 11:59:02 <jirutka_> ncopa: I think that cargo executes that gcc command line and somehow gets the errors below
2018-04-04 11:59:34 <jirutka_> ncopa: I’m really not sure what the heck it’s doing; so it’s not any common problem with cross building?
2018-04-04 11:59:39 <ncopa> what machine is it running on? an arch64 bachine?
2018-04-04 11:59:46 <ncopa> machine*
2018-04-04 11:59:46 <ncopa> with glibc?
2018-04-04 11:59:54 <ncopa> aarch64*
2018-04-04 12:00:00 <jirutka_> no, Alpine x86_64, cross-compiling to aarch64
2018-04-04 12:00:02 <ncopa> i dont have much experience with cross-compiling
2018-04-04 12:00:26 <jirutka_> CBUILD=x86_64-alpine-linux-musl CHOST=aarch64-alpine-linux-musl CTARGET=aarch64-alpine-linux-musl
2018-04-04 12:00:33 <jirutka_> me neither :(
2018-04-04 12:01:11 <jirutka_> and sane projects provide better support for cross compiling, e.g. it was super easy with Crystal, but Rust devs are *****
2018-04-04 12:02:08 <jirutka_> afk
2018-04-04 12:02:13 <ncopa> can you crosscompile a helloworld.c?
2018-04-04 12:20:16 <ncopa> Your branch is ahead of 'origin/master' by 121 commits.
2018-04-04 12:20:41 <ncopa> i think I am around half way through the libressl rebuilds
2018-04-04 12:21:40 <ncopa> i will hopefully be able to push it at the end of the week or early next week
2018-04-04 12:57:41 <jirutka_> ncopa: do you really think it’s worth it? wouldn’t it be easier to migrate to openssl right now instead of updating to newer libressl and patching packages and migrate to openssl just month or two later?
2018-04-04 18:31:35 <HRio> HI somebody with commit rights here? we just merged XEN 4.7.5 into v3.5, but xenproject just pulled 4.7.5, can we revert?
2018-04-04 18:32:12 <_ikke_> No commit rights, but you mean they pulled that version back?
2018-04-04 18:32:22 <HRio> https://lists.xenproject.org/archives/html/xen-announce/2018-04/msg00001.html
2018-04-04 18:32:31 <HRio> yes, no download anymoer
2018-04-04 18:32:47 <HRio> https://xenproject.org/downloads/xen-archives/xen-project-47-series.html 4.7.5 is not listed anymoer
2018-04-04 18:32:52 <HRio> anymore
2018-04-04 18:33:50 <jirutka> hi
2018-04-04 18:33:53 <_ikke_> HRio: At least make a ticket in bugs.a.o
2018-04-04 18:34:52 <jirutka> so reject the last xen commit in 3.5-stable?
2018-04-04 18:35:12 <HRio> https://bugs.alpinelinux.org/issues/8763
2018-04-04 18:35:18 <HRio> yes
2018-04-04 18:35:36 <jirutka> HRio: please open a PR or send me commit as patch here, I’ll push it
2018-04-04 18:36:09 <jirutka> HRio: include ref to #8763 and XEN announcement in the commit msg
2018-04-04 18:40:42 <HRio> https://github.com/alpinelinux/aports/pull/3907
2018-04-04 18:41:29 <kaniini> god
2018-04-04 18:41:58 <kaniini> this is why i dont use Xen anymore
2018-04-04 18:42:05 <Xe> yikes
2018-04-04 18:44:37 <jirutka> holly shit, that abuild is insane, so many patches!
2018-04-04 18:45:33 <jirutka> kaniini: when I just revert this commit, will builders rebuild it, even when pkgver will be older than the last available? IIRC it didn’t work when I’ve tried to downgrade some pkg last time
2018-04-04 18:46:18 <_ikke_> jirutka: You need to always increase pkgrel
2018-04-04 18:46:27 <_ikke_> jirutka: for dl-cdn
2018-04-04 18:47:20 <jirutka> _ikke_: okay, but still $pkgver-v$pkgrel will be lower than available packages
2018-04-04 18:55:03 <terra> can "apk upgrade" revert a package for an older version?
2018-04-04 18:56:13 <_ikke_> jirutka: it seems it builds the packages as long as there are new commits: https://git.alpinelinux.org/cgit/aports/tree/main/aports-build/aports-build#n137
2018-04-04 18:56:17 <_ikke_> if I'm reading it right
2018-04-04 18:56:37 <_ikke_> terra: I think you might need to add --available
2018-04-04 18:57:09 <terra> APKINDEX.gz may contain such information for reverted packages
2018-04-04 19:00:07 <terra> because most users may not aware such revision changes
2018-04-04 19:07:59 <jirutka> kaniini: what do you think about automatic updates of OS?
2018-04-04 19:08:32 <Xe> ie via the wrath of cron?
2018-04-04 19:08:42 <azarus> Is there a way to build linux-vanilla only for x86_64? (not wasting time on other arches)
2018-04-04 19:08:46 <skarnet> I'm not kaniini, but I'm all for automatic updates... as long as they're opt-in
2018-04-04 19:08:49 <jirutka> kaniini: how do other distros handle it? it’s easy to just update packages, but useless when it does not reload affected services; but how to know which to reload or restart? etc.
2018-04-04 19:10:03 <jirutka> also how can I tell apk to upgrade everything except kernel, without explicitly pinning kernel version in world?
2018-04-04 19:11:32 <terra> ALpine devs love kernel upgrades even though there are packages to be waiting for to be reviewed for a month.
2018-04-04 19:11:52 <_ikke_> jirutka: I usually use lsof and grep files which are deleted
2018-04-04 19:12:17 <Xe> terra: which packages?
2018-04-04 19:12:34 <terra> mame
2018-04-04 19:13:03 <Xe> does mame have a test suite?
2018-04-04 19:13:38 <terra> no
2018-04-04 19:13:53 <Xe> that's probably why it's taking so long
2018-04-04 19:14:14 <jirutka> _ikke_: lsof does not list loaded libraries, does it?
2018-04-04 19:14:30 <azarus> argh -- disabled so many useless modules in my kernel config only to have the build slowed down by other "one size fits all" configs :/
2018-04-04 19:14:49 <jirutka> _ikke_: for example, when I run nginx and upgrade libressl, I need to reload nginx to load new libcrypto.so
2018-04-04 19:15:17 <terra> Xe: I think that's not the case because I'he submitted many packages without any test suite.
2018-04-04 19:15:36 <jirutka> azarus: Alpine’s kernel configs are insane, probably better to start from zero and make very custom config for your HW
2018-04-04 19:15:43 <azarus> jirutka: I did
2018-04-04 19:15:52 <azarus> I just want to use abuild for it
2018-04-04 19:16:03 <azarus> and it compiles all other configs as well
2018-04-04 19:16:03 <terra> I noticed many times but snes9x didn't get a final review either.
2018-04-04 19:17:02 <azarus> I'm just following this guide: https://wiki.alpinelinux.org/wiki/Custom_Kernel
2018-04-04 19:17:21 <terra> azarus: arch="x86_86"
2018-04-04 19:17:21 <azarus> Or should one just not use abuild for a custom kernel?
2018-04-04 19:17:32 <terra> azarus: arch="x86_64" <<--
2018-04-04 19:17:35 <azarus> OK.
2018-04-04 19:17:47 <azarus> That's set in /etc/abuild.conf, I presume?
2018-04-04 19:17:55 <terra> APKBUILD
2018-04-04 19:18:02 <azarus> Ah, ok.
2018-04-04 19:18:06 <jirutka> azarus: try Gentoo guide for building kernel; abuild part is simple, but now you need actual .config
2018-04-04 19:18:25 <azarus> I've been using Gentoo for years, kernel config is familiar to me
2018-04-04 19:18:26 <jirutka> azarus: then read https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Setup_your_system_and_account
2018-04-04 19:18:46 <_ikke_> jirutka: I'm not certain it's exhaustive, but it does show linked libraries
2018-04-04 19:19:01 <azarus> jirutka: followed that guide, skipped some parts. whoops
2018-04-04 19:19:25 <_ikke_> jirutka: lsof | grep -e DEL | grep lib | sort -u -k1,1
2018-04-04 19:19:26 <jirutka> azarus: and probably better to write abuild from scratch for that, our kernel abuilds are complex because of support of many flavours and arches, you don’t need most of it
2018-04-04 19:19:28 <_ikke_> that's what I usually run
2018-04-04 19:19:42 <jirutka> _ikke_: aha, I probably need fullblown lsof, not busybox lsof
2018-04-04 19:19:49 <_ikke_> ah, right
2018-04-04 19:20:18 <HRio> jirutka: this is used in Debian https://tracker.debian.org/pkg/needrestart to track what to restart after a library upgrade
2018-04-04 19:20:37 <jirutka> yeah, lsof from lsof pkg provide *much* more information
2018-04-04 19:21:07 <HRio> github project for needrestart https://github.com/liske/needrestart
2018-04-04 19:21:29 <jirutka> HRio: thanks, I’ll look at it
2018-04-04 19:22:06 <jirutka> HRio: however, there still must be some whitelist or blacklist, for example I definitely don’t want automatically restart PostgreSQL server
2018-04-04 19:22:21 <HRio> jirutka: same developer for apt-dater
2018-04-04 19:22:45 <HRio> yes I newer alow it to restat anything xen related on Debian dom0s
2018-04-04 19:24:58 <jirutka> oh my… I’m susprised that JVM haven’t run out of available file descriptors, it has like gazillion opened files o.O
2018-04-04 19:26:10 <azarus> Hrm. Now I'm not sure if I should just write my own APKBUILD for a custom kernel :/
2018-04-04 19:26:36 <_ikke_> azarus: You can probably copy an existing one
2018-04-04 19:26:48 <jirutka> azarus: why not? it’s really easy once you set up environment (all described on the wiki i linked)
2018-04-04 19:27:12 <jirutka> _ikke_: that’s not a good idea, our kernel aports are too complex b/c of support of more flavours and archs
2018-04-04 19:27:32 <azarus> Hmm. Will take some time, but it'd be nice
2018-04-04 19:27:38 <azarus> "linux-custom" or something
2018-04-04 19:28:10 <azarus> I do use ZFS and wireguard, which are out of tree. Any way to build those easily as well?
2018-04-04 19:28:55 <azarus> 
ACTION takes a peek at the zfs and zfs-kmod APKBUILDs

2018-04-04 19:41:30 <HRio> jirutka: looks like pkgs.a.o has noticed the downgrade at least https://pkgs.alpinelinux.org/packages?name=&branch=v3.5  thanks!
2018-04-04 19:42:08 <jirutka> just fyi, this is what VoidLinux use for finding what processes needs to be restarted after library upgrade: https://github.com/chneukirchen/xtools/blob/master/xcheckrestart
2018-04-04 20:12:15 <HRio> jirutka: using gnu grep and "apk info --who-owns" instead of "xbps-query -o" lists /usr/sbin/nginx and /usr/bin/php-fpm5 after upgrading ssl libs
2018-04-04 20:19:39 <jirutka> HRio: yeah; the question is how to find runscript; it’s not sufficient to just query the package files for /etc/init.d/*, because it may be in a subpackage, so we must query all subpackages
2018-04-04 20:19:49 <jirutka> HRio: or grep runscripts instead?
2018-04-04 20:20:27 <jirutka> HRio: that may work for good runscripts that define command variable, but not for crappy ones
2018-04-04 20:21:10 <HRio> yes thats a problem, there will be some runscripts to clean-up if we will grep them for this
2018-04-04 20:21:14 <jirutka> HRio: now I expect skarnet to come in and tell that if I used s6, then I can just kill the process and let supervisor restart it… :)
2018-04-04 20:21:49 <skarnet> I don't need to interfere, since you already know it :P
2018-04-04 20:21:50 <HRio> also /proc/<pid>/exe will give /bin/busybox quite often on alpine :-)
2018-04-04 20:22:13 <HRio> openrc can supervise as well...
2018-04-04 20:22:19 <Xe> skarnet: killall5 pls
2018-04-04 20:22:28 <HRio> https://github.com/OpenRC/openrc/blob/master/supervise-daemon-guide.md
2018-04-04 20:22:28 <jirutka> I know, but… skarnet would tell you :)
2018-04-04 20:22:31 <skarnet> HRio: yes, badly
2018-04-04 20:22:40 <jirutka> as I said ^ :)
2018-04-04 20:22:42 <HRio> skarnet: :-)
2018-04-04 20:23:33 <skarnet> Xe: killall5 stops all processes, scans /proc, then restart processes. You don't want to do that if you can avoid it.
2018-04-04 20:23:49 <jirutka> however, maybe I’m just overengineering  now, because I need whitelist anyway, so I can also specify runscript instead of just process/pkgname
2018-04-04 20:36:27 <jirutka> andypost: https://bugs.alpinelinux.org/issues/8764
2018-04-04 20:37:06 <jirutka> andypost: I wonder what (s)he did wrong, because it’s quite not possible that modules are built against older API, php7-gd is subpackage of php7
2018-04-04 20:37:13 <jirutka> andypost: maybe incomplete upgrade? o.O
2018-04-04 20:37:22 <jirutka> andypost: …on the user side
2018-04-04 22:47:44 <andypost> jirutka, looks like he downloading old packages
2018-04-04 22:53:39 <andypost> btw I'm trying to run alpine:edge docker image but it downloading from 3.7 repo(
2018-04-04 22:54:25 <andypost> that's happens on arm8 scaleway... so probably issue in image
2018-04-04 22:54:38 <andypost> docker pull alpine:edge
2018-04-04 23:25:04 <jirutka> andypost: aha, if he use docker, then it explains a lot… probably yet another broken image
2018-04-04 23:26:37 <jirutka> goosh, why is ARM so slooooow, it takes eternity to build something on it :(
2018-04-05 05:57:34 <bernhardgruen> Good morning together, when a lib goes to /usr/lib64 and when does it go to /usr/lib ? current example: librsync 2.0.2 is in /usr/lib64 - but nothing else on my system.
2018-04-05 07:01:44 <ncopa2> bernhardgruen: we dont support multilib so we dont use /usr/lib64 at all. everything goes into /usr/lib
2018-04-05 07:02:22 <ncopa2> cargo is being annoying on me (again)
2018-04-05 07:02:29 <bernhardgruen[m]> ncopa2: then librsync seems to be packaged wrong.
2018-04-05 07:03:07 <ncopa2> This crate is only compatible with OpenSSL 1.0.1, 1.0.2, and 1.1.0, or LibreSSL
2018-04-05 07:03:07 <ncopa2> 2.5 and 2.6.0, but a different version of OpenSSL was found. The build is now
2018-04-05 07:03:07 <ncopa2> aborting due to this version mismatch.
2018-04-05 08:04:51 <azarus> Okay, trying to write my own APKBUILD for a custom kernel, but I am stuck here: https://bpaste.net/show/c2eeb54f5993 this is my APKBUILD: https://ptpb.pw/86mt and my kernel config: https://ptpb.pw/tsRG
2018-04-05 08:05:11 <azarus> I'm kinda confused at what the dev() function really does, and how subpackages work.
2018-04-05 08:07:29 <_ikke_> azarus: A subpackage just means that some files are placed in a separate package
2018-04-05 08:08:02 <_ikke_> azarus: It does this by having a separate 'root' for the package and subpackages (pkgdir vs subpkgdir)
2018-04-05 08:08:29 <_ikke_> by moving files from the pkgdir to the subpackage dir, abuild can create separate packages for them
2018-04-05 08:09:44 <azarus> _ikke_: Thanks for the explanation.
2018-04-05 08:10:43 <azarus> Now, it tries to cd into main/linux-custom/pkg/linux-custom-dev; should that path be manually created?
2018-04-05 08:11:09 <_ikke_> azarus: did yoy create a custom dev function?
2018-04-05 08:11:16 <azarus> yes.
2018-04-05 08:11:34 <_ikke_> do you want to override the default behaviour, or just supplement it?
2018-04-05 08:12:03 <azarus> Not sure if what I want to do is possible with the default behaviour.
2018-04-05 08:12:34 <azarus> I'd like to create the same -dev packages as linux-vanilla and the such have
2018-04-05 09:06:42 <ncopa> Your branch is ahead of 'origin/master' by 197 commits.
2018-04-05 09:07:00 <ncopa> im making progress with libressl 2.7
2018-04-05 09:07:10 <ncopa> the only blocker so far is cargo
2018-04-05 09:07:35 <azarus> Cool! Is it proving to break many applications? :o
2018-04-05 09:09:17 <ncopa> libressl 2.7 seems to work pretty well
2018-04-05 09:09:48 <ncopa> some patches are needed. mostly it is enough to remove " || defined(LIBRESSL_VERSION_NUMBER)"
2018-04-05 09:10:02 <ncopa> because libressl now implements some of the openssl 1.1. api
2018-04-05 09:10:02 <azarus> ah, some macro faff
2018-04-05 09:10:37 <ncopa> cargo is painful because it fetches code from github which breaks
2018-04-05 09:10:43 <ncopa> so it needs to be fixed upstream
2018-04-05 09:10:51 <ncopa> and it looks like it is already fixed upstream
2018-04-05 09:10:57 <ncopa> but not new release tag
2018-04-05 09:11:13 <ncopa> so i cannot tell cargo to pick the latest version (yet)
2018-04-05 09:11:22 <ncopa> and i cannot patch it
2018-04-05 09:11:43 <ncopa> and #cargo on irc.mozilla.org is quiet about it
2018-04-05 09:17:29 <azarus> argh :/
2018-04-05 09:22:34 <terra> ncopa: libressl seems add DTSL support but uses old DTLSv1_method() instead DTLS_method()
2018-04-05 09:23:06 <ncopa> ok?
2018-04-05 09:23:33 <terra> gst-plugins-bad expects DTLS_method()
2018-04-05 09:23:51 <terra> which is openssl compatible
2018-04-05 09:25:27 <terra> I want to send gtreamer 1.14.0 upgrade but waiting for libressl 2.7.*
2018-04-05 09:25:54 <HRio> ncopa: do you know if anybody is working on a backport of the retpoline stuff for gcc (indirect thunk)? found this branch https://github.com/hjl-tools/gcc/commits/hjl/indirect/gcc-6-branch/master that might be of interest
2018-04-05 09:26:38 <azarus> why not upgrade to gcc 7 ?
2018-04-05 09:28:32 <HRio> azarus: alpine v3.7 and older
2018-04-05 09:28:59 <azarus> ah, then
2018-04-05 09:29:00 <azarus> 
ACTION shrugs

2018-04-05 09:30:04 <ncopa> terra: i have gstream-1.14 in the queue together with libressl-2.7
2018-04-05 09:30:39 <ncopa> HRio: i dont think anyone is workgin on backporting retpoline for gcc6
2018-04-05 09:30:49 <ncopa> i think fabled is working on upgrade to gcc7
2018-04-05 09:31:00 <terra> ncopa: I checked build options and there are some unrelated (or deprecated) configure options and inconsistencies
2018-04-05 09:31:29 <terra> I'm talking about gstreamer btw
2018-04-05 09:32:18 <terra> so I will send aports mail list in case of you may review
2018-04-05 09:32:34 <terra> I updated them in my local repo anyway
2018-04-05 09:34:49 <ncopa> terra: this is my WIP with libressl: https://github.com/ncopa/aports/tree/libressl-2.7
2018-04-05 09:35:24 <ncopa> if you creater a patch on top of that i'll merge it and push it with the libressl-2.7 upgrade
2018-04-05 09:38:54 <terra> Ok. that's just a simple replace of function name. But other thing that I've changed needs to be considered.
2018-04-05 09:39:22 <terra> I added qt and gtk3 plugins for gstreamer-1.0 for example
2018-04-05 09:43:40 <terra> ncopa: I will follow you after.
2018-04-05 09:49:23 <ncopa> hum
2018-04-05 09:49:28 <ncopa> there are some silent errors
2018-04-05 09:49:35 <ncopa> for example python3
2018-04-05 09:49:59 <ncopa> it just didnt build the ssl module
2018-04-05 09:55:33 <ncopa> which blow up in my face later
2018-04-05 10:16:01 <terra> ncopa: I've send the patch to your email
2018-04-05 10:17:05 <ncopa> thanks!
2018-04-05 10:22:25 <ncopa> terra: can you please commit it  and do git format-patch -1
2018-04-05 10:22:33 <ncopa> and attach that in an email?
2018-04-05 10:22:45 <ncopa> or git send-email -1 --to ncopa@alpinelinux.org
2018-04-05 10:22:54 <ncopa> so i can `git am` it
2018-04-05 10:23:05 <ncopa> otherwise i need to commit it with --author etc
2018-04-05 10:23:53 <terra> ok
2018-04-05 10:58:41 <terra> ncopa: I've sent
2018-04-05 12:33:41 <chambbj> Requesting final review (and merge) of https://github.com/alpinelinux/aports/pull/3901. Thx!
2018-04-05 12:38:52 <triple> skarnet: The s6 package has broken dependencies in edge: 'required by: s6-2.7.1.1-r0[so:libexecline.so.2.3]'
2018-04-05 12:40:45 <skarnet> huh?
2018-04-05 12:41:21 <skarnet> the latest execline apk is 2.3.0.4
2018-04-05 12:41:26 <skarnet> so, the so should be there
2018-04-05 12:43:06 <triple> Should not depend on libexecline according to https://pkgs.alpinelinux.org/package/edge/main/x86_64/s6
2018-04-05 12:43:29 <triple> execline even
2018-04-05 12:43:42 <triple> In edge, but is there in 3.7
2018-04-05 12:44:49 <skarnet> has apk changed? the execline dependency isn't declared, but automatically found by apk
2018-04-05 12:45:09 <skarnet> it's weird because the skalibs dep is also autofound
2018-04-05 12:55:39 <triple> skarnet: The latest execline is execline-2.5.0.0-r0 in edge though
2018-04-05 12:57:17 <skarnet> well that's fine, it should work too
2018-04-05 12:57:36 <skarnet> the dep should resolve as libexecline >= libexecline.so.2.3
2018-04-05 12:57:52 <skarnet> ah, I think I know what happened
2018-04-05 12:58:04 <skarnet> the apk was built with execline-2.3.0.4
2018-04-05 12:58:06 <triple> Did you compile with old libs?
2018-04-05 12:58:10 <triple> Aha
2018-04-05 12:58:31 <skarnet> but the dep solver is too strict here
2018-04-05 12:59:25 <skarnet> should I explicitly list package dependencies with >= instead of relying on apk to autolist them?
2018-04-05 13:00:15 <skarnet> because if apk makes deps "this version of .so" instead of "this version or higher" that's not what I want
2018-04-05 13:00:59 <skarnet> alternatively I could just abump to s6-2.7.1.1-r1, it would pick up the new execline version, but it won't solve the underlying issue
2018-04-05 14:03:17 <clandmeter> ncopa, i wonder how much time we could save when using pigz and axel for src files.
2018-04-05 14:09:31 <bernhardgruen> There is a bug in librsync aport (installs to /usr/lib64). PR incoming.
2018-04-05 14:11:31 <ncopa> bernhardgruen: i think i have it in my queue already
2018-04-05 14:11:53 <ncopa> i looked at librsync since you mentioned it and you are right
2018-04-05 14:12:28 <andypost> which is default/suggested version of python?
2018-04-05 14:12:36 <bernhardgruen> ok - shit. I investigated 2 days into this because I first thought my burp update is wrong.
2018-04-05 14:12:48 <ncopa> bernhardgruen: http://tpaste.us/aMqN
2018-04-05 14:13:28 <bernhardgruen> your prefix line is doubled - other than that my file looks identical.
2018-04-05 14:13:55 <ncopa> bernhardgruen: can you git format-patch -1 --stdout | tpaste it?
2018-04-05 14:14:00 <ncopa> then i'll merge it here and now
2018-04-05 14:14:23 <ncopa> andypost: default python is python2, i'd recommend use python3 whenever possible
2018-04-05 14:14:50 <ncopa> and only use python2 whenever python3 is not possible
2018-04-05 14:15:17 <andypost> ncopa, is it fine if I add one more commit to https://github.com/alpinelinux/aports/pull/3916 to make conversion?
2018-04-05 14:15:42 <_ikke_> Any plans making python default to python3?
2018-04-05 14:16:12 <bernhardgruen> ncopa: http://tpaste.us/qoJz
2018-04-05 14:16:21 <ncopa> andypost: yeah i think thats fine
2018-04-05 14:16:32 <andypost> grep shows 2937/1654 in repo for 3 vs 2
2018-04-05 14:17:18 <andypost> oh py3 1654 vs py2 2937
2018-04-05 14:17:43 <ncopa> bernhardgruen: pushed. thanks!
2018-04-05 14:18:32 <ncopa> _ikke_: not really. i think there is some legacy thing that makes most python apps expect /usr/bin/python to be python2
2018-04-05 14:18:51 <ncopa> i also think that the recommendation from upstream is to use python3
2018-04-05 14:18:56 <ncopa> instead of just "python"
2018-04-05 14:19:51 <ncopa> arch linux made python to be python3, and i think they did it before upstream came up with the recommendation
2018-04-05 14:20:01 <_ikke_> right
2018-04-05 14:20:06 <_ikke_> Probably because of it
2018-04-05 14:20:19 <ncopa> possibly
2018-04-05 14:20:23 <_ikke_> But eventually python2 will be EOL
2018-04-05 14:20:32 <ncopa> yup
2018-04-05 14:20:58 <ncopa> and at that point everyone should be using #!/usr/bin/python3 - if i understand things correctly
2018-04-05 14:21:15 <ncopa> but i dont think it will happen :)
2018-04-05 14:21:37 <ncopa> many people insist on staying on python2
2018-04-05 14:21:42 <_ikke_> yup, sadly
2018-04-05 14:22:02 <ncopa> i read the other day that they do that becuase python 3.5 -> 3.6 breaks their apps
2018-04-05 14:24:17 <_ikke_> So you stay on python2 for that reason?
2018-04-05 14:24:28 <ncopa> im saying, some people does
2018-04-05 14:24:30 <_ikke_> Not even sure what would break their apps
2018-04-05 14:24:49 <ncopa> its a tricky thing
2018-04-05 14:24:55 <_ikke_> Possibly the StopIteration change
2018-04-05 14:25:04 <ncopa> make new features -> break backwards compat
2018-04-05 14:25:09 <ncopa> people gets angry
2018-04-05 14:25:18 <ncopa> no new features -> people ge angry too
2018-04-05 14:25:19 <ncopa> :)
2018-04-05 14:25:26 <_ikke_> catch22
2018-04-05 14:26:08 <andypost> at least fail2ban provides conversion script that works now)
2018-04-05 14:27:25 <ncopa> i wonder if i should push libressl-2.7 as soon im done with main and community, and do the testing rebuild while builders are building main and community
2018-04-05 14:27:37 <ncopa> or push everything in one shot
2018-04-05 14:27:58 <Shiz> ncopa: it's not e alegacy thing, it's codified in a PEP that it should be ike that
2018-04-05 14:28:00 <Shiz> :P
2018-04-05 14:28:12 <Shiz> https://www.python.org/dev/peps/pep-0394/
2018-04-05 14:28:30 <ncopa> yeah, i think arch linux did their thing before that PEP was created
2018-04-05 14:29:06 <ncopa> wow 2011
2018-04-05 14:29:21 <bernhardgruen> ncopa, I think fastest way would be to push as soon as possible (only main and community).
2018-04-05 14:29:36 <ncopa> "for the time being, all distributions should ensure that python refers to the same target as python2."
2018-04-05 14:29:46 <ncopa> bernhardgruen: thats what im thinking too
2018-04-05 14:30:06 <ncopa> it means that things in testing will not build
2018-04-05 14:30:18 <bernhardgruen> why do you at all have to change all those APKBUILDs? Isn't there any automatism?
2018-04-05 14:30:42 <ncopa> and that people using testing repo will not get rid of libressl2.6-* til those are rebuilt
2018-04-05 14:30:51 <ncopa> i have a script that helps me
2018-04-05 14:31:05 <ncopa> but there are many things that needs to be patches
2018-04-05 14:31:07 <ncopa> patched
2018-04-05 14:31:34 <_ikke_> I think that PEP wants to prevent people starting to rely on python == python3 like it's now with python2
2018-04-05 14:31:56 <ncopa> $ git diff --name-only origin/master | grep patch | wc -l
2018-04-05 14:31:56 <ncopa> 46
2018-04-05 14:31:58 <bernhardgruen> for testing I wouldn't do all this myself. For some packages the maintainer could also provide a solution, don't they?
2018-04-05 14:32:13 <bernhardgruen> I think it is faster if more people can work in parallel on things.
2018-04-05 14:32:14 <_ikke_> 46 patches
2018-04-05 14:32:32 <ncopa> those 46 patches have been manual work for me the last days
2018-04-05 14:32:38 <ncopa> some patches are found upstream
2018-04-05 14:32:42 <ncopa> some patches are removed
2018-04-05 14:33:09 <bernhardgruen> what an awful work, I have to say.
2018-04-05 14:33:12 <ncopa> some patches are for fix building with boost 1.66
2018-04-05 14:33:23 <ncopa> its work indeed
2018-04-05 14:34:11 <ncopa> the only package that i have decided to skip for now is cargo
2018-04-05 14:34:12 <bernhardgruen> ncopa, which package did not build with boost 1.66? I had tested them and also provided a fix for namecoin which did not build.
2018-04-05 14:35:12 <ncopa> http://tpaste.us/xe6l
2018-04-05 14:36:13 <ncopa> domoticz build error looked like it was due to libressl
2018-04-05 14:36:24 <ncopa> +-                              boost::asio::ssl::context ctx(io_service, boost::asio::ssl::context::sslv23);
2018-04-05 14:36:24 <ncopa> ++                              boost::asio::ssl::context ctx(boost::asio::ssl::context::sslv23);
2018-04-05 14:38:10 <bernhardgruen> domoticz was not even in my list of test candidates. Where is the dependency to boost?
2018-04-05 14:38:40 <ncopa> makedepends="cmake libressl-dev zlib-dev curl-dev boost-dev sqlite-dev
2018-04-05 14:39:23 <bernhardgruen> sorry for that, ncopa. I did only check via pkgs.a.o
2018-04-05 14:39:27 <ncopa> no hurt feelings about it though
2018-04-05 14:39:56 <ncopa> this can happen, and will happen in future
2018-04-05 14:40:36 <ncopa> and we fix it when we discover it
2018-04-05 14:49:21 <bernhardgruen> The next big step will be the test of all sourceforge.net links I assume. ncopa, there surely are many package where you already have fixed the download links, right?
2018-04-05 15:00:20 <ncopa> i havent really looked into that
2018-04-05 15:01:09 <ncopa> you can do: abuild sourcecheck
2018-04-05 15:01:15 <bernhardgruen> So all downloads worked for you, fine.
2018-04-05 15:01:31 <ncopa> i dont know, i may have the download in cache too
2018-04-05 15:01:31 <bernhardgruen> thanks - will do this now.
2018-04-05 16:08:19 <bernhardgruen> seems, all tested (311) sourceforge.net links are working currently.
2018-04-05 16:44:02 <hs3dUBwdmCjy> ncopa libressl 2.7.0 had a bug: https://mail.python.org/pipermail/python-dev/2018-April/152624.html
2018-04-05 16:44:28 <hs3dUBwdmCjy> maybe i'm delivering old news. if so, apologies
2018-04-05 18:27:49 <azarus> In the zfs-vanilla APKBUILD I see "$_kpkg-dev=$_kpgver", what does that mean?
2018-04-05 18:27:53 <azarus> https://git.alpinelinux.org/cgit/aports/tree/main/zfs-vanilla/APKBUILD#n27
2018-04-05 18:43:16 <terra> azarus: strictly depends on a specific version
2018-04-05 19:02:34 <azarus> terra: Thanks.
2018-04-05 19:19:10 <azarus> Now, I'm encountering this error: https://bpaste.net/show/c3ab6b37ff2c
2018-04-05 19:19:27 <azarus> Is there a way to find out where it tries to cd and why that directory doesn't exist yet?
2018-04-05 19:25:27 <azarus> as a reference, here the APKBUILD: https://ptpb.pw/bWbG
2018-04-05 19:25:40 <terra> rm -rf pkg
2018-04-05 19:25:46 <terra> abuild rootpkg
2018-04-05 19:26:09 <azarus> terra: yup, same output
2018-04-05 19:27:21 <terra> probably you removed -dev package from subpackage list or -dev function removed
2018-04-05 19:27:49 <azarus> the _dev function exists, and it's in the subpackage list
2018-04-05 19:31:17 <terra> let me see your APKBUILD
2018-04-05 19:31:29 <azarus> (scroll up a bit)
2018-04-05 19:45:02 <terra> I only see zfs-vanilla build script your error related with custom kernel
2018-04-05 19:45:18 <terra> *but your error...
2018-04-05 19:46:24 <azarus> Hmm. You're getting the same error as me, right?
2018-04-05 19:49:31 <terra> I'm not getting an error neither build a kernel. trying to figure out what you are trying to compile. kernel itself or zfs package.
2018-04-05 19:49:44 <azarus> right now? just the kernel
2018-04-05 19:49:54 <azarus> zfs is coming later
2018-04-05 19:50:03 <ncopa> hs3dUBwdmCjy: old news yes and it was fixed in libressl-2.7.1 (which still was unstable ABI). the stable ABI release was 2.7.2 which is what i have been working on
2018-04-05 19:50:15 <azarus> I can provide a functioning APKBUILD, just the -dev stuff isn't working
2018-04-05 19:52:57 <hs3dUBwdmCjy> oh, good. sorry for bothering then.
2018-04-05 19:53:34 <ncopa> hs3dUBwdmCjy: no problem. good to check up that kind of things
2018-04-05 19:53:50 <azarus> terra: fully functioning APKBUILD: https://ptpb.pw/_lnS
2018-04-05 20:07:11 <terra> azarus: Unable to connect
2018-04-05 20:21:25 <azarus> terra: Connect with what...? :/
2018-04-05 20:33:47 <terra> Ok, it works now. It seems you already comment out -dev subpackage related sections.
2018-04-05 20:35:21 <azarus> terra: yup, because it doesn't work yet
2018-04-05 20:35:31 <azarus> it's what's giving me trouble
2018-04-05 21:02:18 <terra> Guys, can I set patch level in APKBUILD instead of default 1
2018-04-05 22:30:16 <clandmeter> terra, yes you can
2018-04-05 22:52:28 <terra> clandmeter: how ? like some.patch:0
2018-04-05 22:52:40 <terra> ?
2018-04-05 22:55:43 <clandmeter> https://github.com/alpinelinux/abuild/commit/d8dabed4aa0df31a18b9e682e1efe12e1d00b6d1
2018-04-05 22:59:53 <terra> so, but I dont want to define a single patch level globally. I seek something like bla.patch:0 blabla.patch:2
2018-04-06 00:07:25 <Truxx> There is an issue with the ModemManager package (mentioned it here already)
2018-04-06 00:08:33 <Truxx> It was also in an older log, something with "Could not acquire the 'org.freedesktop.ModemManager1' service name"
2018-04-06 00:12:54 <Truxx> It looks like after installation there is a file missing: org.freedesktop.ModemManager1.service in /usr/share/dbus-1/system-services
2018-04-06 00:13:14 <Truxx> Otherwise ModemManger won't start
2018-04-06 00:13:57 <Truxx> Can someone confirm that this file is not there after install? At least it wasn't for me.
2018-04-06 01:17:23 <ollieparanoid[m]> Truxx: looks like this file is not in any package: https://pkgs.alpinelinux.org/contents?file=system-services&path=&name=&branch=edge
2018-04-06 01:25:47 <Truxx> @ollieparanoid Well, I found this: https://bugs.freedesktop.org/show_bug.cgi?id=34526#c1
2018-04-06 01:26:23 <Truxx> "dbus would not only monitor /usr/share/dbus/system-services, but also /etc/systemd/system and /lib/systemd/system/ and implicitly add all service files found in there that start with "dbus-" and end in ".service" as activatable service to the list. That way we don't need the indirection via files in /usr/share/dbus/system-services anymore"
2018-04-06 01:27:15 <Truxx> Maybe because of systemd it is not required otherwise, but with openrc ModemManager won't run without it
2018-04-06 01:29:39 <Truxx> There are tons of posts with that error message (and there can be many other fixes for this isue too)
2018-04-06 01:29:47 <Truxx> *issue*
2018-04-06 01:47:27 <ollieparanoid[m]> I guess modem manager worked before most other distros adapted to systemd. And there's also Gentoo, where one can choose between systemd and OpenRC. The best way would probably be to identify the missing files (either if you know all of them by now, or by looking at Gentoo's ebuild for example) and after it works for you locally, creating a patch for Alpine's aports so everyone benefits from it
2018-04-06 01:49:30 <Truxx> I'm waiting for someone to confirm, but I think there is only that file missing I mentioned.
2018-04-06 01:50:22 <Truxx> It's more about dbus and the init system, not really a modemmanager issue. modemmanager is great
2018-04-06 01:51:12 <ollieparanoid[m]> what is there left to confirm though? the package contents search linked above shows that the file is not present in any package
2018-04-06 01:51:15 <Truxx> And there is the package maintainer, who will read this, so hopefully it'll be an easy fix
2018-04-06 01:51:56 <Truxx> To confirm that it won't start without it - at least it didn't for me
2018-04-06 01:52:13 <ollieparanoid[m]> okay, now I get what you mean
2018-04-06 01:53:28 <ollieparanoid[m]> I'd recommend to tag the package maintainer or send him an e-mail to make sure he's informed though. I don't think every package maintainer reads all of the #alpine-devel backlog
2018-04-06 01:53:57 <Truxx> I mentioned this here before, I was told it's a dbus issue. On the dbus channel they told me that it's got something to do with openrc. On the openrc channel they told me it's a modemmanger problem.
2018-04-06 01:54:19 <awilfox> don't always assume package maintainers are 'he's either
2018-04-06 01:54:19 <Truxx> Ok, I'll let him know
2018-04-06 01:54:29 <awilfox> though probably in alpine
2018-04-06 01:54:44 <awilfox> I think I'm the only not-he here, and I don't do modemmanager
2018-04-06 01:54:48 <ollieparanoid[m]> awilfox: I did not, I looked it up ;)
2018-04-06 01:54:53 <awilfox> ahh ok
2018-04-06 01:55:00 <Truxx> I read his name already, so I knew...
2018-04-06 01:55:07 <awilfox> it's nice to be able to read this channel again
2018-04-06 01:55:14 <awilfox> matrix has been broken for weeks... I gave up and went back to IRC
2018-04-06 01:56:51 <Truxx> Oh, sorry to hear. Matrix is promising, I have high hopes for the project...
2018-04-06 01:57:43 <awilfox> https://bugzilla.redhat.com/show_bug.cgi?id=1109559  hmm, did you try '/usr/sbin/ModemManager --debug' from root prompt?
2018-04-06 01:57:56 <awilfox> maybe paste that on a pastebin like bpaste.net and link here, maybe it will help
2018-04-06 01:58:00 <Truxx> Yes, did try that - it won't start
2018-04-06 01:58:10 <awilfox> or is it just the same error?
2018-04-06 01:59:47 <Truxx> I've seen that thread before, it's just the same error, but in that case it was something about a
2018-04-06 02:00:02 <Truxx> hardware failure - but there are many other cases with this message
2018-04-06 02:03:50 <awilfox> Truxx: very silly question - dbus is running, right?
2018-04-06 02:04:11 <awilfox> Truxx: i.e. /etc/init.d/dbus start returns "* WARNING: dbus has already been started"
2018-04-06 02:04:13 <Truxx> Yes, it's running
2018-04-06 02:04:35 <Truxx> That was the first thing I checked
2018-04-06 02:08:12 <Truxx> CU @ll
2018-04-06 05:19:37 <ncopa> here comes libressl 2.7. hold on to your hats...
2018-04-06 05:19:45 <awilfox> oh noes
2018-04-06 05:20:59 <_ikke_> 
ACTION grabs a rope

2018-04-06 05:24:43 <fabled> ncopa, so early in the morning!?
2018-04-06 05:26:08 <ncopa> morning
2018-04-06 05:26:30 <ncopa> yes, i was hoping everybody was sleeping and nobody noticed...
2018-04-06 05:26:50 <ncopa> i have a few packages left in testing
2018-04-06 05:27:00 <bernhardgruen[m]> Morning. Good luck.
2018-04-06 05:27:36 <bernhardgruen[m]> fabled: any plans for Java 9 or 10 already?
2018-04-06 05:29:11 <bernhardgruen[m]> Or do you wait for 11 (which is LTS)? I ask because I thought I could start packaging them as we need them in my company.
2018-04-06 05:29:44 <fabled> bernhardgruen[m], i don't think anyone is working on it. PRs would be appreciated.
2018-04-06 05:30:22 <fabled> ncopa, what you mean no one notice when algitbot is going to be printing commits until dawn ;)
2018-04-06 05:31:44 <bernhardgruen[m]> Ok. My idea for 9 and 10 is to download the binary version and use it to bootstrap itself. That way no gcj is needed anymore, fabled.
2018-04-06 05:32:05 <fabled> we don't generally like that :/
2018-04-06 05:32:24 <fabled> can you build 9 with 8 ?
2018-04-06 05:32:41 <bernhardgruen[m]> I know but gcj is not an option in future. I think this will work.
2018-04-06 05:33:53 <bernhardgruen[m]> fabled, do you know that the official openjdk docker images are using the Alpine builds of openjdk directly? Therefore there is no Alpine version for 9 and 10 yet.
2018-04-06 05:35:44 <fabled> there is gazillion docker images we don't know about or maintain. i'm sure they do lots of things we would not do in alpine.
2018-04-06 05:36:06 <fabled> but yes, gcj is tricky
2018-04-06 05:36:22 <bernhardgruen[m]> If ok I will first try to bootstrap it with itself and then try the variant with jdk 8. I fear that I can not start with it before sunday.
2018-04-06 05:36:35 <fabled> bootstrap with jdk8 would be ok for us
2018-04-06 05:36:38 <fabled> currently
2018-04-06 05:36:55 <fabled> the bootstrap is gcj -> openjdk7 -> openjdk8
2018-04-06 05:37:11 <bernhardgruen[m]> But that will not work if GCC 7 comes into play, right?
2018-04-06 05:37:28 <fabled> we are working to keep gcj 6 and upgrade gcc to 7
2018-04-06 05:38:13 <fabled> so at least for the time being we plan to keep gcj
2018-04-06 05:38:19 <fabled> not sure if that changes in future
2018-04-06 05:38:41 <fabled> i wonder if there's any other java compilers that don't depend on java and could be used to bootstrap
2018-04-06 05:38:51 <awilfox> not cross-platform
2018-04-06 05:39:03 <awilfox> IBM's Java might work, but it is for powerpc only IIRC
2018-04-06 05:39:58 <bernhardgruen[m]> Fine.
2018-04-06 05:41:01 <bernhardgruen[m]> Cross platform could be problematic. I don't know for which of them exist official builds.
2018-04-06 05:44:10 <fabled> there's Kaffe but it's old
2018-04-06 05:46:16 <bernhardgruen[m]> btw, do you currently double bootstrap? I.e. build jdk8 with jdk7 and then build jdk8 with jdk8 again to package that version?
2018-04-06 05:48:00 <fabled> we are not doing it explicitly, but iirc icedtea might be doing it
2018-04-06 05:49:29 <fabled> this is annoying that increasing amount of languages want to be self-hosting and don't do proper cross builds
2018-04-06 05:49:52 <bernhardgruen[m]> Afaik icedtea doesn't exist for 9 and 10
2018-04-06 05:50:17 <fabled> yes, we had 8 on the vanilla sdk too earlier, and later converted to icedtea
2018-04-06 05:50:50 <fabled> that was in 2016, see commit f1bcc4bce42045795274a1e87ff38fbee0808a6f
2018-04-06 06:01:10 <bernhardgruen[m]> Thank you fabled.
2018-04-06 09:48:47 <azarus> OK, I am a bit farther with my APKBUILD now. It now still fails in the _dev() function, now with this error: http://dpaste.com/07PSE2Z
2018-04-06 09:49:45 <azarus> Now the changed APKBUILD: https://ptpb.pw/JfE8
2018-04-06 09:50:45 <azarus> It fails on the `make silentoldconfig prepare modules_prepare scripts` part, and it complains that the linux source directory isn't clean.
2018-04-06 09:52:02 <azarus> It suggests running "make mrproper", which I tried, but then it complains about the missing ".config". However if I cp that back into the source directory, it complains about being unclean again.
2018-04-06 09:52:07 <azarus> So confusing :/
2018-04-06 09:57:58 <mps> azarus: "make mrproper" cleans config in kernel source tree
2018-04-06 09:58:21 <azarus> mps: Yup.
2018-04-06 09:59:16 <mps> it's purprose is to make source tree to be 'proper'
2018-04-06 09:59:47 <azarus> I understand.
2018-04-06 10:00:19 <mps> if you need to preserve .config you have to move it somewhere before and move back after make mrproper
2018-04-06 10:01:03 <mps> or, just copy desired .config from somewhere where you have appropriate one
2018-04-06 10:01:09 <azarus> I don't need to preserve the config.
2018-04-06 10:02:29 <mps> but you have to have some custom config and not the kernel default, I presume
2018-04-06 10:02:55 <azarus> That's true. I have the config in another place tough, ready to be copied back into the tree if needed
2018-04-06 10:03:48 <mps> oh, then I misunderstood you
2018-04-06 10:04:14 <azarus> Might not even be, I'm completely confused at the _dev() function of linux-vanilla myself.
2018-04-06 10:04:55 <mps> just to note, 'make mrproper' is something like 'make clean' but on steroids
2018-04-06 10:05:25 <azarus> It basically resets the tree to the original state, like it's distributed?
2018-04-06 10:06:10 <azarus> Anyway. Back to making this work.
2018-04-06 10:06:35 <mps> right, but I'm not sure if it clean *.o and *.ko files, long time passed when I was run it
2018-04-06 10:06:49 <azarus> Which target needs the tree to be pristine?
2018-04-06 10:07:42 <mps> make mroprer ; make clean is safe
2018-04-06 10:08:24 <azarus> And after that, copy the config into the tree again?
2018-04-06 10:08:31 <mps> but 'git reset --hard; git clean -f -d' is more appropriate these days
2018-04-06 10:09:08 <azarus> Now I'm double confused. I don't use git to fetch the kernel sources...?
2018-04-06 10:09:29 <mps> right, you have to copy .config  version which you want to build to the tree
2018-04-06 10:10:10 <azarus> Actually, I don't think we're understanding each other right. I have absolutely no trouble building the kernel.
2018-04-06 10:10:23 <azarus> What's failing is providing the kernel -dev files.
2018-04-06 10:10:25 <mps> if you don't use git, then you forget about what I told about git clean, reset
2018-04-06 10:10:48 <azarus> The kernel gets built just fine. I'll show you the output of "abuild":
2018-04-06 10:11:20 <mps> not needed, I'm not versed (yet) to ABUILD
2018-04-06 10:11:37 <mps> s/ABUILD/APKBUILD/
2018-04-06 10:12:10 <mps> though trying to learn in spare time
2018-04-06 10:12:22 <azarus> Is there a way to make abuild just execute the _dev function, so I can debug this quicker?
2018-04-06 10:12:30 <azarus> Rebuilding the entire kernel each time is slow
2018-04-06 10:12:49 <mps> yes, I tried once :)
2018-04-06 10:13:05 <azarus> Oh, how?
2018-04-06 10:13:43 <mps> can't remember exactly, but just 'abuild build' IIRC
2018-04-06 10:14:03 <mps> actually changing .config before that
2018-04-06 10:15:15 <azarus> The kernel config?
2018-04-06 10:15:24 <mps> but moved to make kernel and modules 'kernel way', i.e. unpack, 'make menuconfig', 'make bzImage modules dtbs'
2018-04-06 10:15:25 <azarus> But then the checksums don't match anymore.
2018-04-06 10:15:59 <mps> abuild checksum, I think
2018-04-06 10:16:09 <azarus> But then it builds the entire kernel again.
2018-04-06 10:16:41 <mps> can't remember, but probably you are right
2018-04-06 10:17:06 <azarus> I'm so confused at this :(
2018-04-06 10:18:01 <azarus> I don't see what's different from linux-vanilla to my APKBUILD, my dev() function fails, and the one from linux-vanilla doesn't.
2018-04-06 10:19:02 <mps> really can't help because I don't understand APKBUILD file for so complicated package like kernels
2018-04-06 10:19:24 <azarus> OK. Thanks anyway.
2018-04-06 10:20:51 <mps> I build kernels from upstream source and install it 'by hand' when I need kernels for specific purposes
2018-04-06 10:21:10 <azarus> That works fine on Alpine and mkinitfs?
2018-04-06 10:21:19 <mps> of course
2018-04-06 10:21:34 <azarus> Oh, cool. Might do it that way then.
2018-04-06 10:21:59 <mps> and for ARM machines that is only way to get it work properly
2018-04-06 10:22:07 <azarus> But then I'd also have to compile ZFS and wireguard manually :s
2018-04-06 10:22:42 <mps> yes, I add wireguard to all my kernels
2018-04-06 10:22:53 <azarus> Directly in tree?
2018-04-06 10:24:10 <mps> there is script in wireguard tree (contrib, I think)by which I make patch and then patch kernel tree with that patch which is generated
2018-04-06 10:24:38 <azarus> Cool!
2018-04-06 10:25:37 <mps> I'm not interested in ZFS so didn't tried to make it
2018-04-06 10:26:09 <azarus> Fair enough. I do use it tough, and ending up with a kernel without a support would make it unbootable. :P
2018-04-06 10:26:52 <mps> understand, same is with me, but instead of zfs I use f2fs
2018-04-06 10:27:09 <azarus> At least it's upstream ;)
2018-04-06 10:28:10 <azarus> I really only want to build my own kernel because I'd like to have kernel 4.16, as it has support for a wireless card I want
2018-04-06 10:28:14 <mps> true, but on my chromebooks it must be compiled in kernel (not a externale module) to boot rootfs from f2fs partition
2018-04-06 10:29:00 <mps> and most distro (haha) kernels have it as a module
2018-04-06 10:29:16 <azarus> ... there's always gentoo xP
2018-04-06 10:30:03 <mps> please don't
2018-04-06 10:30:13 <azarus> I once tried F2FS. Had quite bad fsck times, and only few features I'd like out of a FS. :/
2018-04-06 10:30:44 <mps> sorry, fast finger, "please don't speak word gento to me"
2018-04-06 10:30:59 <azarus> Hehe. Gentoo is OK but kinda bloated.
2018-04-06 10:31:59 <mps> I found that the f2fs is best for emmc, sdcard and microsd card, and maybe usb flash disk
2018-04-06 10:32:29 <mps> quite good speed and solid safety
2018-04-06 10:32:44 <azarus> Is it better than ext4 in your usecase?
2018-04-06 10:33:03 <mps> always is better than ext4
2018-04-06 10:33:35 <mps> sorry, I didn't wrote right word (I'm self taught in english)
2018-04-06 10:34:02 <azarus> Don't worry about it. It's not my primary language either.
2018-04-06 10:34:37 <mps> intended to say that is a better solution than ext4 on sdcard/microsd/emmc because of the speed
2018-04-06 10:35:01 <mps> on one machine it is two times faster then ext4
2018-04-06 10:35:29 <azarus> I have trouble believing that, what kind of load are you putting on it?
2018-04-06 10:36:11 <mps> on my son notebook with emmc it is astonishing fast, I simply can't beliewe how fast is it
2018-04-06 10:36:50 <mps> but never was interested in running benchmark to show concrete numbers
2018-04-06 10:37:04 <azarus> Might also just be placebo :P
2018-04-06 10:37:30 <mps> hehe, I'm to old for such kind of placebos
2018-04-06 10:38:40 <mps> azarus: btw, looking your surname on the bugs.a.o looks like you are from country
2018-04-06 10:42:25 <azarus> ;)
2018-04-06 11:01:06 <azarus> My heritage might be, but I've grown up in Switzerland.
2018-04-06 11:14:07 <mps> aha, ok, so we can't talk on my native language
2018-04-06 11:14:33 <azarus> Njet! :P
2018-04-06 11:14:57 <mps> not njet but ne
2018-04-06 11:15:00 <mps> :)
2018-04-06 11:15:26 <azarus> Hvala!
2018-04-06 11:16:03 <mps> danke, merci, gracie
2018-04-06 11:16:29 <mps> s/gracie/grazie/ IIRC italian
2018-04-06 14:19:45 <chambbj> https://github.com/alpinelinux/aports/pull/3901 and https://github.com/alpinelinux/aports/pull/3933 both ready for review. Thanks!
2018-04-06 14:20:32 <vkrishn> ncopa: https://git.alpinelinux.org/cgit/aports/tree/main/wget/APKBUILD#n34 says --with-ssl=openssl, hope its ok
2018-04-06 14:28:31 <vkrishn> this should not an issue in al, https://www.theregister.co.uk/2018/02/12/kde_naming_usb_drive_vuln/ ;) ?
2018-04-06 14:28:39 <vkrishn> not be*
2018-04-06 15:32:13 <ncopa> vkrishn: doesnt look like we have plasma
2018-04-06 15:32:17 <ncopa> wget should be ok
2018-04-06 16:53:56 <TBB> might be an issue for whatsthatspinagain, Adelie?
2018-04-07 06:45:49 <Aerdan[m]> no, adelie is on 5.12
2018-04-08 10:38:00 <liwakura> hello
2018-04-08 10:38:10 <liwakura> yesterday i tried setting up alpine linux on hetzner cloud
2018-04-08 10:38:32 <liwakura> and it seems that the udhcpc script for busybox behaves different from other dhcp implementations
2018-04-08 10:39:00 <liwakura> Constellations to cause is is having the gateway outside of whats covered by the netmask
2018-04-08 10:39:03 <liwakura> https://w1r3.net/zSpeMG.txt
2018-04-08 10:40:01 <liwakura> so i was thinking if it makes sense to run "ip route add $gw dev $interface scope link" before adding the actual default route
2018-04-08 13:52:04 <_ikke_> How would a default gateway outside of the subnet event work?
2018-04-08 13:52:44 <_ikke_> s/event/even
2018-04-08 14:06:10 <liwakura> _ikke_: by explicitly setting a route to that gateway using that interface
2018-04-08 14:08:36 <_ikke_> But that gateway should still be inside the current broadcast domain, right?
2018-04-08 14:29:47 <liwakura> dont know
2018-04-08 14:30:01 <liwakura> i think its outside in that case
2018-04-08 14:30:07 <_ikke_> It can't
2018-04-08 14:30:15 <_ikke_> It needs ARP to get the mac adress of the default gateway
2018-04-08 14:30:50 <liwakura> i didnt know ARP was dependent of any existing IP configuration
2018-04-08 14:31:30 <azarus> ... if busybox udhcpc doesn't work for you, why not use dhcp or dhcpcd?
2018-04-08 14:32:03 <liwakura> azarus: its not udhcpc that doesn't work
2018-04-08 14:32:08 <liwakura> its the glue script that alpine ships
2018-04-08 14:32:24 <azarus> Hmm. Okay then. :/
2018-04-08 14:32:27 <liwakura> _ikke_: im sure the network setup is functional, even if i dont understand how
2018-04-08 14:33:00 <_ikke_> Just trying to think how such a setup even would work
2018-04-08 14:33:19 <_ikke_> The only thing I can think is that the subnet is smaller than it actually is
2018-04-08 14:33:45 <_ikke_> or perhaps arp proxy
2018-04-08 14:33:56 <liwakura> you want to see it via ssh?
2018-04-08 14:34:56 <liwakura> also, somehow debian is able to do this without issues
2018-04-08 14:35:16 <_ikke_> I believe you that it works
2018-04-08 14:35:27 <liwakura> i'll tcpdump and look what ARP does
2018-04-08 14:43:17 <liwakura> the impact is currently, alpine linux is the first ISO for hetzner cloud
2018-04-08 14:43:33 <liwakura> one puts that ISO in, reboots the machine, and dhcp fails
2018-04-08 14:44:08 <liwakura> so i wanted to open for discussion if it makes sense to adapt the udhcpc script
2018-04-08 14:50:00 <_ikke_> What part of dhcp is faling?
2018-04-08 14:52:03 <liwakura> adding the default route
2018-04-08 14:52:41 <liwakura> in the /usr/share/udhcpc/default.script
2018-04-08 14:52:59 <liwakura> fails with "Address unreachable", because it does not know how to reach the default gateway
2018-04-08 14:55:11 <_ikke_> Right, I think it assumed the default gateway is reachable
2018-04-08 14:55:48 <liwakura> fix is sth like "ip route add $gw dev $interface scope link"
2018-04-08 14:55:56 <liwakura> like dhcpcd does
2018-04-08 15:09:18 <bernhardgruen[m]> Do you know if this additional rule has any negative impacts?
2018-04-08 15:10:28 <liwakura> if the gateway ip is also some address on some other network via some other interface, that would break
2018-04-08 15:10:52 <liwakura> which is probably an even dumber idea to have in the first place
2018-04-08 15:25:16 <bernhardgruen[m]> So possible solution would be to first try the standard way and if this does not work switch to the alternative. Shouldn't be too complicated
2018-04-08 17:14:37 <mksully222> I think opensc build fail is because libressl-2.7.patch wasnt integrated into APKBUILD, PR https://github.com/alpinelinux/aports/pull/3966 should fix
2018-04-08 18:16:12 <pickfire> Can we add pre-commit-msg git hook like https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Commit_your_work to generate default commit message with url and pkgdesc?
2018-04-08 18:16:39 <pickfire> Currently there are multiple kinds of "new aport".
2018-04-08 19:18:20 <Fusl> any alpine dev online willing to take a look at https://github.com/alpinelinux/aports/pull/3377 and https://github.com/alpinelinux/aports/pull/3378 or provide information why it's being ignored?
2018-04-08 19:26:16 <terra1> can't install dnscrypt-proxy due to the missing libldns.so.1 but the package exist on repo. https://paste.ee/p/fvFdy
2018-04-08 19:35:45 <_ikke_> terra: Justs a lot of open PRs, and Alpine is currently focusing on getting a stable release out
2018-04-08 19:40:08 <terra> _ikke_: that wasn't me asking. it was Fusl
2018-04-08 19:41:16 <Fusl> thanks _ikke_
2018-04-08 19:50:53 <terra> it seems dnscrypt-proxy 1.9.x is no more maintained. not working website, can't fetch source to compile.
2018-04-08 19:52:09 <terra> some people continue it from 2.0 https://github.com/jedisct1/dnscrypt-proxy
2018-04-09 06:26:42 <fabled> i deleted main/alpine because upstream deleted all tarballs
2018-04-09 06:26:52 <fabled> apparently there's forks, but not sure which one to use
2018-04-09 07:50:40 <ncopa> fabled: looks like alpine (MUA) dont do release tarballs anymore. the git repo seems active though.
2018-04-09 07:55:31 <fabled> yeah
2018-04-09 07:55:51 <fabled> ncopa, see https://github.com/alpinelinux/aports/pull/3947
2018-04-09 08:04:23 <ncopa> i suppose we could create tarball from git tag and upload to dev.a.o
2018-04-09 08:04:26 <fabled> are x86 and armhf builders somehow broke?
2018-04-09 08:04:35 <ncopa> but i am sort of ok with deleting it for now
2018-04-09 08:04:50 <ncopa> i can check the builders...
2018-04-09 08:05:36 <ncopa> it looks like build-edge-armhf is running openvswitch test suite
2018-04-09 08:05:41 <fabled> ok
2018-04-09 08:05:43 <fabled> it's just slow
2018-04-09 08:07:27 <ncopa> hum
2018-04-09 08:08:18 <fabled> ok, armhf just started uploading main
2018-04-09 08:08:43 <ncopa> i did a killall ovsdb-server
2018-04-09 08:09:04 <ncopa> and killall epmd
2018-04-09 08:09:50 <ncopa> the package was built but hanged for some reason
2018-04-09 08:11:09 <ncopa> seems like same thing happened with x86
2018-04-09 08:11:32 <_ikke_> Yeah, I fixed x86/x86_64 the other day
2018-04-09 08:12:14 <ncopa> Executing busybox-1.28.2-r3.trigger
2018-04-09 08:12:14 <ncopa> OK: 281 MiB in 110 packages
2018-04-09 08:12:14 <ncopa> >>> openvswitch: Updating the main/x86 repository index...
2018-04-09 08:12:14 <ncopa> >>> openvswitch: Signing the index...
2018-04-09 08:12:21 <ncopa> according the log the package is built
2018-04-09 08:12:44 <ncopa> but apparently it hangs due to testsuite not cleaning up properly
2018-04-09 08:13:23 <ncopa> i killed ovsdb-server now on build-edge-x86
2018-04-09 08:28:35 <ncopa> i alos had to kill ovsdb-server on s390x
2018-04-09 09:03:41 <ncopa> kaniini: you assigned yourself https://github.com/alpinelinux/aports/pull/3377 https://github.com/alpinelinux/aports/pull/3378 what is the status there?
2018-04-09 11:29:41 <chambbj> Requesting review of PR at https://github.com/alpinelinux/aports/pull/3901. Thanks!
2018-04-09 11:30:45 <mepholic> the url in that PR is 404 :(
2018-04-09 11:30:53 <mepholic> >  +url="http://people.cs.ubc.ca/~mariusm/flann"
2018-04-09 11:31:08 <mepholic> seems it's moved here
2018-04-09 11:31:11 <mepholic> https://www.cs.ubc.ca/research/flann/
2018-04-09 11:31:46 <mepholic> chambbj: what does the testsuite on that do?
2018-04-09 11:34:51 <chambbj> mepholic: you mean generally speaking? or what was the issue that I chose to skip them for now?
2018-04-09 11:35:47 <mepholic> yeah, what was the specific issue?
2018-04-09 11:36:00 <chambbj> and thanks for pointing out the URL change, i can fix
2018-04-09 11:44:43 <chambbj> The tests require quite a few additional dependencies (python, ruby, etc.), which I know that I can install via checkdepends. Even so, many were failing. There is no overarching make test (well, there is, but it doesn't do what you would expect). Documentation in general is poor (e.g., there is actually a data download step as well). I'd be happy to continue to work through the issues and bump the build number once complete, if
2018-04-09 11:45:32 <mepholic> there weren't any failures in any math functions, were there?
2018-04-09 11:46:21 <chambbj> no, it wasn't a matter of the tests building and failing - it was more a matter of the configuration of the tests themselves prevented it from ever getting to more meaningful tests
2018-04-09 11:47:07 <chambbj> I'd have to setup the environment and try again to give any more information. Don't remember all the details.
2018-04-09 11:58:43 <mepholic> ok
2018-04-09 11:58:46 <mepholic> not really important then
2018-04-09 12:00:23 <chambbj> ok, URL fix is done
2018-04-09 14:13:17 <korylprince> What's the best way to request a package be updated? (The update is in aports, it just hasn't been built yet...)
2018-04-09 14:17:56 <_ikke_> Hmm, if it's in aports, it should've been built
2018-04-09 14:18:06 <_ikke_> It's an automatic process
2018-04-09 14:23:25 <korylprince> I had two patches accepted for community/rt4 on April 2nd, but the package hasn't been rebuilt since March 26th.
2018-04-09 14:23:53 <korylprince> I was supposed to update pkgrel in APKBUILD, wasn't I? I'm new to this...
2018-04-09 14:28:16 <korylprince> Urg, that's what happened. My first patch I submitted, someone corrected the pkgrel for me it looks like. But that didn't happen when I submitted the next two patches. I'll submit another patch to fix the pkgrel...
2018-04-09 14:33:48 <clandmeter> korylprince, i bumped it.
2018-04-09 14:37:12 <korylprince> Thanks! I'll make sure to do that next time. I really like alpine's aports/build system. Still learning all the ins and outs
2018-04-09 16:04:39 <wener> please review this pr https://github.com/alpinelinux/aports/pull/3976 add mirrors for China
2018-04-09 18:14:43 <duncan^> Hmm
2018-04-10 04:39:21 <_ikke_> We might want to pull zfs-0.7.7 (https://news.ycombinator.com/item?id=16797932)
2018-04-10 04:42:16 <_ikke_> https://bugs.alpinelinux.org/issues/8779
2018-04-10 07:48:26 <ncopa> _ikke_: looks like we are using 0.7.5 atm
2018-04-10 08:22:16 <clandmeter> ncopa, https://pkgs.alpinelinux.org/package/edge/main/x86_64/zfs
2018-04-10 08:23:32 <ncopa> that is userspace tool
2018-04-10 08:23:34 <ncopa> not kernel module
2018-04-10 08:24:04 <ncopa> https://git.alpinelinux.org/cgit/aports/tree/main/zfs-vanilla/APKBUILD#n6
2018-04-10 08:29:51 <clandmeter> ah ok, bit unclear like this :)
2018-04-10 08:30:10 <azarus> hrm, has me quite worried
2018-04-10 16:09:47 <ncopa> jirutka: i have some WIP for cargo in case you are interested: http://tpaste.us/jD8e
2018-04-10 16:10:03 <ncopa> i didnt know how to deal with it while upgrading libressl so i disabled it for now
2018-04-10 16:10:17 <ncopa> we should fix it
2018-04-10 16:13:09 <jirutka> ncopa: does this actually work…?
2018-04-10 16:14:00 <ncopa> no
2018-04-10 16:14:03 <ncopa> it fails to build
2018-04-10 16:14:30 <ncopa> i saw that rust-openssl upstream added support for libressl-2.7 not too long ago
2018-04-10 16:14:34 <jirutka> ncopa: I need to know this: how is it currently done that rust can use previously built rust on builders?
2018-04-10 16:15:36 <jirutka> ncopa: b/c when I add “rust” to rust’s makedepends, it’s ignored, even when I specify rust<=1.8.0 or something like that
2018-04-10 16:15:51 <jirutka> ncopa: I know why it is ignored, but don’t know how it actually work now
2018-04-10 16:17:10 <ncopa> hm
2018-04-10 16:17:22 <ncopa> i think we have depends=gcc in gcc
2018-04-10 16:17:45 <jirutka> yes
2018-04-10 16:18:06 <jirutka> but when I try to build this locally with abuild -r, it does not install rust
2018-04-10 16:18:12 <jirutka> (or gcc or whatever)
2018-04-10 16:18:20 <ncopa> is rust already installed?
2018-04-10 16:18:36 <jirutka> how can I know? I don’t have access to builders
2018-04-10 16:18:57 <jirutka> (that’s why I’m asking, it’s really very annoying that it’s more or less like a blackbox for me)
2018-04-10 16:18:58 <ncopa> i thought you tried build locally?
2018-04-10 16:19:05 <jirutka> yes
2018-04-10 16:19:18 <jirutka> does this mean that rust is always installed on builders?
2018-04-10 16:19:22 <ncopa> no
2018-04-10 16:19:36 <ncopa> i would expect the builders to install rust when they try build rust
2018-04-10 16:19:46 <jirutka> when I’ve tried abuild -r, ofc rust was not installed before
2018-04-10 16:19:57 <ncopa> it is just ignored?
2018-04-10 16:20:01 <jirutka> yes, but I can’t reproduce such behaviour locally
2018-04-10 16:20:05 <ncopa> same here
2018-04-10 16:20:09 <ncopa> ERROR: unsatisfiable constraints:
2018-04-10 16:20:09 <ncopa>   so:libcrypto.so.42 (missing):
2018-04-10 16:20:09 <ncopa>     required by: cargo-0.22.0-r4[so:libcrypto.so.42] cargo-0.22.0-r4[so:libcrypto.so.42]
2018-04-10 16:20:09 <ncopa>   so:libssl.so.44 (missing):
2018-04-10 16:20:09 <ncopa>     required by: cargo-0.22.0-r4[so:libssl.so.44] cargo-0.22.0-r4[so:libssl.so.44]
2018-04-10 16:20:09 <ncopa>   .makedepends-rust-0:
2018-04-10 16:20:09 <ncopa>     masked in: cache
2018-04-10 16:20:10 <ncopa>     satisfies: world[.makedepends-rust]
2018-04-10 16:20:16 <jirutka> and I even found code in abuild script that removes it from dependencies to be installed
2018-04-10 16:20:27 <jirutka> that’s different problem
2018-04-10 16:20:37 <ncopa> yes, i realize that
2018-04-10 16:21:01 <ncopa> where does it remove it from dependencies?
2018-04-10 16:22:54 <ncopa> i think it removes it from depends (runtime), but it should not remove it from makedepends (bulid time)
2018-04-10 16:24:06 <jirutka> ncopa: https://github.com/alpinelinux/abuild/blob/master/abuild.in#L1981
2018-04-10 16:24:43 <jirutka> and this https://github.com/alpinelinux/abuild/blob/master/abuild.in#L1979
2018-04-10 16:25:03 <jirutka> (btw there’s missing `~` operator)
2018-04-10 16:26:19 <jirutka> anyway, I have to somehow build static binary of newer cargo, but that’s quite complicated ’cause this crap is broken on multiple levels
2018-04-10 16:40:53 <_ikke_> ncopa: So that issue can be closed again?
2018-04-10 16:41:41 <ncopa> _ikke_: i think so yes
2018-04-10 16:42:56 <_ikke_> Done
2018-04-10 16:43:51 <jirutka> ncopa: so how does it work on builders?
2018-04-10 16:44:08 <jirutka> ncopa: and what is that manual step that is required first time?
2018-04-10 16:44:46 <_ikke_> jirutka: fyi, this is the script that is used (not sure if it helps): https://git.alpinelinux.org/cgit/aports/tree/main/aports-build/aports-build
2018-04-10 16:46:12 <ncopa> im looking at it and at the git blame/log
2018-04-10 16:49:16 <jirutka> _ikke_: it’s not; actually more interesting is buildrepo, but it just calls `abuild -r` https://github.com/alpinelinux/lua-aports/blob/master/bin/buildrepo.lua#L100, so does not reveal anything
2018-04-10 16:49:43 <jirutka> there must be some configuration or whatever on the builders, so that damn part that is total black box for anyone who don’t have access to these damn systems
2018-04-10 16:53:56 <ncopa> the code is in main/aports-build
2018-04-10 16:55:34 <jirutka> yes, the code…
2018-04-10 16:56:01 <ncopa> config is there too
2018-04-10 16:56:05 <jirutka> so can you explain me how does it work? because I can only explain why it does not work and can show how it does not work on local system, but I have no idea how it works on the builder
2018-04-10 16:56:14 <ncopa> mqtt-exec.aports-build.confd
2018-04-10 16:57:07 <jirutka> my suspicious is that rust is actually preinstalled on builders…
2018-04-10 16:57:22 <ncopa> i think we remove the $pkgname and subpackages from makedepends because otherwise we get problem during bootstrap an empty repository
2018-04-10 16:57:53 <ncopa> its not
2018-04-10 16:58:38 <jirutka> it starting to be really funny…
2018-04-10 16:58:54 <ncopa> i think we have a problem really
2018-04-10 16:59:02 <ncopa> i dont think it works at all in the rust case
2018-04-10 16:59:10 <jirutka> but it actually works…
2018-04-10 16:59:11 <jirutka> magic?
2018-04-10 16:59:23 <jirutka> I’ve upgraded rust few times and it works on the x86_64 builder
2018-04-10 16:59:58 <jirutka> and yet you’re not able to answer me how does it work and I found only evidence for why it cannot work, no idea how it actually work
2018-04-10 17:00:54 <jirutka> also I know that there’s _some_ manual step that someone had to do to get it work, but I have no idea what this step actually is
2018-04-10 17:01:01 <ncopa> i think it "works" due to the the circular dep
2018-04-10 17:01:02 <jirutka> oh dear, our infra is so transparent!
2018-04-10 17:01:13 <jirutka> hmm
2018-04-10 17:01:22 <jirutka> b/c rust depends on cargo and cargo depends on rust
2018-04-10 17:01:32 <jirutka> yeah, this is probably the explanation!
2018-04-10 17:01:39 <ncopa> and it needs manual bootstrapping
2018-04-10 17:01:57 <jirutka> okay, but how it works for gcc?
2018-04-10 17:02:05 <jirutka> there’s some circular dep too?
2018-04-10 17:02:08 <ncopa> gcc is always installed on the builders
2018-04-10 17:02:13 <jirutka> aha, right
2018-04-10 17:02:47 <jirutka> and what about ghc?
2018-04-10 17:02:51 <_ikke_> We need some kind of bootstrap paradigm, there are more packages with the issue
2018-04-10 17:03:01 <ncopa> i think ghc is broke
2018-04-10 17:03:42 <ncopa> hm
2018-04-10 17:03:48 <ncopa> ghc does not have ghc in makedepends
2018-04-10 17:03:52 <jirutka> https://github.com/alpinelinux/aports/blob/master/community/ghc/APKBUILD#L18
2018-04-10 17:03:53 <jirutka> it has
2018-04-10 17:04:05 <jirutka> ghc is self-hosted, it needs ghc to build itself
2018-04-10 17:04:16 <ncopa> and its broken
2018-04-10 17:04:20 <ncopa> configure: error: GHC is required.
2018-04-10 17:04:20 <ncopa> >>> ERROR: ghc: all failed
2018-04-10 17:04:31 <jirutka> oh great
2018-04-10 17:05:28 <ncopa> so, how this worked originally
2018-04-10 17:05:37 <ncopa> build-base was expeced to be on the builder
2018-04-10 17:05:57 <ncopa> that is why you dont need to add make gcc musl-dev to makedepends
2018-04-10 17:06:43 <ncopa> while buildind buildbase (eg bootstrapping aports), we needed to exclude everythign in the dependency chain for build-base
2018-04-10 17:07:07 <ncopa> you cannot makedepend on yourself in otherworkds
2018-04-10 17:07:26 <_ikke_> yeah, that makes kind of sense if you start from scratch everytime
2018-04-10 17:08:15 <ncopa> what is interesting though
2018-04-10 17:08:21 <ncopa> is that it apparently works with go
2018-04-10 17:08:59 <jirutka> yeah, another example of self-hosted thing… so how does it work?
2018-04-10 17:09:09 <ncopa> and the reason for that is that go does not depend on go, but on go-bootstrap
2018-04-10 17:09:17 <jirutka> aha
2018-04-10 17:09:26 <ncopa> and go has a provides=go-bootstrap
2018-04-10 17:09:29 <jirutka> and go-bootstrap builds go using gcc?
2018-04-10 17:09:36 <ncopa> it used to
2018-04-10 17:09:48 <ncopa> but not anymore
2018-04-10 17:09:49 <jirutka> used to?
2018-04-10 17:09:56 <jirutka> so go-bootstrap is not used at all?
2018-04-10 17:10:33 <ncopa> gcc-go can only build go-1.4
2018-04-10 17:11:02 <ncopa> or...
2018-04-10 17:11:03 <ncopa> huim
2018-04-10 17:12:02 <ncopa> we have a go-bootstrap APKBUILD
2018-04-10 17:12:18 <jirutka> yes, we know that…
2018-04-10 17:12:33 <ncopa> i wonder how go-bootstrap is built
2018-04-10 17:12:51 <_ikke_> 1.4
2018-04-10 17:13:14 <jirutka> using gcc…
2018-04-10 17:13:17 <jirutka> apparently…
2018-04-10 17:13:23 <jirutka> go-bootstrap is 1.4.3
2018-04-10 17:13:35 <_ikke_> so gcc -> go 1.4 -> go <latest>
2018-04-10 17:13:53 <jirutka> can you really build latest go with go 1.4?
2018-04-10 17:15:11 <_ikke_> http://build.alpinelinux.org/buildlogs/build-edge-armhf/community/go/go-1.10.1-r0.log
2018-04-10 17:15:24 <_ikke_> That just downloads the latest version of go
2018-04-10 17:15:31 <_ikke_> (go 1.10)
2018-04-10 17:15:40 <jirutka> yes
2018-04-10 17:15:47 <jirutka> not downloads, but installs from aports
2018-04-10 17:15:54 <ncopa> due to provides=go-bootstrap
2018-04-10 17:16:02 <_ikke_> jirutka: yeah, meant that
2018-04-10 17:16:06 <jirutka> so it’s basically a hack
2018-04-10 17:16:20 <ncopa> ok. it looks like go-1.4 had a small bootstrap go implemented in C
2018-04-10 17:16:26 <bernhardgruen[m]> Isn't this the same sh*t as with openjdk? At some point there must be some (binary) base download to bootstrap with it.
2018-04-10 17:16:35 <ncopa> bernhardgruen[m]: it is
2018-04-10 17:16:42 <ncopa> with openjdk we have gcc-java
2018-04-10 17:16:44 <jirutka> bernhardgruen[m]: openjdk7 is built using gcc-java
2018-04-10 17:16:48 <ncopa> that builds openjdk7
2018-04-10 17:16:55 <ncopa> which builds openjdk8
2018-04-10 17:16:58 <jirutka> yeah
2018-04-10 17:17:00 <_ikke_> bootstrapping languages is always fun
2018-04-10 17:17:14 <ncopa> not when it comes to rust :)
2018-04-10 17:17:36 <ncopa> rust is a completely different animal
2018-04-10 17:17:41 <jirutka> it’s okish when authors of the lang are sane and understand the bootstrapping problem… not the case of Rust
2018-04-10 17:18:10 <_ikke_> with Rust they assume you always have a Rust available, right?
2018-04-10 17:18:13 <ncopa> care about the bootstrapping problem
2018-04-10 17:18:14 <ncopa> yes
2018-04-10 17:18:44 <bernhardgruen[m]> Btw. Thanks for this discussion it gives a deep insight into Alpine.
2018-04-10 17:19:13 <jirutka> great example is Crystal, they provide binary that does not depend on glibc, it’s not static binary (don’t know why, it’s possible to build crystal as stativ binary with musl), but it doesn’t have any dependencies and don’t use any glibc extensions
2018-04-10 17:19:25 <jirutka> and it’s damn easy to cross-compile crystal for other arches or systems
2018-04-10 17:19:40 <jirutka> https://crystal-lang.org/docs/syntax_and_semantics/cross-compilation.html
2018-04-10 17:20:09 <_ikke_> jirutka: but you still have to rely on a binary blob then
2018-04-10 17:20:16 <jirutka> it’s really these two commands, no hidden craps: on build system: `crystal build your_program.cr --cross-compile --target "x86_64-unknown-linux-gnu"`, on host system: `cc your_program.o -o your_program -lpcre -lrt -lm -lgc -lunwind`
2018-04-10 17:20:33 <bernhardgruen[m]> IMHO the most compatible way would be to somehow compile a statical build (or download) that is used to bootstrap a language. These magical packages shouldn't be available in the normal repositories.
2018-04-10 17:20:47 <ncopa> i disagree
2018-04-10 17:21:11 <jirutka> _ikke_: yes… the only other way for self-hosted langs is to have some different implementation in other language, at least minimal compiler
2018-04-10 17:21:18 <ncopa> i think the best thing would be to have a boostrap implementation in C
2018-04-10 17:21:30 <ncopa> like go did
2018-04-10 17:21:41 <jirutka> like go had…
2018-04-10 17:21:46 <_ikke_> right, but that requires you to keep two versions of the compiler up-to-date, right?
2018-04-10 17:21:55 <jirutka> can you actually build latest go with go 1.4…?
2018-04-10 17:22:02 <ncopa> dunno
2018-04-10 17:22:47 <_ikke_> Most languages want to become self-hosted as soon as possible
2018-04-10 17:22:51 <bernhardgruen[m]> You always have some form of binary that is not visible (on the builders it is the system the builders use).
2018-04-10 17:22:59 <jirutka> _ikke_: and that’s very bad thing IMHO
2018-04-10 17:23:28 <jirutka> _ikke_: self-hosted is real pain for portability
2018-04-10 17:23:34 <_ikke_> yes, understand
2018-04-10 17:23:37 <ncopa> jirutka: re rust, i think the easiest way forward is to update cargo as a separate package
2018-04-10 17:23:44 <ncopa> for now
2018-04-10 17:23:56 <jirutka> no
2018-04-10 17:24:22 <ncopa> i think it will require manual intervention at this point
2018-04-10 17:24:38 <ncopa> regardless what we do
2018-04-10 17:24:45 <jirutka> there’s no Cargo.lock anymore in cargo repository, they told me that it’s not expected to be built separately, it should be built with rust; without Cargo.lock you don’t have locked versions
2018-04-10 17:24:56 <ncopa> argh
2018-04-10 17:24:59 <_ikke_> I wonder what we would do if we only had aports left
2018-04-10 17:25:08 <jirutka> you can create it yourself to provide reproducibility between your builds, but this is still not very good
2018-04-10 17:25:46 <jirutka> but okay, even if we decide to do that, we still have a problem
2018-04-10 17:26:00 <jirutka> the last version I was able to build statically is 1.17.0
2018-04-10 17:26:16 <ncopa> of rust?
2018-04-10 17:26:20 <jirutka> no, cargo
2018-04-10 17:26:30 <jirutka> long story short, you can’t build some cargo deps with that version
2018-04-10 17:26:32 <jirutka> anymore
2018-04-10 17:26:59 <jirutka> there should be a way to build static cargo, but I have to do some hackery AGAIN
2018-04-10 17:27:01 <ncopa>  oh /o\
2018-04-10 17:27:18 <jirutka> (I need some alcohol, but I’m at work now)
2018-04-10 17:27:20 <bernhardgruen[m]> Is it possible to "cross-compile" from glibc to musl? Like the compiler runs with glibc and then generates musl objects and libs?
2018-04-10 17:27:29 <jirutka> (it really increases me blood pressure when talking about Cargo)
2018-04-10 17:27:37 <jirutka> heh, cross-compile from glibc to musl
2018-04-10 17:27:48 <jirutka> I was doing that for aarch64
2018-04-10 17:28:09 <jirutka> hm, I spent on it 2 or 3 days (not continuous work, it takes forever to compile)?
2018-04-10 17:28:21 <ncopa> x86_64 machine?
2018-04-10 17:28:25 <ncopa> or aarch64 machine
2018-04-10 17:28:28 <jirutka> no, aarch64
2018-04-10 17:28:30 <jirutka> I _think_ that I finally have cross-compiled rustc
2018-04-10 17:28:33 <jirutka> but not cargo!
2018-04-10 17:28:47 <ncopa> and cargo does not build without cargo?
2018-04-10 17:28:48 <jirutka> and I have to hack it to make it cross-compile for musl
2018-04-10 17:29:05 <jirutka> yes, of course it does not!! this is perfectly normal and sane, what did you though?
2018-04-10 17:29:11 <jirutka> (sarcasm)
2018-04-10 17:29:36 <ncopa> so you need to crosscompile cargo as well?
2018-04-10 17:29:37 <jirutka> really, burn this shit with fire… or napalm
2018-04-10 17:29:40 <jirutka> yes
2018-04-10 17:30:02 <jirutka> but I‘ll try to use minicargo from mrustc
2018-04-10 17:30:21 <jirutka> I don’t have success with mrustc on aarch64 yet, it works only on x86_64 for now
2018-04-10 17:30:29 <jirutka> but minicargo should be usable
2018-04-10 17:31:03 <jirutka> I still know about some ways how to do that, but I hate doing it, it’s so desperate that it’s not even challenging anymore, it’s just sad
2018-04-10 17:31:08 <ncopa> https://github.com/thepowersgang/mrustc
2018-04-10 17:31:13 <jirutka> yes, that mrustc
2018-04-10 17:31:14 <ncopa> looks like salvation to me
2018-04-10 17:31:18 <jirutka> not yet
2018-04-10 17:32:05 <jirutka> I’m watching this project, sent few PRs, tried it on musl etc.
2018-04-10 17:32:06 <bernhardgruen[m]> Are there any usage statistics (per package and arch)? That could make the decision easier, doesn't it?
2018-04-10 17:32:52 <jirutka> but as I said, other than x86_64 is currently not really supported, there are some bugs; I/clandmeter have provided aarch64 machine to the author to test it on aarch64 (he don’t have access to any aarch64 machine himself)
2018-04-10 17:34:40 <jirutka> at least we can use it to bootstrap rust and cargo on x86_64, but… there’s a but… it currently works for rust 1.19.0, the latest rust is 1.25.0… and generally you need (V - 1) to build V
2018-04-10 17:34:53 <jirutka> so it’s not something we really want to do know
2018-04-10 17:35:02 <bernhardgruen[m]> If it isn't supported why one would use it on an unsupported platform? The person would not even get help from upstream.
2018-04-10 17:35:08 <jirutka> but at least there’s some way already, that’s huge successs
2018-04-10 17:35:22 <_ikke_> bernhardgruen[m]: afaik, alpine doesn't do any statistics tracking
2018-04-10 17:35:24 <jirutka> you don’t understand
2018-04-10 17:35:36 <jirutka> it’s not supported in terms that author didn’t test it at all yet
2018-04-10 17:35:46 <jirutka> there’s some basic support like defined targets etc.
2018-04-10 17:35:55 <jirutka> but as I and some other guy already found, it’s buggy
2018-04-10 17:36:07 <jirutka> we were not able to build rustc using it
2018-04-10 17:36:32 <bernhardgruen[m]> Ok, I understand. Thanks
2018-04-10 17:36:56 <jirutka> and thepowersgang is kinda inactive since Easter, so don’t know how about that aarch64 machine I gave him access to yet
2018-04-10 17:37:18 <jirutka> you can read some issues here https://github.com/thepowersgang/mrustc/issues
2018-04-10 17:37:44 <ncopa> jirutka: so you think mrustc and minirutc is the way forward for v3.8?
2018-04-10 17:37:46 <jirutka> and as you can see in https://github.com/thepowersgang/mrustc/issues/2, I’m watching this project for almost 2 years :)
2018-04-10 17:38:23 <jirutka> ncopa: no, it’ll be better to use our latest rustc/cargo built to bootstrap rustc/cargo for v3.8
2018-04-10 17:38:57 <ncopa> what we do with libressl 2.7 and cargo?
2018-04-10 17:39:11 <ncopa> you want ship cargo with rust package, right?
2018-04-10 17:39:17 <jirutka> ncopa: and about cargo, I needed to know how this dependency on itself etc. works to decide how to proceed
2018-04-10 17:39:40 <ncopa> i think we can have a rust-bootstrap package
2018-04-10 17:39:46 <jirutka> ncopa: and also to get reminded about this nasty libressl problem, this means that we need static cargo anyway :(
2018-04-10 17:39:57 <ncopa> or add provides="rust-bootstrap"
2018-04-10 17:40:13 <ncopa> i guess we can build it with static libressl like we did with libgit2
2018-04-10 17:40:35 <jirutka> I thought that I will just resign on static cargo and make rust/cargo package depend on itself, but I cannot do that because of libressl incompatibility
2018-04-10 17:41:01 <jirutka> provides="rust-bootstrap" is just a hack… what about making this exclusion of self optional?
2018-04-10 17:41:37 <ncopa> that means we need refactor abuild to do proper bootstraping
2018-04-10 17:41:41 <ncopa> i mean understand bootstrap
2018-04-10 17:41:46 <ncopa> not abuild
2018-04-10 17:41:50 <ncopa> but aports-build
2018-04-10 17:42:00 <jirutka> this hack is absurd, we’re just workarounding our own arbitrary limit
2018-04-10 17:43:11 <jirutka> ncopa: btw could you please summarize what we have discovered now into wiki page?
2018-04-10 17:43:24 <jirutka> ncopa: for future us and other people
2018-04-10 17:43:38 <ncopa> up til now (with rust) and ghc we did not allow new languages that required special bootstrap
2018-04-10 17:44:14 <_ikke_> Other languages don't depend on itself?
2018-04-10 17:44:23 <_ikke_> python/ruby/perl/...
2018-04-10 17:44:27 <ncopa> so its a relatively new issue, since java and go could bootstrap from C++
2018-04-10 17:44:41 <ncopa> python, ruby and perl are implemented in C
2018-04-10 17:44:44 <_ikke_> ok
2018-04-10 17:45:04 <skarnet> ISTR ruby was implemented for a large part in ruby
2018-04-10 17:45:22 <skarnet> but it may bootstrap from C indeed
2018-04-10 17:45:27 <jirutka> skarnet: CRuby (MRI) is implemented in C
2018-04-10 17:45:37 <jirutka> skarnet: the ruby interpret/compiler and part of stdlib
2018-04-10 17:45:44 <ncopa> actually, even gcc is problematic nowdays since its implemented in C++
2018-04-10 17:45:45 <skarnet> same for Perl, which is a lot of perl, but miniperl is C
2018-04-10 17:45:56 <skarnet> and yes, gcc is C++ :/
2018-04-10 17:46:07 <_ikke_> Right, so they have a basic version in C which can be used to build the complete environment
2018-04-10 17:46:25 <jirutka> skarnet: only the second part of stdlib is implemented in Ruby, but that’s irrelevant for this; Rubinius is the Ruby implementation written in Ruby, but we don’t even package that (it’s more or less, practically dead)
2018-04-10 17:46:30 <ncopa> _ikke_: exactly, thats what make the easy to work with
2018-04-10 17:47:12 <_ikke_> This probably also means that the language itself cannot use every feature to build the language, right?
2018-04-10 17:47:32 <jirutka> that’s how I solved it for crystal: https://github.com/alpinelinux/aports/blob/master/community/crystal/APKBUILD#L40-L63
2018-04-10 17:47:34 <_ikke_> feature / library
2018-04-10 17:47:54 <jirutka> I just build statically linked crystal and upload it to dev.a.o to be used for building next crystal versions
2018-04-10 17:48:14 <jirutka> the problem with rustc is that you cannot even built static rustc, it’s simple not possible!
2018-04-10 17:48:29 <jirutka> and it’s currently somehow fucked up with cargo
2018-04-10 17:54:52 <jirutka> btw OpenJDK JVM is implemented in C/C++, not in Java… probably the build process is just so complicated, that it’d be hard to build it incrementally without javac for bootstrapping
2018-04-10 17:57:37 <bernhardgruen[m]> Thanks, did not know that.
2018-04-10 17:59:32 <jirutka> Java is not low-level lang and it’s not compiled to native code, so it’s simply not usable for writing real VMs ;)
2018-04-10 18:17:12 <bernhardgruen[m]> Well there is a Java native interface and the library itself has some binary parts. I now use Java since 19 years but I never looked into javac.
2018-04-10 18:23:11 <jirutka> JNI is for “native extensions”; how is it related to this?
2018-04-10 18:23:13 <jirutka> what binary parts?
2018-04-10 18:28:25 <bernhardgruen[m]> The security extensions are binary .so files for example
2018-04-10 18:30:32 <bernhardgruen[m]> I have to search on my system. But I saw some .so files for java.
2018-04-10 18:31:47 <jirutka> yes
2018-04-10 18:32:14 <jirutka> but again, how is it related to bootstrapping? these binaries are built from C or C++ code
2018-04-10 18:37:07 <bernhardgruen[m]> I thought those jar files were compiled / generated during bootstrap.
2018-04-10 18:38:25 <bernhardgruen[m]> like rt.jar, jce.jar or resources.jar
2018-04-10 18:39:05 <jirutka> bernhardgruen[m]: is this history complete? https://dpaste.de/p4Og/raw ; it looks for me like I’ve missed some messages or something
2018-04-10 18:40:21 <bernhardgruen[m]> Yes.
2018-04-10 18:41:18 <jirutka> the question is if you need existing rt.jar, jce.jar etc. to build these or if it’s possible to compile .java sources to bytecode without them
2018-04-10 18:42:05 <jirutka> rt.jar is runtime, not sure if you need it to compile stuff
2018-04-10 18:44:04 <bernhardgruen[m]> I think it is needed to compile too. As I was not able to compile JDK 8 with jdk 6 installed for example (some years ago on arch)
2018-04-10 18:44:24 <bernhardgruen[m]> But I will test it now.
2018-04-10 18:44:54 <jirutka> that’s quite a different question
2018-04-10 18:45:18 <jirutka> we know that the current OpenJDK build system needs existing JDK to bootstrap
2018-04-10 18:45:44 <jirutka> but it may be just limitation of the build system, not that it’s really not possible to bootstrap it without it
2018-04-10 18:48:12 <jirutka> it’s easy to prove that it’s not possible for Rust and similar, because the compiler is really written in Rust, but Java compiler and even VM is not written in Java, but C/C++; it may depend on some parts written in Java, but I don’t know that and the fact that you need Java to build OpenJDK does not prove that
2018-04-10 18:50:13 <jirutka> however, you surely need some Java parts (e.g. rt.jar) to _run_ Java bytecode…
2018-04-10 18:54:53 <bernhardgruen> you can't even start javac: 'java/lang/NoClassDefFoundError: java/lang/Object'
2018-04-10 18:56:09 <jirutka> hm, aha
2018-04-10 18:58:38 <jirutka> actually, that makes sense, javac needs at least base java classes to compile anything depending on them (that’s basically, well, anything), to load signatures
2018-04-10 18:59:20 <bernhardgruen> compiler: com.sun.tools.javac (in tools.jar)
2018-04-10 19:00:31 <bernhardgruen> but you are right. there are some binary parts - written in c - like awt
2018-04-10 19:01:24 <jirutka> I think that jdk.compiler (replacement of com.sun.tools.javac) is just Java binding for native javac
2018-04-10 19:04:21 <jirutka> there is HotSpot, a lot of C and C++ code, plus some assembly: https://github.com/dmlloyd/openjdk/tree/jdk8u/jdk8u/hotspot
2018-04-10 19:06:02 <bernhardgruen> I think this still needs a working java environment as it needs the mercurial repository.
2018-04-10 19:09:52 <jirutka> hm, I found some notes that Java compiler is actually written in Java, just VM is in C/C++
2018-04-10 19:11:40 <jirutka> (btw I assume OpenJDK in all discussion, it may be different for other JVM/JDK implementations; for example I found that part of IBM JVM is (was?) written in SmallTalk!)
2018-04-10 19:12:22 <bernhardgruen> I don't think variants other than OpenJDK are really relevant anymore.
2018-04-10 19:12:36 <jirutka> agree
2018-04-10 19:13:17 <jirutka> well, except Dalvik or how it is called nowadays
2018-04-10 19:13:35 <jirutka> ART
2018-04-10 19:13:59 <bernhardgruen> But I fear building jdk will need some form of binary too. Maybe like this -> jdk 7 builds jdk 8 builds jdk 9 builds jdk 10. They intend to release half yearly.
2018-04-10 19:14:42 <bernhardgruen> ART - right. But is this still java? Or does it just use the DSL of java?
2018-04-10 19:14:52 <jirutka> that’s a good question! :)
2018-04-10 19:16:07 <jirutka> yea, half yearly… IMO just Oracle’s strategy how to get some money from Java, b/c Java is used mostly in enterprise and it’s total sci-fi to update JDK twice a year for enterprises, so they will pay for extended support
2018-04-10 19:19:16 <bernhardgruen> They release long term releases every 2 years. The idea behind the changed release plan was that in the past it took to long to bring the good features into the public. But I don't think that every company can afford to update twice a year. We intend to do this and I agreed to bring jdk 9 and 10 in an alpine version to use it as docker image.
2018-04-10 19:20:03 <jirutka> how it looks? do you already have abuild for JDK 9 or 10 ? :)
2018-04-10 19:20:04 <bernhardgruen> But I really don't know if I can finish this before 3.8.
2018-04-10 19:20:50 <bernhardgruen> I have built it on arch - but I use the "hack" that I download a bootstrap-jdk (as described in the building document for jdk 9 and 10) to bootstrap it.
2018-04-10 19:21:16 <bernhardgruen> Otherwise I would have to depend on 9 to build 10 to build 11 ...
2018-04-10 19:21:44 <bernhardgruen> and if gcj goes away - java is unbuildable with that strategy
2018-04-10 19:32:04 <bernhardgruen> jirutka: https://bugs.openjdk.java.net/browse/JDK-8188863 (jdk N-1 to build jdk N)
2018-04-10 19:33:13 <jirutka> hm, this will be “fun” :(
2018-04-10 19:37:08 <bernhardgruen[m]> Therefore I had the "idea" to download the binary (static) to build the JDK with it.
2018-04-11 01:14:11 <jirutka> why I’m doing that… https://github.com/rust-lang/rust/issues/48967 :( it’s just broken all the way down, like dancing on mine field; how the hell I can know where’s problem somewhere on my side and when I just hit yet yet yet another bug
2018-04-11 10:36:30 <terra> I have updated a init.d service file it has installed as .apk-new. How can I force to owerwrite it?
2018-04-11 10:37:32 <azarus> apk fix <pkg>
2018-04-11 10:37:55 <azarus> .apk-new files are files that haven't completed writing yet by apk
2018-04-11 10:38:21 <azarus> most likely you interrupted a apk process
2018-04-11 10:39:00 <_ikke_> More likely they are newer versions of config files that apk won't overwrite
2018-04-11 10:39:15 <azarus> Might also be.
2018-04-11 10:39:29 <_ikke_> but an init.d file is not a configuration file
2018-04-11 10:40:24 <jirutka> _ikke_: I guess terra probably modified this init.d file…
2018-04-11 10:40:35 <terra> I wanted to notice that.. I think apk treats all files in etc as config files. Am I wrong?
2018-04-11 10:40:51 <azarus> use the source, luke
2018-04-11 10:40:57 <azarus> 
ACTION digs into apk-tools

2018-04-11 10:40:59 <jirutka> yes, it does
2018-04-11 10:41:15 <jirutka> and why have you modified init.d file?
2018-04-11 10:41:48 <terra> because old one is not working anymore :)
2018-04-11 10:41:55 <jirutka> ??
2018-04-11 10:42:03 <jirutka> so we have broken init.d file in some package?
2018-04-11 10:42:06 <terra> major version upgrade
2018-04-11 10:42:21 <terra> I'm just testing
2018-04-11 10:42:24 <jirutka> aha
2018-04-11 10:42:45 <jirutka> well, then just run update-conf and accept/reject changes
2018-04-11 10:43:32 <terra> jirutka: it may require for config files but service files are just script files
2018-04-11 10:43:54 <terra> it doesn't make any sense
2018-04-11 10:43:58 <jirutka> no, they are treat the same as other files in /etc
2018-04-11 10:44:32 <jirutka> update-conf is just a trivial merge tool, it shows difference between <file> and <file>.apk-new and let you decide what to do
2018-04-11 10:44:49 <jirutka> you can also just mv <file>.apk-new <file>
2018-04-11 10:44:53 <terra> jirutka: it still doesn't make any sense
2018-04-11 10:44:58 <jirutka> not a rocket science
2018-04-11 10:45:01 <jirutka> why?
2018-04-11 10:45:21 <terra> service files aren't meant to altered
2018-04-11 10:45:29 <terra> but config files do
2018-04-11 10:45:58 <jirutka> but you have altered the “service” file, sigh
2018-04-11 10:46:05 <azarus> terra: well, send a patch to make apk-tools register /etc/init.d as "non" mutable or so
2018-04-11 10:46:15 <jirutka> azarus: no
2018-04-11 10:46:26 <azarus> (only for it to be rejected, of cours ;) )
2018-04-11 10:46:29 <azarus> course*
2018-04-11 10:46:50 <terra> jirutka: just like altering /bin/ls
2018-04-11 10:46:58 <jirutka> terra: have you manually modified /etc/init.d/foo? you said that yes
2018-04-11 10:47:22 <terra> jirutka: no. I installed as package
2018-04-11 10:47:52 <jirutka> [12:41:16]  jirutka:	and why have you modified init.d file?
2018-04-11 10:47:52 <jirutka> [12:41:49]  terra:	because old one is not working anymore :)
2018-04-11 10:48:24 <terra> jirutka: I meant via APKBUILD
2018-04-11 10:48:28 <jirutka> aha
2018-04-11 10:48:48 <terra> maybe I had modified the actual file.
2018-04-11 10:49:03 <terra> I stumbled
2018-04-11 10:50:38 <terra> during a normal package upgrade, it should owerwrite the old one, right? I'm asking this.
2018-04-11 10:51:06 <terra> if it is nor modified, of course
2018-04-11 10:51:11 <terra> *not
2018-04-11 10:51:20 <terra> by hand on system
2018-04-11 11:04:42 <jirutka> yes
2018-04-11 11:05:55 <mps> changed init files must not be overwritten by upgrade, IMHO
2018-04-11 11:06:46 <jirutka> yes
2018-04-11 11:06:55 <jirutka> and that’s exactly how it currently works
2018-04-11 11:09:39 <mps> hope it will stay as such :)
2018-04-11 12:29:43 <chambbj> Would appreciate a merge of PR at https://github.com/alpinelinux/aports/pull/3901 if everything looks to be in order.
2018-04-11 14:00:04 <clandmeter> A fresh, clean look for Gmail on the web.. hmm
2018-04-11 14:01:27 <clandmeter> finally getting gmail blue :)
2018-04-11 14:10:02 <_ikke_> New compared to what?
2018-04-11 14:13:51 <clandmeter> idk,  i just received an email regarding an upcoming change.
2018-04-11 17:58:51 <okeuday_bak> Please merge or comment on https://github.com/alpinelinux/aports/pull/3847
2018-04-11 18:02:14 <_ikke_> okeuday_bak: the devs are quite busy atm
2018-04-11 18:38:10 <okeuday_bak> _ikke_: np, just wanted to make sure it is in the queue somewhere
2018-04-11 18:46:41 <_ikke_> the queue is basically the list of open PRs :)
2018-04-11 22:04:08 <terra> fcolista: do you able to run dnscrypt-proxy as service with default install ?
2018-04-11 22:12:58 <terra> fcolista: I installed it and it doesn't launch. I have a working version and I want to submit it.
2018-04-11 22:42:43 <terra> fcolista: setup-dnscrypt script seems not working either -> tr: invalid character class ‘[0-9]’
2018-04-11 23:10:27 <Aerdan[m]> build-testing nheko 0.3.0, will submit PR for updating soon.
2018-04-11 23:12:47 <Aerdan[m]> aaaaand build failure.
2018-04-11 23:13:16 <Aerdan[m]> good ol' wall-of-text error messages, thanks gcc.
2018-04-11 23:15:27 <Aerdan[m]> https://dpaste.de/5ueL oh my fucking god.
2018-04-11 23:15:42 <Aerdan[m]> if that's what nontrivial C++ is going to spew when something goes wrong, I want nothing to do with it.
2018-04-11 23:51:05 <jirutka> rc-service-pid – find OpenRC service by PID and vice versa : https://gist.github.com/jirutka/5cf0e731dbc39a3a6fff0c6e1a2c2510
2018-04-12 02:21:18 <mitchty> ok question for a port, I have ghc 8.0 with x86 and armhf fine, but with 8.2.x and 8.4.x I apparently have found a bug i don't have the time/brains to figure out right now
2018-04-12 02:21:20 <mitchty> https://ghc.haskell.org/trac/ghc/ticket/14739
2018-04-12 02:21:55 <mitchty> would it be ok to split ghc into ghc80 and ghc84 where the latter provides an updated x86_64 that always seems to work and debug armhf on the side?
2018-04-12 02:23:00 <mitchty> that way at least there is an armhf ghc, if a bit out of date
2018-04-12 02:23:53 <awilfox> it's better to fix it.  maintaining multiple versions just for one or two arches seems silly when the older one works fine everywhere.
2018-04-12 02:23:58 <awilfox> see main/guile for another example of this.
2018-04-12 02:24:25 <mitchty> well the more i debug this the more down the rabbit hole I go
2018-04-12 02:24:51 <mitchty> now i'm into debugging binutils trying to figure out that its linking right
2018-04-12 02:25:23 <mitchty> i'll have a look at guile quick though
2018-04-12 02:26:01 <mitchty> its more i don't see why armhf should hold up x86_64 working fine
2018-04-12 02:27:42 <okeuday_bak> mitchty: ghc hasn't worked on armhf when attempting to use binary, last I tried
2018-04-12 02:28:15 <mitchty> i use ghc on armhf now, hold on, let me make sure i don't have a test apk i installed
2018-04-12 02:29:36 <okeuday_bak> arm support for ghc is still experimental, based on their website, for all versions
2018-04-12 02:30:37 <okeuday_bak> (e.g., https://www.haskell.org/ghc/download_ghc_8_4_1.html)
2018-04-12 02:32:10 <mitchty> sure, I know that all too well :) but ghc on armhf seems to work ok, albeit slow https://gist.github.com/mitchty/aedcad6d673bb8f5cd457693ee82c8eb
2018-04-12 02:32:45 <mitchty> yes my test machine's hostname is derp
2018-04-12 02:33:24 <okeuday_bak> mitchty: https://github.com/kolmodin/binary/issues/145 is what I ran into on Alpine Linux armhf
2018-04-12 02:34:33 <mitchty> let me see what I get on that
2018-04-12 02:36:30 <mitchty> there's a chance i'm also running patched binutils on this box, i've been testing a bit too much lately
2018-04-12 02:37:20 <mitchty> while i like learning, i'm not a fan of linkers right now :)
2018-04-12 03:06:16 <chambbj> https://github.com/alpinelinux/aports/pull/3887 now complete and ready for review
2018-04-12 03:26:03 <mitchty> okeuday_bak: so I got that to build fine, i'm guessing you're hitting that binutils bug
2018-04-12 03:26:46 <okeuday_bak> mitchty: ok, is it a problem with an older version of binutils, or is it a special binutils fix that needs to be added to binutils in the future?
2018-04-12 03:27:41 <mitchty> okeuday_bak: https://sourceware.org/bugzilla/show_bug.cgi?format=multiple&id=16177
2018-04-12 03:28:16 <mitchty> i had a look and have a binutils that has that patch applied to it
2018-04-12 03:28:42 <mitchty> which should be the only change between a stock alpine linux binutils
2018-04-12 03:29:12 <mitchty> that said, its not in the upstream yet either
2018-04-12 03:29:38 <okeuday_bak> mitchty: ok, thanks for the information though
2018-04-12 03:30:07 <mitchty> it does help with arm linking at least, i'm just trying to make sure its right for 8.2.2/8.4.1 too
2018-04-12 03:30:31 <mitchty> but i've gone down a very deep rabbit hole in llvm/arm/binutils/linkers
2018-04-12 03:33:40 <mitchty> i'll keep working on it, but i feel like i know less than I did every day looking at this
2018-04-12 03:51:47 <deedubs> is there a subpackage_install="" that I can use to setup post-install for subpackages? or do I need to make them standalone at that point?
2018-04-12 03:56:10 <awilfox> deedubs: put install="" in the subpackage's function.  see main/apache2's ssl() for an example.
2018-04-12 03:57:09 <deedubs> ah I believe I had put install = "..."
2018-04-12 03:59:06 <deedubs> thanks awilfox
2018-04-12 04:11:25 <awilfox> np :)
2018-04-12 06:28:06 <skarnet> <mitchty>well the more i debug this the more down the rabbit hole I go
2018-04-12 06:28:10 <skarnet> story of our lives
2018-04-12 06:29:17 <skarnet> and for a lot of software, I'm in no hurry to meet the giant-ass rabbit who dug this hole to Earth's core
2018-04-12 06:34:14 <bernhardgruen[m]> jirutka: would you mind moving py3-llfuse to community? It is needed by borgbackup in case you like to use borg mount (via fuse). Don't know if that dependency should be added to borgbackup by default.
2018-04-12 11:16:52 <clandmeter> busybox wget doesnt like proxy settings?
2018-04-12 11:47:11 <bernhardgruen[m]> I am quite sure it does as I use it in chroot of alpine-make-vm-image to generate the images faster.
2018-04-12 11:47:54 <bernhardgruen[m]> Oh no wait. I don't use wget from busybox. Sorry.
2018-04-12 12:50:55 <clandmeter> https doesnt seems to work. probably the way its implemented in bb is causing it.
2018-04-12 12:51:52 <jirutka> busybox wget is quite weird, it happened to me on some versions that it doesn’t work with https, but don’t know why
2018-04-12 13:21:43 <jirutka> fcolista: kaniini: how can I perform `apk upgrade` and exclude specific packages (e.g. kernel) from upgrading, without temporarily modifying /etc/apk/world to pin it to the current version?
2018-04-12 13:46:15 <terra> jirutka: tagging as dummy repos?
2018-04-12 13:46:29 <jirutka> terra: ?
2018-04-12 13:46:50 <fcolista2> jirutka, there's no --exclude..so I wouldd check with --simulate which one are going to be upgraded, and then upgrade the single packages with a script
2018-04-12 13:47:11 <fcolista2> apk upgrade --simulate > list-of-packages
2018-04-12 13:47:34 <jirutka> fcolista2: is it safe to do so?
2018-04-12 13:47:42 <terra> jirutka: I mean, treat it like installed from a custom repo tagged like @custom
2018-04-12 13:47:58 <fcolista2> what you mean with "safe"?
2018-04-12 13:48:13 <jirutka> fcolista2: hm, actually, if A depends on B, then apk upgrade A will automatically upgrade even B
2018-04-12 13:48:43 <fcolista2> yes. If you use apk add -u $package inside list-of-packages
2018-04-12 13:49:02 <jirutka> fcolista2: I mean that apk does not perform upgrade in some arbitrary order and also it runs all post-* scripts _after_ upgrading whole bunch, right?
2018-04-12 13:49:33 <jirutka> 
ACTION openning issue for adding --exclude to apk upgrade

2018-04-12 13:50:11 <fcolista2> post-* scripts ran after upgrading, yes. Regarding the order, it depends on the order you ran apk add -u $package
2018-04-12 13:51:01 <fcolista2> for p in $(cat list-of-packages | grep -v grsec), do apk add -u $p, done
2018-04-12 13:57:36 <jirutka> fcolista2: okay, this is probably the best I can do for now, thanks!
2018-04-12 13:59:48 <fcolista2> np
2018-04-12 13:59:55 <fcolista2> no other ideas from my side :)
2018-04-12 14:14:55 <clandmeter> jirutka, busybox has its own unsafe https feature.
2018-04-12 14:15:10 <jirutka> clandmeter: unsafe https feature??
2018-04-12 14:15:12 <clandmeter> if it cannot find ssl it will use its own without any cert verification
2018-04-12 14:15:40 <clandmeter> similar like wget --allow-untrusted or something like that.
2018-04-12 14:17:45 <jirutka> uh, WHAT?!
2018-04-12 14:17:59 <jirutka> can we disable this?
2018-04-12 14:18:25 <jirutka> not let it NEVER do ssl without cert verification
2018-04-12 14:18:37 <clandmeter> i talked to ncopa about it
2018-04-12 14:19:01 <clandmeter> it should probably be implemented with --no-check-certificate similar like regular wget does
2018-04-12 14:21:03 <clandmeter> jirutka, https://github.com/mirror/busybox/blob/master/networking/wget.c#L90 ;-)
2018-04-12 14:23:01 <jirutka> i don’t care about this long explanation, this is simply wrong to do HTTPS without TLS verification, i.e. making TLS totally useless, without even warning user about it
2018-04-12 14:23:18 <jirutka> it’d be okish with explicit option --no-check-certificate, but definitely not as default
2018-04-12 14:23:42 <clandmeter> send the patch :)
2018-04-12 14:24:43 <clandmeter> for now we could simply disable it.
2018-04-12 14:24:58 <jirutka> yes, +1 for disabling
2018-04-12 16:17:43 <terra> Any dev could check this: http://lists.alpinelinux.org/alpine-aports/5301.html
2018-04-12 16:18:01 <terra> vlc can't do https without tls support
2018-04-12 16:20:13 <terra> this patch is waiting for to be reviewed for 3 weeks
2018-04-12 16:27:52 <mitchty> skarnet: yep :) its fun and all, but a bit of a learning curve
2018-04-12 17:26:46 <jirutka> Find processes that use (maps into memory) files which have been deleted or replaced on disk: https://gist.github.com/jirutka/9f7624a54b6f44b31c1dce3bea11708f
2018-04-12 17:34:29 <liwakura> im currently doing PXE booting with alpine, and i noticed that none of the install media support networking from the initrd
2018-04-12 17:34:46 <liwakura> is that because of the NIC drivers?
2018-04-12 17:46:23 <terra> liwakura: maybe your initramfs has to rebuilt with networking modules
2018-04-12 17:46:34 <terra> liwakura: mkinitfs -L
2018-04-12 17:46:46 <liwakura> yes it has
2018-04-12 17:46:57 <liwakura> but that requires an full rootfs checkout
2018-04-12 17:47:40 <liwakura> i was wondering why its not enabled per default
2018-04-12 17:48:01 <liwakura> rebuilding the initramfs is why network install doesn't work on alpine per upstream binaries
2018-04-12 17:48:16 <liwakura> binaries as in, whats on the downlade page
2018-04-12 17:49:12 <terra> did you rebuilt for yourself?
2018-04-12 17:49:27 <liwakura> only for native arch
2018-04-12 17:50:20 <liwakura> just appending another cpio archive to the existing initrd doesn't work because of depmod..
2018-04-12 17:50:28 <terra> i understand but maybe you can manually add required modules via unpack/repack initramfs manually
2018-04-12 17:50:59 <terra> I mean a mirror copy of your native arch initramfs
2018-04-12 17:51:24 <liwakura> adding network support to the initramfs is just a matter of shipping the installer images with initfs_features="... network"
2018-04-12 17:51:36 <liwakura> but not sure about the driver situation
2018-04-12 17:51:39 <liwakura> thats why i asked
2018-04-12 17:52:32 <terra> that's where initramfs comes in: the modules
2018-04-12 17:52:55 <liwakura> does it make sense to include some of the modules?
2018-04-12 17:53:10 <liwakura> the virt image works when generated with initfs_features="... network"
2018-04-12 17:59:06 <terra> you can try other distro kernel+initramfs with alpine rootfs
2018-04-12 17:59:27 <liwakura> *sigh*
2018-04-12 17:59:32 <liwakura> terra: let me explain my usecase
2018-04-12 17:59:41 <liwakura> i launch qemu, give an kernel and an initramfs in
2018-04-12 17:59:52 <liwakura> in the bootparams, i set the modloop url and the apkovl url
2018-04-12 18:00:00 <liwakura> intention is to launch a system using that apkovl
2018-04-12 18:00:22 <liwakura> = quick cross-arch provisioning of workers
2018-04-12 18:01:22 <liwakura> to do that for armhf, i'd need to generate an armhf boot media from an x86_64 system
2018-04-12 18:01:30 <liwakura> so i'd need an qemu setup
2018-04-12 18:03:28 <liwakura> i can try again, but last time i checked the qemu-user-arm alpine ships was not working reliable
2018-04-12 18:04:42 <liwakura> so, how do i best bootstrap an armhf initramfs with networking enabled?
2018-04-12 18:32:42 <tmh1999> recently BOOT_IMAGE kernel parameter was introduced. Anyone knows which commit ?
2018-04-12 18:35:33 <liwakura> tmh1999: afaik syslinux put that there since ever
2018-04-12 19:48:15 <clandmeter> liwakura, boot.a.o
2018-04-12 19:48:58 <liwakura> clandmeter: godsend!
2018-04-12 19:50:00 <liwakura> this needs to be documented somewhere, i've seen people asking for it quite often
2018-04-12 19:50:46 <liwakura> but still no armhf..
2018-04-12 19:51:27 <clandmeter> It's new
2018-04-12 19:51:55 <liwakura> right
2018-04-12 19:51:59 <clandmeter> I'm going offline
2018-04-12 19:52:15 <clandmeter> Ping me later if you have issues
2018-04-12 19:52:40 <clandmeter> Or suggestions
2018-04-13 03:01:51 <awilfox> can you use `echo` in a .post-install?
2018-04-13 03:02:17 <Shiz> I don't see why not, particularly
2018-04-13 03:02:20 <Shiz> sup awilfox \o
2018-04-13 03:02:23 <Shiz> been a bit
2018-04-13 03:04:03 <awilfox> Shiz: hi!  we were just talking about you on #adelie.
2018-04-13 03:04:15 <Shiz> o?
2018-04-13 03:04:20 <Shiz> not missing out on any juicy rumours i hope
2018-04-13 03:04:28 <awilfox> Shiz: do you have any further info to share about OtherOS?  I actually acquired a "fat" PS3, but it has that firmware update applied so it won't do the native OtherOS :/
2018-04-13 03:04:41 <awilfox> we were wondering if you had any info about how to make it work
2018-04-13 03:04:42 <Shiz> awilfox: you can do a soft-downgrade these days
2018-04-13 03:04:44 <Shiz> :)
2018-04-13 03:04:57 <Shiz> and for CFW's ps3s there's otheros++
2018-04-13 03:04:58 <awilfox> ohh!  okay.  I didn't find that on a quick web search.  I guess I'll look again
2018-04-13 03:05:24 <awilfox> the other problem with it is that the red laser doesn't work (it won't read CDs or DVDs, just Blu-Rays)
2018-04-13 03:05:51 <pseudor00t> hello campain #deletefacebook https://www.facebook.com/buh0x/videos/2059902154334625/UzpfSTE0ODg3ODYxNTE1MzMyOToyMDA5MzAxMzQ5MTExMDM3/
2018-04-13 03:06:00 <Shiz> nah
2018-04-19 04:01:53 <tmh1999> looks like busybox's mount does not support mounting option uid and gid ?
2018-04-19 04:02:43 <tmh1999> it's ext4
2018-04-19 05:26:12 <clandmeter> terra, the problem with contributions is that we have too many new contributors and too little reviewers. And ATM the more experienced have personal and business like things to take care of.
2018-04-19 05:27:50 <clandmeter> But I'm sure the pace will get back soon. It always have been.
2018-04-19 07:46:27 <ncopa> hi
2018-04-19 07:46:47 <ncopa> we get a ton of PRs and we do close many of them
2018-04-19 07:47:13 <ncopa> unfortunally, many of the PRs needs feedback and explain how to be improved
2018-04-19 07:47:22 <ncopa> sometimes i just fix it myself
2018-04-19 07:47:44 <ncopa> sometimes a PR takes days of work (mariadb upgrade did)
2018-04-19 07:48:06 <ncopa> we are very well aware of the problem
2018-04-19 08:07:36 <ncopa> also, previously i could do a number of trivial maintenance updates half automatic
2018-04-19 08:08:05 <ncopa> now i need to manually check every update with the PR queue, so i dont step on someone elses toes
2018-04-19 08:08:46 <ncopa> and importing a PR is more manual work than doing update myself
2018-04-19 08:09:26 <clandmeter> yes i agree
2018-04-19 08:09:49 <clandmeter> its more complex now.
2018-04-19 08:10:04 <clandmeter> but its an investment i hope.
2018-04-19 08:10:19 <clandmeter> train other people how to make good contributions.
2018-04-19 08:11:27 <ncopa> what can help that is documentation
2018-04-19 08:11:54 <ncopa> also, the amount of work is increasing with the added tests
2018-04-19 08:12:23 <ncopa> i spend some time on investigating if the test itself is broken or if its a real bug in the app
2018-04-19 08:12:32 <ncopa> often the test is broken
2018-04-19 08:12:34 <clandmeter> yes, and that reminds me.
2018-04-19 08:12:53 <clandmeter> regarding new release, we should include it in the notes.
2018-04-19 08:13:22 <ncopa> about tests?
2018-04-19 08:13:37 <clandmeter> yes some figures about how much aports have tests now.
2018-04-19 08:14:24 <clandmeter> i think it shows our quality has increased.
2018-04-19 08:14:39 <ncopa> yeah, good point
2018-04-19 08:15:13 <clandmeter> when did we actually start to implement them?
2018-04-19 08:25:49 <andypost[m]> yep, tests are main pita, most of them require some server runnnig but no idea how to implement it in travis
2018-04-19 09:33:15 <ncopa> often the tests are buggy: https://github.com/jbboehr/handlebars.c/commit/931ad82c17984935decebacd3283e9d26ca749cb
2018-04-19 09:33:29 <ncopa> i have another test failure in same project now
2018-04-19 10:33:08 <_ikke_> "Should fix problem" :D
2018-04-19 11:52:08 <mepholic> https://flux.hxx.in/meph/screenshots/Screenshot-20180419-045144.079406154.png
2018-04-19 12:27:20 <terra> ncopa, clandmeter: thanks for response. hope pace will increase and even get better soon.
2018-04-19 12:42:27 <andypost> any idea why x86 build fails http://build.alpinelinux.org/buildlogs/build-edge-x86/community/yarn/yarn-1.6.0-r0.log
2018-04-19 12:43:06 <andypost> looks like patch was trying to be applied to already patched build
2018-04-19 12:48:53 <terra> andypost: looks like the patch is already applied to upstream. better compare the patch and original file (package.json) line by line.
2018-04-19 12:49:58 <terra> or track the git commits of upstream related with this file (git log of this file on upstream)
2018-04-19 12:51:59 <andypost> terra, it is not applied, I did test on 86-64 & arm 8
2018-04-19 12:52:35 <andypost> terra, it fails only at build time
2018-04-19 12:54:07 <terra> to make sure, compile the package step by step: abuild unpack ; abuild prepare; abuild build
2018-04-19 12:54:45 <terra> "abuild clean" first of course
2018-04-19 12:56:04 <andypost> terra, yes, that's done but it fails on build server only
2018-04-19 12:59:04 <terra> andypost: I don't see any error except wrong patch. If you put a patch file in APKBUILD that doesn't apply correctly, build process will fail.
2018-04-19 13:01:21 <andypost> terra, other arches build fine, so patch valid
2018-04-19 13:27:14 <chambbj> @clandmeter https://github.com/alpinelinux/aports/pull/3887 should now be ready. All enabled tests pass locally with abuild rootbld.
2018-04-19 13:27:46 <terra> interesting.. i tested on my x86 build machine and it built and packed just fine.
2018-04-19 13:28:02 <terra> andypost ^
2018-04-19 13:28:32 <andypost> terra, thanks, that's why I wondered... maybe @ncopa knows
2018-04-19 13:29:54 <terra> andypost: but I don't use cross-compiling.
2018-04-19 13:30:28 <andypost> ah... it could be it! I know nothing about build "factory"
2018-04-19 13:42:30 <duncan^> Hmm. I've made a PR: https://github.com/alpinelinux/aports/pull/4069. However, from CI, it fails on "Setting up Alpine Linux". I am not sure if this is my fault or if it of the CI's own violution.
2018-04-19 13:42:50 <duncan^> (FWIW it works fine locally, so I'm guessing this is something upstream)
2018-04-19 14:06:54 <clandmeter> chambbj: ok ill try a local build.
2018-04-19 14:08:37 <ncopa> duncan^: yes looks like cryptsetup is broken
2018-04-19 14:22:01 <ncopa> ok i know what happened
2018-04-19 14:32:39 <clandmeter> ncopa: what happend to the builder(s)?
2018-04-19 14:32:47 <ncopa> i broke them
2018-04-19 14:32:54 <clandmeter> yes i figured that :)
2018-04-19 14:32:54 <ncopa> by pushing json-c
2018-04-19 14:33:10 <ncopa> json-c 0.13 provides so:libjson-c.so.4
2018-04-19 14:33:12 <clandmeter> nice verbose error msg.
2018-04-19 14:33:33 <ncopa> previous json-c (0.12) had so:libjson-c.so.2
2018-04-19 14:33:48 <ncopa> cryptsetup depends on json-c
2018-04-19 14:33:59 <ncopa> and mkinitfs depends on cryptsetup
2018-04-19 14:34:05 <ncopa> mkinitfs is installed on the builders
2018-04-19 14:34:36 <ncopa> so what happened was that json-c got upgraded
2018-04-19 14:34:49 <ncopa> but old json-c was still in the local repository
2018-04-19 14:35:30 <clandmeter> do we need to have it installed? mkinitfs?
2018-04-19 14:35:53 <ncopa> when the rebuild of cryptsetup was done (and the others), the json-c-dev dependency pulled in the 0.12 version because 0.13 would conflict with the installed so:libjson-c.so.2
2018-04-19 14:36:13 <ncopa> so everything in main got rebuilt with the old json-c instead of the new
2018-04-19 14:36:30 <ncopa> when main was built the old json-c was removed
2018-04-19 14:36:50 <ncopa> so when building community we got an error
2018-04-19 14:37:04 <clandmeter> i guess we have it installed to create releases?
2018-04-19 14:37:12 <ncopa> correct
2018-04-19 14:37:17 <ncopa> that is why it is installed
2018-04-19 14:37:29 <ncopa> but technically, we dont need it to be installed all the time
2018-04-19 14:37:33 <clandmeter> right
2018-04-19 14:37:37 <clandmeter> that was my thought.
2018-04-19 14:37:42 <clandmeter> add it to the script
2018-04-19 14:38:07 <clandmeter> i think i bumped into it on boot.a.o
2018-04-19 14:54:05 <terra> I was getting this error when building packages: "no names found cannot describe anything"
2018-04-19 14:54:22 <terra> But abuild wasn't throwing error
2018-04-19 14:54:59 <terra> Now it fails at building index stage
2018-04-19 15:08:18 <tmh1999> hi ncopa, just checking if matrix/freenode problem. wonder if you got my messages.
2018-04-19 15:08:39 <ncopa> no i didnt get your messages
2018-04-19 15:08:43 <ncopa> i logged out from matrix
2018-04-19 15:08:55 <tmh1999> oh boy matrix ...
2018-04-19 15:09:03 <tmh1999> okay I will pm you in freenode here
2018-04-19 15:09:16 <ncopa> yeah.. matrix has been painful
2018-04-19 15:15:04 <Adran> ncopa: clandmeter: is there a way for me to use a apkovl / lbu to save state and restore after setup is complete? i have a custom live image that runs stuff live, and after install is finished I want to sync with the install
2018-04-19 15:15:27 <Adran> also matrix was the most unresponsive thing i've seen in a long time heh
2018-04-19 15:18:53 <terra> networkmanager package says add your user to "plugdev" group..but this group is created by udevil package
2018-04-19 15:19:26 <terra> and network manager is not any related with udev package
2018-04-19 15:19:58 <terra> *udevil
2018-04-19 15:27:09 <ncopa> Adran: hum, i know you can do setup-disk to install the running system to disk
2018-04-19 15:28:07 <ncopa> terra: im not sure who has responsibility to create plugdev group
2018-04-19 15:28:51 <Adran> ncopa: how?
2018-04-19 15:28:59 <Adran> -o?
2018-04-19 15:30:48 <terra> I did a "grep -rn" in aport tree only udevil and librdti are adding plugdev group to system via pre-install
2018-04-19 15:31:49 <terra> modemmanager is also needs plugdev group
2018-04-19 15:32:32 <ncopa> Adran: at setup-alpine select no disk, then add you packages and rc-update add your services
2018-04-19 15:32:37 <ncopa> finally you run setup-disk
2018-04-19 15:32:38 <terra> sorry *libftdi1
2018-04-19 15:32:59 <ncopa> terra: i wonder if we should ship it with alpine-baselayout
2018-04-19 15:32:59 <Adran> ncopa: great! thats saving me a ton of time
2018-04-19 15:36:08 <terra> I think plugdev must have added by eudev/udev
2018-04-19 15:38:49 <terra> didn't see this: i wonder if we should ship it with alpine-baselayout
2018-04-19 15:39:06 <terra> ncopa: yes probably
2018-04-19 15:47:59 <Truxx> Hmmm... I have modemmanager, but no plugdev - or was this remark regarding not the recent version?
2018-04-19 15:51:55 <terra> Truxx : probably no. plugdev group seems became out of scope somehow.
2018-04-19 15:53:33 <Truxx> I see, thx. It looks like that as modemmanager is working fine. Groups: root bin daemon sys adm disk wheel
2018-04-19 15:53:44 <Truxx> floppy dialout tape video
2018-04-19 15:54:20 <terra> Truxx : can you manage it with normal user on desktop?
2018-04-19 15:55:28 <Truxx> Yes, via mmcli, without networkmanager
2018-04-19 15:55:32 <terra> For instance I can't manage NetworkManager with normal user.
2018-04-19 15:57:15 <Truxx> Hmmm, there is something called nmcli if I remember correctly
2018-04-19 15:58:26 <Truxx> Doesn't it work without privileges? At least mmcli works fine, only setting the ip with ifconfig and the gateway with route require root
2018-04-19 16:02:38 <terra> NM uses policykit. I'm not using modemmanager.
2018-04-19 16:02:54 <terra> At least not yet.
2018-04-19 16:07:21 <Truxx> Somehow I started to dislike networkmanager and realised there is a beautiful life without it :)
2018-04-19 16:08:02 <Truxx> Btw what I meant regarding mmcli from userspace works after modemmanager was started as root.
2018-04-19 16:17:48 <duncan^> ncopa: Thanks, I think the fix for json-c has fixed that issue.
2018-04-19 16:18:27 <duncan^> but I',m now getting "fatal: No names found, cannot describe anything.
2018-04-19 16:18:29 <duncan^> "
2018-04-19 16:19:01 <duncan^> I wonder what that means?
2018-04-19 16:19:39 <terra> Truxx : I like to see small network icon at toolbar. Nothing vital with NM if you use a highly customized desktop.
2018-04-19 16:21:07 <duncan^> seems to be related to git describe
2018-04-19 16:21:27 <duncan^> I'm recalling how Travis broke a few days ago...
2018-04-19 16:22:32 <tmh1999> duncan^: maybe you clone aports with no tags ?
2018-04-19 16:22:58 <tmh1999> I usually have that error when runnning abuild and "git --describe" cant find no tag
2018-04-19 16:23:16 <tmh1999> clone --depth 1 maybe
2018-04-19 16:29:07 <Truxx> @terra I can understand. For me it was the exact opposite: I started to consider that icon as bloat and it was just annoying me. For info I prefer something like bmon, slurm or a simple conky.
2018-04-19 16:36:08 <terra> duncan^ I got same "fatal: No names found, cannot describe anything." error. Package building fails during indexing stage due to this.
2018-04-19 16:36:37 <terra> duncan^ : U was getting this error before but nothing was vital.
2018-04-19 16:36:43 <terra> *I
2018-04-19 17:40:51 <duncan^> tmh1999: No. If you read what I said, you'd understand that's not the case.
2018-04-19 19:04:16 <deedubs> Has anyone taken a kick at the SCAP / OpenSCAP can yet?
2018-04-19 19:12:43 <deedubs> I know we have cvechecker
2018-04-19 19:29:46 <liwakura> deedubs: i really /dislike/ how CVE enables a whole lot of management and attention crap
2018-04-19 19:30:03 <liwakura> i wish computer security was again about actually /fixing bugs/
2018-04-19 19:31:31 <deedubs> as a developer sure, as someone operating a mission critical high security system it'd be nice to know what I should expect developers to be /fixing bugs/ in
2018-04-19 19:31:54 <liwakura> hopefully all bugs
2018-04-19 19:32:24 <liwakura> not just the ones that can possible exploited by malicious actors
2018-04-19 19:33:00 <liwakura> security also includes system operating safety, which CVE's largely ignore
2018-04-19 19:35:35 <deedubs> again thats a nice sentiment, but if there is a bug in some software that could allow data to be exfiltrated vs a machine occasically strokes-out and needs to replaced I'm going to want to be able to attempt to mitigate any exfiltrated even if I have to operate my cluster in a reduced capacity / functionality until its resolved
2018-04-19 19:36:15 <liwakura> *shrugs*
2018-04-19 20:33:22 <duncan^> The issue with No names foound seems related to https://github.com/alpinelinux/abuild/commit/fd97176583ed7a13d4fd63ecbad3d436fe14d0b2
2018-04-19 20:45:35 <silverwillow> hello all, any tips on how to go about creating a personalised distro based on Alpine? Looking to keep musl/busybox/toybox and Alpine package management but want to fiddle with a new type of installer, preconfigured applications and documentation.
2018-04-19 20:46:23 <terra> duncan^: Thanks. Setting git=true at abuild.conf is discarding the problem. But not sure that it is a safe workaround.
2018-04-19 20:50:36 <TBB> silverwillow: isn't that all there is to it really? :)
2018-04-19 20:56:03 <silverwillow> You're quite right. I'll stop being lazy and go RTFM and peruse the Alpine installer sources :)
2018-04-19 20:59:07 <TBB> if it gives you any further motivation, I went through the same thing about two years ago
2018-04-19 20:59:17 <TBB> and it was a fascinating journey
2018-04-19 21:18:29 <Shiz> silverwillow: most of the interesting stuff can just be cherry-picked
2018-04-19 21:18:36 <Shiz> in fact
2018-04-19 21:18:42 <Shiz> most of the interesting stuff is in aports that you may wanna modify
2018-04-19 21:18:54 <Shiz> so maybe start by forking aports for yourslf and see which packages you wanna modify/keep/drop
2018-04-19 21:18:56 <Shiz> :)
2018-04-19 21:20:32 <silverwillow> Have dived into the wiki. Expect to be 'living' there for a few days/weeks :)
2018-04-19 21:23:54 <terra> Guys, I added "provide net" to networkmanager service script to make it launch before services requiring network.
2018-04-19 21:24:00 <silverwillow> TBB: I'm comcing back to GNU/Linux after a decade... i'm old, started on SLS and kernel 1.2.x. First MILO on Alpha AXP 'Jensen' arch was something I had a hand in.... But in my 'dotage' i've gotten lazy. Well, I'll get back under my rock. If I ever finish this project I'll let the Alpine community know about it. Take care!
2018-04-19 21:25:05 <terra> when I restart it, it stopped chronyd but didn't trigger chronyd start
2018-04-19 21:25:34 <TBB> you too silverwillow, and happy journey :)
2018-04-19 21:25:57 <TBB> (oop, that could be taken the wrong way)
2018-04-19 22:53:29 <skarnet> silverwillow, if you start on this road, don't expect it to ever finish
2018-04-19 22:53:39 <skarnet> making a distro is a long, neverending project
2018-04-19 22:53:50 <skarnet> and it has to be maintained
2018-04-19 22:54:07 <skarnet> you need to know what you're signing up for before you jump into it. :)
2018-04-20 00:29:52 <terra> most of linux-firmware packages are not needed by mt system. How can I select only needed ones to be updated?
2018-04-20 00:30:04 <terra> *my system
2018-04-20 08:32:39 <ncopa> tmh1999: can you help me with liboil? I have a patch which makes the tests pass, but i'm not 100% sure about the s390x asm: https://dpaste.de/B5JF/raw
2018-04-20 08:35:24 <ncopa> it fixes a stack overflow, but im not sure it returns same value
2018-04-20 08:40:53 <awilfox> been meaning to poke about that: liboil is no longer used in gst1.0
2018-04-20 08:40:55 <awilfox> only 0.10
2018-04-20 08:41:17 <awilfox> since liboil is not maintained any more and does not support ppc64 I've been trying to prune it
2018-04-20 08:41:22 <ncopa> yeah
2018-04-20 08:41:24 <ncopa> i saw
2018-04-20 08:41:28 <awilfox> ah good :)
2018-04-20 08:41:41 <ncopa> i think even gst 0.10 uses orc
2018-04-20 08:42:04 <ncopa> there is one single package, swfdec that uses liboil
2018-04-20 08:42:18 <ncopa> and there was one in testing that depends on swfdec
2018-04-20 08:42:52 <ncopa> i dont think any of our users will ever try play flash videos on s390x
2018-04-20 08:48:07 <awilfox> hm
2018-04-20 08:48:27 <awilfox> it was mlt, wasn't it
2018-04-20 08:49:18 <awilfox> yeah.  we removed the swfdec dep from mlt in our build for adelie, which removes the ability to process flash files with it, but also made it build on ppc64 :p
2018-04-20 08:53:32 <ncopa> swfdec does not build on ppc64le?
2018-04-20 08:53:44 <ncopa> i think adobe flash is dead anyways
2018-04-20 08:55:29 <awilfox> adelie does not target ppc64le
2018-04-20 08:55:32 <awilfox> we target ppc64
2018-04-20 08:55:34 <awilfox> that is BE
2018-04-20 08:55:54 <awilfox> and liboil does not build on ppc64 BE, with musl, because it does JIT
2018-04-20 08:56:02 <awilfox> and ppc64 BE has two ELF ABIs
2018-04-20 08:56:09 <awilfox> musl implements ELFv2 only, which is not supported by liboil
2018-04-20 08:57:55 <ncopa> ok
2018-04-20 11:39:53 <tboerger[m]> have the travis builds been fixed? :)
2018-04-20 11:47:00 <tboerger[m]> yay, hugo tests are failing on alpine -.-
2018-04-20 11:49:59 <tboerger[m]> clandmeter: should i start patching out specific test cases to get hugo building on alpine?
2018-04-20 11:50:36 <tboerger[m]> https://github.com/gohugoio/hugo/blob/v0.39/hugolib/page_test.go#L913-L939 this test fails because i think the used tarball lacks any git information.
2018-04-20 12:30:09 <duncan^> I think Travis builds have not been fixed somehow
2018-04-20 12:39:39 <tboerger[m]> the build itself works fine again (at least for me) but i really have to patch out test cases... deleting single test files results in different tests are failing because they use definitions of the deleted test file -.-
2018-04-20 12:52:51 <clandmeter> tboerger[m], do you know why they fail?
2018-04-20 12:53:03 <clandmeter> you could patch them and report it upstream.
2018-04-20 12:53:24 <tboerger[m]> clandmeter: as said above, some tests require the .git folder
2018-04-20 12:53:36 <clandmeter> sorry i didnt have time to check
2018-04-20 12:54:00 <clandmeter> git folder?
2018-04-20 12:54:06 <clandmeter> what does it test for?
2018-04-20 12:54:56 <tboerger[m]> https://github.com/gohugoio/hugo/blob/master/releaser/git_test.go this requires git ingormations that are not available within the release tarballs
2018-04-20 12:55:15 <tboerger[m]> same for https://github.com/gohugoio/hugo/blob/v0.39/hugolib/page_test.go#L913-L939
2018-04-20 12:56:00 <tboerger[m]> this works fine for a git clone of the project, but not for the tarball from the releases.
2018-04-20 12:58:26 <clandmeter> maybe add a check to see if it exists?
2018-04-20 13:08:24 <tboerger[m]> yeah maybe, but that won't help to get the current version built :)
2018-04-20 13:08:43 <tboerger[m]> that's something i can communicate upstream when i've pinned all required changes... for now i will prepare a patch that drops the broken test cases.
2018-04-20 14:12:34 <chambbj> I've recently submitted a PR (https://github.com/alpinelinux/aports/pull/4075) that appears to build, test, and install correctly, and yet Travis fails with a not-so-obvious error message (https://travis-ci.org/alpinelinux/aports/builds/369142357#L2274-L2275). I'd appreciate it if someone can provide some feedback.
2018-04-20 14:13:57 <chambbj> Regarding the same, everything checks out fine locally with abuild rootbld.
2018-04-20 14:19:08 <terra> I want to reinstall somepackage's new build with 'same' version and release number via "apk fix --reinstall".
2018-04-20 14:20:34 <terra> But apk doesn't fetsh the new build from my local repo. But I don't want to reinstall cached one.
2018-04-20 14:21:35 <terra> Even if I remove the cached one, it doesn't fetch the new apk from repo.
2018-04-20 14:23:35 <terra> In short, how can I reinstall the packages with same version and release number but different hash.
2018-04-20 14:23:59 <TBB> del and add?
2018-04-20 14:24:30 <terra> there are dependencies and I don't want to reinstall them.
2018-04-20 14:25:27 <terra> This should be typical case for contributors who want to test their build. Isn't it?
2018-04-20 14:47:06 <tmh1999_> ncopa: me neither. I'll see if I can confirm...
2018-04-20 14:47:33 <ncopa> tmh1999_: i removed liboil package so it does not matter
2018-04-20 14:47:49 <tmh1999> ncopa: yeah I hope it'll be gone. project is really dead.
2018-04-20 14:48:08 <ncopa> but i am slightly curious if i did the s390x asm correct
2018-04-20 14:48:13 <tmh1999> :D
2018-04-20 14:48:18 <ncopa> or copy/pasted it correct :)
2018-04-20 14:48:32 <tmh1999> recorded it. I'll have my chance to improve my asm skill soon.
2018-04-20 20:02:34 <PureTryOut[m]> hmm, can't subpackages install init files?
2018-04-20 20:02:36 <PureTryOut[m]> I get permission denied
2018-04-20 20:40:26 <mksully22> I submitted PR https://github.com/alpinelinux/aports/pull/4077 to fix the testing/pcl build issue. Problem is that it fails Travis because the job takes too long to build. Is there a parameter or something I can add to the pull request to make sure Travis gives it enough time to build the package?
2018-04-20 20:41:32 <clandmeter> nope
2018-04-20 20:41:51 <clandmeter> but you can leave a msg so dev can try it local.
2018-04-20 20:43:42 <mksully22> clandmeter, K Thanks
2018-04-20 21:37:43 <bratchley> how do I ensure the files my .apk package installs had the same UID during build as will need to be the case on the final system?
2018-04-20 21:38:47 <bratchley> for example, I want the app's /etc directory to be owned by the daemon user and the user is successfully added in post-install but with a different UID than the system I build my .apk on because chronyd is using that UID
2018-04-20 21:47:08 <bratchley> Do I just try to find a UID that doesn't appear to be used in any of the aports APKBUILD dirs and supply that to the adduser command in my pre-install script?
2018-04-20 21:51:34 <bratchley> if so how do I make sure pkgusers also has that uid? just chown to the UID in the build process?
2018-04-20 21:52:49 <bratchley> er I mean in my "pre-install" not post-install
2018-04-20 22:23:19 <terra> I was trying to make lightdm launch propely. But failed.
2018-04-20 22:23:48 <terra> I updated lightdm package for latest version along with all patches & fixes. I also created the required port: lightdm-gtk-greeter.
2018-04-20 22:25:05 <terra> But lightdm still refuses to work. I tested other DM's on repo: slim and lxdm then realized that none of them working.
2018-04-20 22:25:43 <awilfox> bratchley: it preserves ownership in the apk file
2018-04-20 22:25:54 <terra> I think there's something wrong with my or Alpine's default X setup.
2018-04-20 22:26:11 <awilfox> bratchley: so if the files in the pkg dir are owned by 'food' user, then the foo daemon is who owns those files on the installed system
2018-04-20 22:26:26 <awilfox> bratchley: note that it uses /name/, not uid, so that it doesn't matter what uid it uses on the installed system
2018-04-20 22:42:49 <terra> Anyone can launch X with "root" ? Because I can't.
2018-04-20 22:45:16 <liwakura> i can
2018-04-20 22:45:49 <liwakura> alternatively, you could describe what you did, what you did expect, and what you got instead
2018-04-20 22:45:53 <terra> liwakura : are you on edge?
2018-04-20 22:46:04 <liwakura> yeah
2018-04-20 22:47:41 <liwakura> https://nero.github.io/2018/04/12/hints-on-getting-irc-support.html#too-generic
2018-04-20 22:47:56 <terra> liwakura : I was able to run xinit as root before. But realized that can't do anymore.
2018-04-20 22:48:37 <terra> echo "exec xterm" > .xinitrc ; xinit
2018-04-20 22:48:49 <terra> should work as root.
2018-04-20 22:50:19 <liwakura> i havent tested with xterm, but starting i3 works for me
2018-04-20 22:50:44 <terra> liwakura : as root ?
2018-04-20 22:51:00 <liwakura> yeah
2018-04-20 22:51:07 <liwakura> starting xterm also works for me
2018-04-20 22:52:49 <terra> you launching a new X session from console as root, right? I'm not talking about running apps as root.
2018-04-20 22:53:09 <liwakura> via startx
2018-04-20 22:53:19 <terra> ok thanks.
2018-04-20 22:58:17 <liwakura> terra: ah i see, i never called xinit
2018-04-20 22:59:10 <liwakura> maybe thats the culprit
2018-04-20 23:00:53 <liwakura> terra: do you still need help?
2018-04-20 23:43:19 <terra> duncan^ liwakura : I was struggling with X server.
2018-04-20 23:43:45 <terra> Realized that problem doesn't occur if I use fbdev instead radeon.
2018-04-20 23:44:48 <liwakura> it was rather difficult understanding your issue
2018-04-20 23:45:19 <terra> r600_dri.so cause segfault if I launch with root user. but operates normally with regular user.
2018-04-20 23:46:07 <terra> maybe upstream problem because void linux has same issue. but I use same kernel for Void.
2018-04-20 23:48:22 <liwakura> terra: for the next issue, take care that you are communicating 3 points: what you did, what you expected and what you got
2018-04-20 23:48:42 <liwakura> "XYZ does not work" leaves the reader to guessing
2018-04-20 23:49:57 <terra> I thought I was clear enough because I just need some feedback regarding repeatability of my problem.
2018-04-20 23:50:13 <liwakura> alright, if its only that
2018-04-20 23:50:19 <terra> I didn't request any help to solve my problem.
2018-04-20 23:51:34 <terra> thank you anyway. you're quite gentle.
2018-04-20 23:53:47 <liwakura> like, if i see someone having some issue, im genuinely interested in solving it
2018-04-20 23:55:44 <terra> me too. It helps to develop some skills.
2018-04-20 23:55:47 <liwakura> but if the error situation isn't elaborated, its rather frustrating
2018-04-20 23:55:59 <liwakura> how did you find out? Xorg log?
2018-04-20 23:56:13 <terra> gdb X
2018-04-20 23:56:45 <liwakura> after seeing the segfault?
2018-04-20 23:57:12 <terra> doesn't any segfault message to log
2018-04-20 23:57:23 <liwakura> mh..
2018-04-20 23:58:39 <terra> I started to update networkmanager to latest version but ended up with Xorg segfault problem.
2018-04-20 23:58:59 <liwakura> i try to avoid network manager
2018-04-21 00:00:10 <terra> Avoidable for you and me but I want to contribute on enhance on Alpine's desktop experience
2018-04-21 00:00:27 <terra> Because new users seek for easiness at first.
2018-04-21 00:01:41 <liwakura> mh..
2018-04-21 00:02:29 <terra> I was trying to figure out why I can't manage network manager with regular user. Thus try to launch session with a Display Manager instead of bare xinit.
2018-04-21 00:03:24 <liwakura> ah mh..
2018-04-21 00:04:06 <terra> I spend lots of time to update lightdm's aport then create a new greeter aport.
2018-04-21 00:05:03 <terra> Then this problem occur. "occured ?" sorry English is not my native language.
2018-04-21 00:07:24 <liwakura> occured, past tense
2018-04-21 00:07:44 <liwakura> i learned most of my english from reading imgur and shitposting in irc :P
2018-04-21 00:08:40 <terra> But every word doesn't accept "-ed" right. English is weird.
2018-04-21 00:09:34 <terra> You have to know which one gets past tense extension, which one not.
2018-04-21 04:02:25 <_ikke_> try dutch
2018-04-21 06:00:32 <danieli> english is my fourth language
2018-04-21 06:00:36 <danieli> can't really complain there
2018-04-21 11:27:30 <PureTryOut[m]> are subpackages not allowed to install init files to `/etc/init.d/`? I just get permission denied
2018-04-21 11:27:56 <liwakura> are you installing to / instead of $pkgdir ?
2018-04-21 11:37:57 <PureTryOut[m]> no, "$subpkgdir"/etc/init.d/
2018-04-21 17:33:20 <ollieparanoid[m]> PureTryOut: try to echo that line if you're still stuck there to see where it is really trying to write
2018-04-21 17:34:05 <ollieparanoid[m]> http://build.alpinelinux.org/ only displays 4 servers now, is that a bug?
2018-04-21 17:35:01 <PureTryOut[m]> it is trying to write to `/etc/init.d/`, which it's supposed to, but seemingly not allowed to
2018-04-21 17:35:16 <ollieparanoid[m]> what I mean is, maybe it doesn't replace the variable like it should for some reason
2018-04-21 17:35:19 <ollieparanoid[m]> typo etc
2018-04-21 17:37:17 <PureTryOut[m]> nope, it uses the correct location
2018-04-21 17:41:37 <_ikke_> ollieparanoid[m]: might be because msg.a.o has been restarted, so the build servers that are idle haven't posted a new message yet
2018-04-21 17:42:01 <_ikke_> (due to power outage + deffective ups)
2018-04-21 17:55:20 <ollieparanoid[m]> _ikke_: I see, thanks
2018-04-21 18:40:55 <PureTryOut[m]> ok update, it seems I had a typo and had `$subpkdir` instead of `$subpkgdir`. it's fixed now
2018-04-21 18:41:33 <PureTryOut[m]> however, my package is basically a meta package to build and install 2 subpackages. this means that the package itself has no `package()` function. abuild seems to complain here, saying it can't cd to `$pkgdir` (not that it needs to)
2018-04-21 18:41:55 <PureTryOut[m]> I worked around this by making a `package() { mkdir -p '$pkgdir' }` function, but that seems like an awful hack
2018-04-21 18:42:34 <PureTryOut[m]> so is it maybe better to just make 2 separate apkbuilds with basically the same content, just the parts for the subpackages switched? they use the same source code so it doesn't feel right to me
2018-04-21 18:54:57 <ollieparanoid[m]> I would go with one package, as it's easier to maintain/update
2018-04-21 19:09:40 <clandmeter> ollieparanoid[m], arm builders are down atm
2018-04-21 19:10:05 <clandmeter> aarch64 will return shortly
2018-04-21 19:10:36 <PureTryOut[m]> same. I guess I'll just submit it and see what comments I get
2018-04-21 19:14:24 <PureTryOut[m]> thing is though. `apk install <metapackage>` will not actually install any files. is that a problem?
2018-04-21 19:17:19 <ollieparanoid[m]> PureTryOut: if you want that, make the main package depend on the subpackages
2018-04-21 19:18:13 <PureTryOut[m]> ah, good call
2018-04-21 19:19:04 <PureTryOut[m]> btw, does a subpackage inherit all the depends of the main package?
2018-04-21 19:26:05 <PureTryOut[m]> I guess not, doesn't make sense to me.
2018-04-21 19:26:11 <PureTryOut[m]> different question, can a subpackage use .post-install files?
2018-04-21 19:28:11 <ollieparanoid[m]> I think it inherits them unless you override them, and post-install files probably work
2018-04-21 19:31:03 <PureTryOut[m]> hmm ok
2018-04-21 20:17:53 <PureTryOut[m]> what is the procedure for moving a package from testing to community? what's required for it?
2018-04-21 20:19:37 <_ikke_> PureTryOut[m]: the package must be up to standard
2018-04-21 20:19:40 <_ikke_> and tested
2018-04-21 20:21:08 <PureTryOut[m]> why does a package get accepted into the testing repository if it's not up to standard in the first place?
2018-04-21 20:22:31 <PureTryOut[m]> also, what does "tested" enhold? like have a few people run the package for a few days or something? is there a standard test plan?
2018-04-21 20:35:10 <_ikke_> PureTryOut[m]: there is no standard test plan
2018-04-21 20:35:20 <_ikke_> just confirmation that it works, rather than just that the package builds
2018-04-21 20:39:11 <PureTryOut[m]> oh, that sounds easy enough. I wonder why so many packages are still in testing if it's that eas
2018-04-21 20:39:30 <_ikke_> PureTryOut[m]: nobody bothers to pull them forwards
2018-04-21 20:39:43 <PureTryOut[m]> ah ok
2018-04-21 20:39:52 <_ikke_> or still known issues
2018-04-22 08:39:38 <tboerger[m]> clandmeter: now i patched the tests to get them working properly, but travis fails for another reason i don't know...
2018-04-22 10:20:19 <clandmeter> tboerger[m], ok, first i need to fix infra before i get back to prs
2018-04-22 10:20:58 <tboerger[m]> sure, just wanted to give an update... maybe somebody else can also give me a hint why the build fails?
2018-04-22 11:36:20 <duncan^> look at the other PRs
2018-04-22 11:36:30 <duncan^> they're mostly failing on the same point
2018-04-22 11:46:23 <andypost[m]> Travis builds are dead mostly for last week
2018-04-22 12:01:32 <_ikke_> "fatal: No names found, cannot describe anything."
2018-04-22 12:05:21 <andypost[m]> it means git fails but I see no reason
2018-04-22 12:05:38 <andypost[m]> looking at https://github.com/alpinelinux/aports/blob/master/.travis/build-pkgs
2018-04-22 12:06:06 <_ikke_> yea, was looking as well
2018-04-22 12:06:53 <_ikke_> like no tags were fetched
2018-04-22 12:17:16 <andypost[m]> 
ACTION trying https://github.com/alpinelinux/aports/pull/4079/commits/f275fb56a466de869024dd13dfba387fab6125e7

2018-04-22 12:48:17 <andypost[m]> https://github.com/alpinelinux/aports/pull/4079/commits/050707252c6be20d94e33e73c35843a27208b8f2 works, will cherry-pick asap
2018-04-22 12:54:14 <andypost[m]> @ji
2018-04-22 12:54:46 <andypost[m]> jirutka: are you agree to make this workaround? something really weird with travis
2018-04-22 12:56:10 <_ikke_> right, I was suspecting depth
2018-04-22 12:58:57 <_ikke_> But I wonder why it only just now causes issues
2018-04-22 12:59:07 <_ikke_> 50 commits is not a lot
2018-04-22 13:00:10 <andypost[m]> I did commit and rebased 2 PRs to see
2018-04-22 13:10:37 <clandmeter> andypost[m], if you disable dept it will clone aports completely each run?
2018-04-22 13:11:43 <andypost[m]> clandmeter: yes, and that's a main problem with this workaround but it's a matter of 10 seconds
2018-04-22 13:12:16 <_ikke_> with --depth, you are not getting any tags that are more deeper, so you cannot describe
2018-04-22 13:13:31 <clandmeter> and that exaplains the error we are getting?
2018-04-22 13:13:39 <andypost[m]> clandmeter: actually 20s according https://travis-ci.org/alpinelinux/aports/builds/369740723
2018-04-22 13:24:05 <clandmeter> _ikke_, what is doing git describe?
2018-04-22 13:30:17 <clandmeter> andypost[m], i dont think its dept thats causing the issue
2018-04-22 13:30:46 <clandmeter> abuild has been updated and has some changes in that direction.
2018-04-22 13:32:03 <andypost[m]> clandmeter:
2018-04-22 13:32:36 <andypost[m]> where I can find abuild changes that added git describe?
2018-04-22 13:34:53 <clandmeter> i just restarted hugo build on travis but it still fails.
2018-04-22 13:35:11 <clandmeter> andypost[m], https://github.com/alpinelinux/abuild
2018-04-22 13:36:41 <clandmeter> i dont have the energy to debug this now/
2018-04-22 13:45:51 <jirutka> andypost: what workaround?
2018-04-22 13:45:57 <jirutka> andypost[m]: ^
2018-04-22 13:47:17 <clandmeter> jirutka, check aports latest commit
2018-04-22 13:47:47 <jirutka> aha
2018-04-22 13:48:11 <clandmeter> My guess is abuild is not in aports dir
2018-04-22 13:50:29 <jirutka> "fatal: No names found, cannot describe anything." … I see this quite often, IIRC there’s pending PR fixing this in aports…?
2018-04-22 13:50:49 <jirutka> btw I didn’t read past 200+ messages, so I’m quite out of context now
2018-04-22 13:50:55 <andypost[m]> @clandmeter:alpinelinux.org: btw ffmpeg now works
2018-04-22 14:03:13 <clandmeter> andypost[m], https://travis-ci.org/alpinelinux/aports/builds/369565315
2018-04-22 14:03:18 <clandmeter> still fails
2018-04-22 14:05:03 <andypost[m]> clandmeter: I bet it because it was not rebased on fresh master
2018-04-22 14:06:10 <jirutka> what the heck did you changed that now “No names found, cannot describe anything.” is fatal?
2018-04-22 14:06:40 <jirutka> this is stupidity in abuild that assumes it always run inside git repository
2018-04-22 14:06:56 <clandmeter> there is a new abuild out
2018-04-22 14:07:02 <jirutka> this is here since I remember, but it hasn’t been causing abort until now
2018-04-22 14:07:06 <jirutka> aha, so somehow broke it
2018-04-22 14:07:06 <jirutka> great
2018-04-22 14:07:12 <clandmeter> with changes in that field
2018-04-22 14:07:32 <jirutka> then instead of hacking travis scripts, fix abuild
2018-04-22 14:08:29 <jirutka> it’s unnecessary for abuild to access git data at all on CI, so it should be at least possible to disable this feature
2018-04-22 14:09:31 <jirutka> aha, I forgot to do review :( https://github.com/alpinelinux/abuild/pull/32 so partially my fault
2018-04-22 14:10:31 <nmeum> yeah, an error seems to have slipped in there
2018-04-22 14:10:52 <nmeum> `git describe || true` was changed to `git describe`
2018-04-22 14:13:00 <clandmeter> looks like that issue yes
2018-04-22 14:13:24 <jirutka> yes
2018-04-22 14:14:24 <clandmeter> but im enjoying my beer, so dont mind me.
2018-04-22 14:15:12 <nmeum> cheer 🍻
2018-04-22 14:17:09 <jirutka> as someone already noted, `git describe` by default looks for tags, but `git clone` does not clone tags by default
2018-04-22 14:17:58 <nmeum> 
ACTION changes the git describe invocation back to `git describe || true`

2018-04-22 14:18:05 <jirutka> so it’s wrong to assume that `git describe` will always succeed
2018-04-22 14:18:33 <jirutka> please also add 2>/dev/null
2018-04-22 14:19:20 <nmeum> alright
2018-04-22 14:19:25 <clandmeter> and please revert the aports commit
2018-04-22 14:19:35 <clandmeter> saves a few seconds on travis
2018-04-22 14:19:41 <jirutka> yes
2018-04-22 14:20:22 <nmeum> which aports commit? the abuild upgrade?
2018-04-22 14:20:53 <jirutka> no, change in .travis.yml
2018-04-22 14:21:08 <nmeum> ah
2018-04-22 14:21:12 <nmeum> sure, will do
2018-04-22 14:21:45 <nmeum> pushed the fix to the abuild repo
2018-04-22 14:21:56 <nmeum> will backport it and revert the travis.yml change in a sec
2018-04-22 14:28:36 <nmeum> pushed
2018-04-22 14:28:43 <nmeum> travis ci should work as expected again
2018-04-22 14:29:48 <jirutka> +1 :)
2018-04-22 14:34:42 <andypost> nmeum, is it possible to make abuild to install packages build in the same travis run, or better split PR in 2?
2018-04-22 14:37:10 <jirutka> andypost: it should use packages built in the same run, if it does not do that, then it’s a bug
2018-04-22 14:40:07 <andypost> jirutka, it does not https://travis-ci.org/alpinelinux/aports/builds/369758643#L1506
2018-04-22 18:49:00 <danieli> is there no pkg-config in alpine?
2018-04-22 18:52:31 <danieli> oh nvm my bad
2018-04-22 18:52:41 <danieli> i was like "pkgconf should be the one i'm looking for"
2018-04-22 18:52:48 <danieli> i completely misread the stderr
2018-04-22 19:29:33 <_ikke_> clandmeter: not sure: https://travis-ci.org/alpinelinux/aports/builds/369585516#L783
2018-04-22 19:29:37 <_ikke_> looks like abuild
2018-04-22 19:30:41 <danieli> i'm making a polybar package in testing
2018-04-22 19:30:52 <danieli> should build all X extensions and split it up but i'll do that later on
2018-04-22 19:32:52 <clandmeter> _ikke_, ?
2018-04-22 19:33:21 <_ikke_> Do you see that message there?
2018-04-22 19:33:34 <_ikke_> clandmeter: was a response to your question what was doing a git describe
2018-04-22 19:34:50 <clandmeter> yes, its fixed already.
2018-04-22 19:36:19 <_ikke_> ok
2018-04-22 19:56:58 <terra> Question: In mesa and xorg-server, there's option for enabling libunwind but not used by Alpine.
2018-04-22 19:57:16 <terra> Do you foresaw any benefit to enable it?
2018-04-22 22:07:31 <duncan^> _ikke_: I thought the issue with No names found might be related to https://github.com/alpinelinux/abuild/commit/fd97176583ed7a13d4fd63ecbad3d436fe14d0b2 or  https://github.com/alpinelinux/abuild/commit/4501a6cc5eb5542c5172ae391ae54377b440e1bf
2018-04-22 22:08:00 <duncan^> (git=true supposedly fixes it)
2018-04-22 22:43:29 <terra> How can I define a particular patch file's level explicitly in APKBUILD?
2018-04-22 22:44:18 <terra> I mean something like this: source="fix1.patch:p0"
2018-04-23 00:55:26 <awilfox> terra: you don't, you have to make it -p1
2018-04-23 01:57:39 <terra> awilfox: I have get some big patches from upstream. But they're p0. That's why I asked.
2018-04-23 13:13:07 <terra> Question about PR's: Are devs prefer whole new patch for each fix request, or incrementing patches for each update in order to see changes?
2018-04-23 13:15:01 <Gottox> ncopa: what's the state of libressl-2.7 at alpine?
2018-04-23 13:16:05 <Gottox> currently doing the update and about 1/3 of the packages are broken. It's great fun :D
2018-04-23 13:22:13 <clandmeter> Gottox, we are already upgraded.
2018-04-23 13:22:50 <Gottox> wow
2018-04-23 13:22:55 <Gottox> :D
2018-04-23 13:26:46 <clandmeter> i think i bumped into it on mongodb, had to revert the previous libressl patch to make it build.
2018-04-23 13:27:43 <Gottox> I have todo that a lot.
2018-04-23 13:29:37 <Gottox> It's great that they've added 1.1 APIs, but this fucks up packages that have introduced their own compatibility layers.
2018-04-23 13:29:43 <clandmeter> i think some of them have already gone upstream, i guess we need to adjust them to target only < 2.7
2018-04-23 13:32:25 <Gottox> Didn't care about upstreaming the packages yet. A lot of useful patches come from openbsd atm.
2018-04-23 13:32:54 <chambbj> Could someone review https://github.com/alpinelinux/aports/pull/4075? (builds without issue and all tests pass when running abuild rootbld locally)
2018-04-23 13:33:00 <terra> clandmeter: Why my new commit on my aports fork appeared in my previous Pull Request?
2018-04-23 13:34:22 <terra> I PR'ed mame and just pushed my changes (snes9x) on my forked repo but it appeared as new patch on my mame PR.
2018-04-23 13:34:24 <terra> https://github.com/alpinelinux/aports/pull/4088
2018-04-23 13:34:58 <liwakura> terra: updating a branch also updates associated pull requests
2018-04-23 13:35:01 <clandmeter> terra, do you craete seperate branches?
2018-04-23 13:35:04 <liwakura> you need a new branch for each pull request
2018-04-23 13:35:20 <liwakura> no he didnt, both commits on master
2018-04-23 13:35:41 <clandmeter> never use master, follow gh procedure if possible please.
2018-04-23 13:48:38 <terra> clandmeter liwakura: Thanks for pointing out. That was my first PR on github. I was using email before.
2018-04-23 13:55:08 <terra> Is it ok to leave the PR as it is now?
2018-04-23 13:59:58 <clandmeter> terra, if the commits dont belong to each other move the commit to its own branch.
2018-04-23 14:00:28 <clandmeter> keep the original in master and move the new one to a new branch.
2018-04-23 14:00:56 <clandmeter> push it to github and it will ask you if oyu want to make a new pr for the new branch.
2018-04-23 14:01:06 <terra> clandmeter: will PR enty will update accordingly?
2018-04-23 14:01:27 <clandmeter> if you remove the commit from master and force push it will.
2018-04-23 14:01:44 <terra> thanks.
2018-04-23 14:10:45 <ncopa> Gottox: i think i needed ~40 patches for libressl 2.7
2018-04-23 14:10:50 <ncopa> some patches could be removed
2018-04-23 14:11:13 <Gottox> for main or the complete aports?
2018-04-23 14:12:45 <ncopa> dont remember
2018-04-23 14:12:52 <ncopa> i think main+community
2018-04-23 14:13:05 <ncopa> some patches made it upstream
2018-04-23 14:13:25 <ncopa> i think freebsd have done a great job to upstream lots of it
2018-04-23 14:14:43 <Gottox> I'm currently feeding mostly openbsd patches to void-packages.
2018-04-23 14:15:03 <ncopa> feel free to take from alpine too
2018-04-23 14:15:13 <Gottox> hehe, I do, I do :)
2018-04-23 14:16:24 <Gottox> It looks like we're the only distribution that ships qt-5.10 with musl and libressl though
2018-04-23 14:16:57 <ncopa> ah
2018-04-23 14:16:58 <ncopa> yes
2018-04-23 14:17:06 <ncopa> we didnt upgrade qt 5.10
2018-04-23 14:17:30 <ncopa> yet...
2018-04-23 14:17:58 <Gottox> feel free to take patches from void :P
2018-04-23 14:18:10 <ncopa> i also had some problem with chromium which seem to require clang/libc++
2018-04-23 14:18:25 <terra> ncopa: will you compile new qt-multimedia with gstream-1.0 series. Current one depends on 0.10 series.
2018-04-23 14:18:41 <ncopa> terra: good reminder
2018-04-23 14:19:43 <terra> Void has chromium 65 btw.
2018-04-23 14:20:15 <ncopa> ok, i need to look at that too then
2018-04-23 14:22:25 <terra> https://github.com/voidlinux/void-packages/tree/master/srcpkgs/chromium
2018-04-23 16:14:24 <Ganwell> On travis: ERROR: http://nl.alpinelinux.org/alpine/edge/main: network error (check Internet connection and firewall)
2018-04-23 19:37:10 <tmh1999> If I want to create/update a wiki page, who should I ping to ask for write access ?
2018-04-23 20:04:39 <clandmeter> tmh1999: you should have write access
2018-04-23 20:05:51 <tmh1999> clandmeter: interesting. Creating an account now.
2018-04-23 20:06:14 <clandmeter> there are some limits on url posting i believe
2018-04-23 20:06:20 <clandmeter> x hours delay
2018-04-23 20:06:49 <tmh1999> thanks, good to know
2018-04-23 21:28:58 <tmh1999> ncopa: are you still there ? I have 1 more patch for aports/scripts, then s390x is all good. About to push in a couple of min.
2018-04-23 21:29:19 <tmh1999> then we just need the server and good to go
2018-04-23 21:39:32 <tmh1999> done
2018-04-24 10:38:54 <ncopa> nice
2018-04-24 12:50:53 <fcolista2> cannot start firefox in edge..process got defunct immediatedly
2018-04-24 12:50:59 <fcolista2> firefox-esr and firefox
2018-04-24 12:51:02 <fcolista2> tried both
2018-04-24 12:51:08 <fcolista2> anyone has the same problem?
2018-04-24 13:19:13 <ncopa> starts for me
2018-04-24 13:19:16 <ncopa> which arch?
2018-04-24 13:19:24 <fcolista2> x86_64
2018-04-24 13:19:53 <fcolista2> even by running: firefox --new-instance
2018-04-24 13:20:14 <fcolista2> fcolista  7382  1.0  0.4 290876 35620 pts/3    S+   13:20   0:00 /usr/lib/firefox-52.7.2/firefox --new-instance
2018-04-24 13:20:14 <fcolista2> fcolista  7384  2.5  0.0      0     0 pts/3    Z+   13:20   0:00 [firefox] <defunct
2018-04-24 13:21:55 <ncopa> huh
2018-04-24 13:21:58 <ncopa> thats weird
2018-04-24 13:26:58 <azarus> the alpine linux logo on the wiki appears messed up for me, all clipped on the sides, anyone else seeing the same?
2018-04-24 13:35:40 <mps> fcolista2: works here with 'firefox --new-instance' but says that I should close already running one
2018-04-24 13:36:01 <fcolista2> mps, yes
2018-04-24 13:38:10 <mps> fcolista2: yes, x86_64
2018-04-24 13:38:24 <fcolista2> mps so you have the same issue?
2018-04-24 13:39:19 <mps> not sure, I can close this one with --new-instance, it is not defunct
2018-04-24 14:42:55 <chambbj> Could a maintainer take a look at https://github.com/alpinelinux/aports/pull/4075 when you get a chance? Thanks!
2018-04-24 18:51:10 <jirutka> automatic (unattended) updates (or should I say upgrades?) for apk: https://github.com/jirutka/apk-autoupdate (for early adopters, documentation is WIP)
2018-04-24 18:52:55 <jirutka> handling of packages_blacklist is currently hackish and fragile, it’s waiting for implementing apk upgrade --exclude
2018-04-24 18:53:47 <jirutka> C coders pls review https://github.com/jirutka/apk-autoupdate/blob/master/src/procs-need-restart.c, it’s my first standalone program written in C
2018-04-24 19:20:17 <jirutka> ^ waiting for `apk upgrade --exclude` to make apk-autoupdate more reliable and faster ;)
2018-04-24 19:20:24 <jirutka> ^ kaniini
2018-04-24 19:24:29 <Adran> jirutka: fancy
2018-04-24 19:24:33 <Adran> i've just been doing a crontab shells cript
2018-04-24 19:24:52 <jirutka> Adran: this is supposed to be executed from cron…
2018-04-24 19:26:15 <Adran> jirutka: I meant apk update && apk upgrade as my script, not something anywhere safe at all
2018-04-24 19:26:34 <jirutka> Adran: aha, so totally useless, b/c you don’t restart affected services ;)
2018-04-24 19:26:44 <jirutka> Adran: that’s the hard part of automatic updates
2018-04-24 19:27:02 <Adran> && reboot? ;)
2018-04-24 19:27:05 <jirutka> but necessary to make it more than just a placebo effect
2018-04-24 19:27:17 <Adran> which is fun and fine until alpine has a broken initramfs for a while and then you have a dead box
2018-04-24 19:27:23 <jirutka> uh, okay, if you can reboot your servers after each update…
2018-04-24 19:27:30 <jirutka> yes
2018-04-24 19:28:01 <jirutka> that’s the main reason why I’ve implemented support for blacklist, to exclude linux-* from automatic upgrades
2018-04-24 19:28:51 <jirutka> I never wanna upgrade kernel automatically and never automatically reboot machine
2018-04-24 19:29:21 <Adran> you need to live a little
2018-04-24 19:29:22 <Adran> :D
2018-04-24 19:29:37 <Adran> i agree though, and you have a much more sane way even if its brittle
2018-04-24 19:30:06 <jirutka> that’s the reason why I don’t wanna automatic kernel upgrades, so I can be sometimes offline without worrying about catastrophe caused by automatic upgrades XD
2018-04-24 19:32:15 <jirutka> the bad think is that I don’t remember in which way it’s brittle :/ I just remember that I came to conclusion that it must be implemented in apk-tools to make it fully functional and reliable, but I forgot the exact use case when this hack does not work well (b/c you know, don’t need to write it, I will _surely_ remember this… oh how many times I did that mistake :( )
2018-04-24 19:33:38 <jirutka> afk
2018-04-24 20:58:18 <tmh1999> haven't seen kaniini for a while
2018-04-24 21:01:37 <[[sroracle]]> they're busy with pleroma stuff
2018-04-24 21:03:05 <jirutka> pleroma? o.O
2018-04-24 21:22:08 <tmh1999> after the libressl/openssl discussion iirc
2018-04-24 21:24:19 <jirutka> hm, I probably missed that discussion
2018-04-24 21:24:35 <jirutka> what was the result? I thought that we’ve already agreed about dropping LibreSSL
2018-04-24 21:24:48 <jirutka> I’m really looking forward it
2018-04-24 21:34:42 <awilfox> as am I
2018-04-24 21:38:26 <tmh1999> kaniini mentions openssl development has been better recently and wanted to libressl but ncopa wanted to stay for a while to really look into the technical details before making any move. everyone was so busy since then.
2018-04-24 21:38:38 <tmh1999> jirutka: ^
2018-04-24 21:39:01 <tmh1999> *wanted to drop libressl
2018-04-24 21:43:52 <tmh1999> pleroma seems very cool
2018-04-24 21:51:32 <jirutka> aha :/
2018-04-24 22:14:43 <Adran> what is pleroma
2018-04-24 22:20:51 <[[sroracle]]> it's an activitypub implementation
2018-04-24 22:23:41 <jirutka> aha XD I looked it in dictionary and was quite confused… (in Gnosticism) the spiritual universe as the abode of God and of the totality of the divine powers and emanations. (in Christian theology) the totality or fullness of the Godhead that dwells in Christ.
2018-04-24 22:28:27 <tmh1999> I mean, you can just look at kaniini github page :D
2018-04-24 22:32:11 <tmh1999> should've mentioned it earlier though :D
2018-04-24 22:34:32 <jirutka> I took a breath that kaniini is not involved in some crazy religion bullshit XD
2018-04-24 22:34:53 <jirutka> (or we just don’t know about it)
2018-04-24 22:49:09 <tmh1999> lol
2018-04-24 22:49:14 <tmh1999> that hurts
2018-04-24 22:49:15 <tmh1999> lol
2018-04-24 22:55:12 <tmh1999> not a fan boy, but I find it interesting reading his blog posts.
2018-04-24 22:55:29 <jirutka> he has a blog?
2018-04-24 22:57:11 <jirutka> aha, http://kaniini.dereferenced.org/
2018-04-24 22:57:15 <jirutka> I have to add it to my RSS
2018-04-24 22:57:50 <tmh1999> yeah some cool stuffs
2018-04-24 23:54:48 <terra> Is there an APKBUILD option to reduce build log?
2018-04-24 23:55:51 <terra> ignore #channel * CRAP
2018-04-24 23:56:05 <terra> sorry
2018-04-24 23:56:24 <terra> jnt: you're right. As long as abuild doesn't depend on rpcgen, this shouldn't cause things.
2018-04-25 00:46:24 <jirutka> terra: no and if the problem is just that it does not finish on Travis, then write a comment that it fails due to log size, so we have to try it locally
2018-04-25 00:46:50 <jirutka> terra: definitely don’t redirect build logs to /dev/null just b/c of this
2018-04-25 00:51:03 <terra> jirutka : ok. thanks.
2018-04-25 05:56:06 <Gottox> I must say, that libressl update is really intense work for void, fixing packages for 3 days now :D
2018-04-25 10:16:17 <mps> whom I should contact regarding the this issue? https://bugs.alpinelinux.org/issues/8660
2018-04-25 10:44:45 <_ikke_> mps: best if possible is to just provide a PR or patch to the APKBUILD
2018-04-25 10:47:11 <mps> _ikke_: I know, but didn't set github account yet :( and the issue is just one line in APKBUILD. And, with enabled NLS tcsh works for me more than two months without any issue
2018-04-25 10:47:45 <liwakura> mps: you are the first person i know to use tcsh
2018-04-25 10:49:20 <mps> liwakura: well, nearly 30 years of habit and practice is hard to left :)
2018-04-25 10:50:42 <liwakura> >30 years of experience
2018-04-25 10:50:47 <liwakura> im not even 30 years old yet
2018-04-25 10:51:46 <mps> one day you will have these time behind you, and even longer I hope :)
2018-04-25 10:53:45 <mps> and, no, I'm not only person who use tcsh in AL. there are other people which uses it. and on other distros of course
2018-04-25 15:09:07 <terra> Guys, if a new package appears on repo with replaces="old-package", will "apk upgrade" detect and install/replace new one?
2018-04-25 15:17:39 <jirutka> terra: replaces has different semantics, it’s about replacing files owned by particular package, not packages
2018-04-25 15:17:49 <jirutka> terra: what you need is provides and it’s typically combined with replaces
2018-04-25 15:21:22 <terra> jirutka : replaces="" didn't work as expected. I will add provides="" too.
2018-04-25 15:22:03 <jirutka> https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#replaces and https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#provides
2018-04-25 15:22:19 <jirutka> if you’re renaming package, you typically need both
2018-04-25 15:26:11 <terra> bit still, it isn't visible to "apk upgrade". manual install required, apk removes old one itself.
2018-04-25 15:26:16 <terra> *but
2018-04-25 15:26:29 <terra> feature request?
2018-04-25 15:34:20 <jirutka> it is
2018-04-25 15:35:03 <jirutka> when you have package A installed, package A is renamed to B and it provides="A" and replaces="B", then when you run apk upgrade, it automatically purges A and installs B
2018-04-25 15:35:21 <jirutka> iff A doesn’t exist anymore, otherwise it ofc keeps A
2018-04-25 15:37:20 <terra> replaces="B" ?
2018-04-25 15:39:24 <terra> wayland package is maintaining wayland-egl libs afterwards 1.15. I'm trying to pass this lib from mesa-wayland-egl to wayland-libs-egl.
2018-04-25 15:39:38 <terra> with a transparent way.
2018-04-25 17:35:19 <clandmeter> Provides and replaces probably
2018-04-25 17:37:42 <wener_> can I static link with out ld-musl-x86_64.so.1 dependency ?
2018-04-25 17:38:28 <jirutka> wener_: ofc
2018-04-25 17:38:42 <wener_> jirutka: ?
2018-04-25 17:38:50 <azarus> (of course)
2018-04-25 17:39:30 <jirutka> wener_: but if you want fully statically linked binary without PIE, then you have to use option -no-pie
2018-04-25 17:40:14 <jirutka> otherwise you get static binary with PIE, which is still portable and everything, but it looks a bit unusual… a file command does not recognize it as static binary
2018-04-25 17:40:49 <wener_> add -no-pie still need libc.musl-x86_64.so.1
2018-04-25 17:42:53 <wener_> jirutka: how can i get a fully statically linked binary ? i also add  -l:/usr/lib/libc.a option
2018-04-25 17:43:00 <jirutka> what are you compiling and what options do you pass to compiler and linker?
2018-04-25 17:44:52 <wener_> same option for linker and compiler -static -static-libgcc -l:/usr/lib/libc.a -l:/usr/lib/libstdc++.a -no-pie
2018-04-25 17:45:00 <wener_> xmrig
2018-04-25 17:46:29 <jirutka> I’ve never seen -l:/path/to/static/lib
2018-04-25 17:46:57 <wener_> https://stackoverflow.com/a/4423324/1870054
2018-04-25 17:50:47 <andypost[m]> this upgrade needs backport asap https://github.com/alpinelinux/aports/pull/4118
2018-04-25 18:05:11 <terra> <jirutka> when you have package A installed, package A is renamed to B and it provides="A" and replaces="B", then when you run apk upgrade, it automatically purges A and installs B
2018-04-25 18:05:34 <jirutka> andypost[m]: yet another drupal CVE?!
2018-04-25 18:05:44 <jirutka> andypost[m]: maybe ppl should finally stop using these craps
2018-04-25 18:05:44 <terra> Is this typo? I didn't understand the logic with replaces="B"
2018-04-25 18:06:19 <jirutka> andypost[m]: I’m currently writing bug report for Nextcloud… it looks like typical problem with concurrency… so it’s screwed, b/c php “devs” don’t have any clue about concurrency
2018-04-25 18:06:39 <jirutka> andypost[m]: I’m finishing it and need to go home after it, so I’ll deploy it in an hour or so
2018-04-25 18:15:40 <terra> I have package A is installed. New B package will take its files and replace the existing A package. so B package will has provides="A" replaces="A"
2018-04-25 18:16:32 <terra> Right? I don't have to do anything on A side which never will get an update?
2018-04-25 18:20:38 <terra> jirutka said replaces="B" and this struggled me. B is the new package here.
2018-04-25 18:52:06 <terra> I'm just trying to figure out the things.
2018-04-25 18:52:23 <terra> Anyway, I think this the fix related with wayland-egl: https://paste.ubuntu.com/p/5KQFPbmYyZ/
2018-04-25 18:52:35 <terra> Thanks.
2018-04-25 19:05:40 <fcolista2> lxc stable turned in 3.0.0
2018-04-25 20:08:17 <clandmeter> fcolista2: yes i know
2018-04-25 20:08:21 <clandmeter> did you see the changelog
2018-04-25 20:12:45 <fcolista2> Yes...lot of things pulled out
2018-04-25 20:12:53 <fcolista2> templates, language bindings etc
2018-04-25 20:13:25 <fcolista2> gonna be a big change
2018-04-25 20:23:13 <clandmeter> yes, probably not suitable for 3.8.
2018-04-25 20:23:46 <clandmeter> ncopa: what is on our todo for 3.8?
2018-04-25 20:24:10 <ncopa> write release notes
2018-04-25 20:24:14 <ncopa> get builders up and run
2018-04-25 20:24:19 <ncopa> fix bugs
2018-04-25 20:24:26 <ncopa> upgrade qt to 5.10
2018-04-25 20:25:02 <terra> ncopa: chromium?
2018-04-25 20:25:08 <fcolista2> what's the ETA for 3.8?
2018-04-25 20:25:21 <ncopa> help with fix bugs and updates with be appreciated
2018-04-25 20:25:26 <ncopa> eg PR reviews
2018-04-25 20:25:31 <ncopa> i am working on build servers
2018-04-25 20:25:52 <ncopa> i was hoping beginning of May
2018-04-25 20:26:04 <ncopa> first week in May would be nice
2018-04-25 20:26:08 <ncopa> but i doubt
2018-04-25 20:26:15 <ncopa> yes, chromium should be updated as well
2018-04-25 20:26:26 <ncopa> i started with chroium but it is pretty time consuming
2018-04-25 20:26:42 <ncopa> need to get the builders bootstrapped and running
2018-04-25 20:27:27 <ncopa> clandmeter configured a new xgene server for armhf
2018-04-25 20:27:38 <ncopa> i am moving the containers over as we speak
2018-04-25 20:27:51 <ncopa> with significantly more ram and with ssd disks
2018-04-25 20:28:22 <clandmeter> i think i got a few mails regarding chromium
2018-04-25 20:28:31 <ncopa> ok?
2018-04-25 20:28:38 <clandmeter> ncopa: would be nice if you take over maintainership
2018-04-25 20:28:41 <ncopa> chromium needs clang nowdays
2018-04-25 20:28:44 <ncopa> yeah
2018-04-25 20:28:48 <clandmeter> so they start to bug you
2018-04-25 20:28:50 <clandmeter> :p
2018-04-25 20:28:53 <ncopa> i guess that would make sense :)
2018-04-25 20:29:12 <clandmeter> i think we dont package some of the stuff in lib dir
2018-04-25 20:29:35 <ncopa> possible
2018-04-25 20:30:07 <clandmeter> Hi Carlo,
2018-04-25 20:30:07 <clandmeter> It’s me again. I have problems getting chromium running on alpine. It says that it can not find the swiftshader stuff.
2018-04-25 20:30:08 <clandmeter> Here is the stack trace. Maybe you have an idea.
2018-04-25 20:30:43 <clandmeter> # /usr/bin/chromium-browser --headless --no-sandbox --disable-gpu --use-gl=osmesa http://www.google.de
2018-04-25 20:31:10 <clandmeter> Failed to load /usr/lib/chromium/swiftshader/libGLESv2.so: Error loading shared library /usr/lib/chromium/swiftshader/libGLESv2.so: No such file or directory
2018-04-25 20:32:01 <clandmeter> there are more so files not packaged. not sure what they are all needed for.
2018-04-25 20:42:50 <liwakura> jirutka: you've been following that issue?
2018-04-25 20:43:02 <liwakura> on alpine-chroot-install ?
2018-04-25 20:44:45 <liwakura> maybe you could implement a bail out if BIND_DIR == "/"
2018-04-25 20:44:58 <liwakura> that guy is difficult because he goes like, implement first, understand second
2018-04-25 20:58:26 <jirutka> ncopa: I definitely wanna upgrade nginx before v3.8, Andy is working on it, will check out how it goes
2018-04-25 20:58:57 <ncopa> good
2018-04-25 20:59:41 <jirutka> liwakura: what issue? chromium? definitely not, chrome/ium is forbidden on my systems :)
2018-04-25 20:59:58 <liwakura> jirutka: on alpine-chroot-script
2018-04-25 21:00:20 <liwakura> *install
2018-04-25 21:01:39 <jirutka> liwakura: I’m totally out of context, what issue are you talking about?
2018-04-25 21:01:46 <liwakura> https://github.com/alpinelinux/alpine-chroot-install/issues/8
2018-04-25 21:01:56 <liwakura> i somewhat assumed you get notifications
2018-04-25 21:02:42 <jirutka> hm, i did not, don’t know why :/
2018-04-25 21:04:50 <jirutka> it really hurts when someone is running this script in docker, when one of the reasons why i wrote it was to avoid docker…
2018-04-25 21:07:09 <ncopa> :D
2018-04-25 21:07:48 <jirutka> hm, `mount: / bound on /alpine` is really not good… I wonder how it can happen
2018-04-25 21:08:20 <jirutka> next time someone tell me using https://github.com/jirutka/alpine-make-rootfs inside docker and i’ll right shoot myself XD
2018-04-25 21:08:54 <liwakura> jirutka: script run with PWD = /
2018-04-25 21:09:00 <liwakura> BIND_DIR is $PWD per default
2018-04-25 21:09:19 <jirutka> liwakura: i figured this, but still it should not mount to / inside chroot
2018-04-25 21:09:19 <liwakura> fix would be to just bail out if BIND_DIR == "/"
2018-04-25 21:09:44 <liwakura> because it does not make sense for an chroot
2018-04-25 21:09:56 <jirutka> ahaa, i see now
2018-04-25 21:10:46 <jirutka> I don’t mount PWD into /mnt as in alpine-make-rootfs and IIRC alpine-make-vm-image
2018-04-25 21:11:09 <liwakura> i dont know how travis is called tho
2018-04-25 21:11:14 <jirutka> but into /, b/c I designed it for CI and wanted to have exactly the same PWD in and outside of chroot
2018-04-25 21:11:23 <jirutka> travis call it from /home/travis/something
2018-04-25 21:11:37 <jirutka> so it mounts /home/travis/something in chroot
2018-04-25 21:11:55 <jirutka> I didn’t expected that someone will start it from /, but that’s my fault that I didn’t add check for this
2018-04-25 21:12:11 <liwakura> just one wish, please dont do it as he suggestd
2018-04-25 21:12:28 <liwakura> grepping for diff output is horribly unportable
2018-04-25 21:12:51 <jirutka> eh, `if diff "$BIND_DIR" "$BIND_TARGET_DIR" | grep "Common"; then` … really?!
2018-04-25 21:13:02 <liwakura> jirutka: some story
2018-04-25 21:13:03 <jirutka> how is this supposed to work? o.O
2018-04-25 21:13:19 <liwakura> that guy wanted a docker container for sabotage linux
2018-04-25 21:13:30 <liwakura> and opened a ticket and asked in irc... and when it was done
2018-04-25 21:13:37 <liwakura> he was disappointed that its bigger than alpine
2018-04-25 21:13:48 <liwakura> and we had to explain him what "self-hosting" means
2018-04-25 21:13:49 <jirutka> heh, not surprised XD
2018-04-25 21:14:35 <liwakura> so, im not sure if he understands what he is doing
2018-04-25 21:14:45 <liwakura> or why he is trying to run that script inside of an docker container
2018-04-25 21:15:26 <liwakura> back to your question, jirutka
2018-04-25 21:15:47 <jirutka> docker users almost never understand what are they doing… disclaimer: my observation, your experience may vary
2018-04-25 21:15:49 <liwakura> some diffs output a description of what they see when its not a file
2018-04-25 21:16:18 <liwakura> the string happens if there are directories of the same name in both arguments to diff
2018-04-25 21:16:43 <jirutka> aha
2018-04-25 21:16:46 <liwakura> >Common subdirectories: abc/foo and def/foo
2018-04-25 21:16:48 <liwakura> like that
2018-04-25 21:17:07 <liwakura> jirutka: also, i think you are right with the observation on docker users
2018-04-25 21:19:57 <jirutka> btw just today’s story… my colleague came to me and asked if I can test something on busybox … to run `/bin/sh sh`… well… I tried it and ofc it doesn’t work, b/c sh expects path to script, not a command… he told me that it works elsewhere… so i tried it with /bin/sh on macOS (and was surprised that it’s bash, I thought… nevermind), then on bash, zsh… behaves exactly the same everywhere
2018-04-25 21:20:35 <liwakura> :<
2018-04-25 21:21:03 <liwakura> 
ACTION tries to stay a friendly person

2018-04-25 21:21:57 <jirutka> short story long, problem was some misunderstanding about docker’s entrypoint vs. command and some stupid image from dockerhub
2018-04-25 21:22:08 <jirutka> eh, long story short XD
2018-04-25 21:26:47 <jirutka> I’m thinking about how to fix it, the current behaviour is actually quite bad for every $PWD except /home/…
2018-04-25 21:28:11 <liwakura> make it explicit
2018-04-25 21:28:21 <liwakura> that the bind mount only happens if the option is explicitly specified
2018-04-25 21:28:44 <jirutka> that would break all existing usages in travis.yml
2018-04-25 21:29:50 <jirutka> maybe do bind by default only when PWD is inside /home
2018-04-25 21:30:17 <liwakura> what about inside $HOME ?
2018-04-25 21:30:36 <liwakura> i know excentrics who have their home directory directly in /
2018-04-25 21:31:59 <liwakura> anyways, im off to bed
2018-04-25 21:32:00 <liwakura> gn8
2018-04-25 21:32:25 <jirutka> this script must be run as root, so $HOME would be /root, that’s not what i want
2018-04-25 21:33:09 <jirutka> the problem is that binding $PWD into chroot as default was bad idea at the first place; it makes sense for Travis, but not in general
2018-04-25 21:35:58 <terra> Do you foresaw any benefit to enable libunwind on mesa?
2018-04-25 21:36:15 <terra> or xorg-server
2018-04-25 21:45:22 <jirutka> as I’m more thinking about it, binding $PWD to /$PWD in chroot by default was extremely bad idea, it’s quite dangerous…
2018-04-25 22:05:46 <terra> s33r : want to compile packages for yourself? so what can beat a 16 core/thread processor? 32 core processor.
2018-04-25 22:06:16 <terra> ^ upps wrong channel
2018-04-26 01:11:05 <tmh1999> say I have 2 openrc services, who use large amount of same code (variable + functions), so I made a third file that in each of these services at /etc/init.d/service_name, I add . /etc/init.d/base_file to load variables + functions.
2018-04-26 01:11:13 <tmh1999> It just doesn't work
2018-04-26 01:11:32 <tmh1999> looks like loading variables is ok, but loading function doesn't work
2018-04-26 01:12:01 <tmh1999> does ash allow loading functions from external (script) file ?
2018-04-26 01:31:04 <tmh1999> no shit, it's PEBKAC
2018-04-26 07:51:13 <Zajt> Hey! Is make installed by default on Alpine linux distros?
2018-04-26 07:51:24 <liwakura> no
2018-04-26 07:52:46 <Zajt> damn alright thanks
2018-04-26 10:23:27 <_ikke_> build-base is a meta package that should give you a core build system
2018-04-26 10:46:21 <jirutka> tmh1999: eh, if you need to extract some common scripts, definitely don’t put them to /etc/init.d, but somewhere into /usr/share/
2018-04-26 10:46:39 <jirutka> tmh1999: /etc/init.d/ should contain only and only runscripts (shebang openrc-run)
2018-04-26 10:46:49 <jirutka> tmh1999: btw what package…?
2018-04-26 10:55:19 <_ikke_> jk
2018-04-26 10:55:35 <_ikke_> ignore that
2018-04-26 11:27:09 <agowa338> Hello, just noticed, that the "." folder within the tar.gz has the permissions set to 700 instead of 755, is there a specific reason for it? It causes trouble with su and sudo for me.
2018-04-26 11:41:53 <ncopa> agowa338: hi, which tar.gz?
2018-04-26 11:43:41 <agowa338> I used the x86_64 "Mini Root Filesystem from https://alpinelinux.org/downloads/
2018-04-26 11:44:38 <agowa338> ncopa: ^^
2018-04-26 11:44:56 <ncopa> i dont think that was intentional
2018-04-26 11:46:23 <agowa338> Hoped so, took me half a day of debugging. At some pont i did "chmod -R 777 /" and then sudo suddenly worked. strace was not really helpfull for that as it just sayed "Access denied" on the requested binary.
2018-04-26 11:48:49 <liwakura> agowa338: never do chmod -R on /, please
2018-04-26 11:49:11 <jirutka> agowa338: you may find https://github.com/jirutka/alpine-make-rootfs useful
2018-04-26 11:49:59 <agowa338> liwakura: was only for debugging purposes.
2018-04-26 11:52:10 <agowa338> jirutka: I don't want to create my own root fs, if there is an official tar.gz already available. It is better for my current usecase. I'm back online in a view minutes, have to swich the network.
2018-04-26 11:52:57 <jirutka> agowa338: what are you gonna use with the downloaded rootfs…?
2018-04-26 11:53:11 <jirutka> s/use/do/g
2018-04-26 11:59:13 <agowa338> jirutka: Now I'm back online.
2018-04-26 12:08:26 <jirutka> agowa338: what are you gonna do with the downloaded rootfs…?
2018-04-26 12:10:01 <agowa338> Thats a long storry and I may wont make any frends here. Do you really want to know?
2018-04-26 12:10:56 <jirutka> agowa338: hm, if it involves docker, then I’d rather don’t know; otherwise tell me, I’m curious person :)
2018-04-26 12:11:43 <agowa338> no it does not involve docker, but after I told you, you would like it was docker :p
2018-04-26 12:12:16 <jirutka> what can be worse?
2018-04-26 12:12:25 <jirutka> Windows Subsystem for Linux?
2018-04-26 12:12:26 <agowa338> jirutka: First of all there is another operating system called windows and windows has a new feature called WSL "Windows Subsystem for Linux"
2018-04-26 12:12:49 <jirutka> heh, I thought that it will be WSL XD
2018-04-26 12:13:09 <agowa338> jirutka: If you want to have a look at my project: https://github.com/agowa338/WSL-DistroLauncher-Alpine
2018-04-26 12:13:38 <agowa338> The binary is up and running, but the store app is currently untested, but "should work"
2018-04-26 12:13:55 <jirutka> well, I’m actually a bit interested in this; I found some issues on WSL about adding Alpine as supported distro, but didn’t have time yet to look at it and asked Windows guys at work if they can help with that
2018-04-26 12:14:33 <agowa338> jirutka: Just compile it, it is now fully working (atleast from what I've tested).
2018-04-26 12:15:52 <jirutka> on one side I think it may be good for people who are forced to support Windows users, so they can prepare for them something based on same sane distro running on WSL, on the other side I’m afraid of Windows users coming to our support channels… but now we have docker users, so it probably can’t be much worse
2018-04-26 12:17:06 <agowa338> jirutka: I'm doing it like: 1. Download the tar.gz from the website 2. Generate a sha256 hash of the downloaded file and compare it with the one specified at compile time 3. throw it against the api to bootstrap the filesystem 4. "Invoke some bugfixing commands", like "chown 755 /" and "apk --no-cache add shadow" 5. ask for a username and add it to the system 6. add some necessary groups 7. tell the api we're done
2018-04-26 12:17:21 <jirutka> where can I find changes made for Alpine support? I don’t see it in the repository
2018-04-26 12:17:34 <agowa338> jirutka: currently I'm looking at verifying the gpg signature, bit it seams to be a pain to implement in c++
2018-04-26 12:18:13 <jirutka> IIRC rootfs from alpinelinux.org does not contain apk, doesn’t it?
2018-04-26 12:18:35 <agowa338> it does ;-)
2018-04-26 12:19:32 <jirutka> hm, but you still need to verify signature of rootfs to avoid hardcoding hash of specific release, right?
2018-04-26 12:20:49 <agowa338> jirutka: Yes, the goal is to dynamically get the latest version and download the sha256, tar.gz and asc. Than verify if the download was successfull by using the sha256 and after that check the integrity by using the asc and hardcoded pgp key.
2018-04-26 12:21:55 <agowa338> jirutka: But as i saied it is a pain to verify pgp signatures in pure c++, also windows does not have any pgp libraries by default...
2018-04-26 12:22:39 <agowa338> jirutka: and if I ship it with my launcher I would be responsible for updating it, that's something I wanto to avoid as much as possible.
2018-04-26 12:22:45 <jirutka> it’d be simpler to use the same approach I use in alpine-make-rootfs and other similar scripts
2018-04-26 12:24:34 <jirutka> I hardcode into the script URL and checksum of prebuilt statically linked apk-tools and Alpine keys, then use this static apk to install Alpine into a directory (apk verifies signatures of packages it downloads), so I get always the latest versions of all packages and apk itself
2018-04-26 12:25:40 <jirutka> it doesn’t matter when you use older apk-tools for bootstrapping, so it’s not needed to be updated regularly
2018-04-26 12:26:02 <agowa338> jirutka: The WSL is extracting the tar.gz for me, and the download happens over plain http, the alpinelinux.org guys don't have tls for it, so I'm not executing it before verifying its integrity.
2018-04-26 12:26:06 <jirutka> but I don’t know if this can work on WSL
2018-04-26 12:27:37 <jirutka> the problem is that you can’t use simple checksum verification for verifying minirootfs tar.gz, because of versions
2018-04-26 12:27:47 <agowa338> Invoking apk before bootstraping the root filesystem would not work, as I don't have a userland to execute apk in.
2018-04-26 12:28:47 <jirutka> there are many mirrors that support TLS, e.g. https://nl.alpinelinux.org/alpine/v3.7/releases/x86_64/alpine-minirootfs-3.7.0-x86_64.tar.gz
2018-04-26 12:30:09 <jirutka> however, TLS is quite weak, especially on Windows where people use various crapwars (called “antivirus software”) basically doing MitM attack on TLS connections, using fake certs etc.
2018-04-26 12:30:10 <agowa338> using https, would solve the man in the middle problem, so verifying the pgp signature would become less important. So only verifying the checksum would be enough.
2018-04-26 12:31:27 <agowa338> jirutka: AV is running with System privileges, so it is out of scope for me. Also on linux you could have imported a bad root ca. So it is not one of my but one of your local admins problem :p
2018-04-26 12:32:45 <jirutka> sane Linux users do not import random root CAs…
2018-04-26 12:33:09 <agowa338> jirutka: Sane Windows users also don't import random root CAs..
2018-04-26 12:33:27 <jirutka> no, they just use software that do it for them without them know…
2018-04-26 12:34:06 <agowa338> jirutka: But isn't that also true for Linux?
2018-04-26 12:34:08 <jirutka> anyway, if you can’t run apk somehow on WSL directly, then downloading minirootfs over TLS is probably the option
2018-04-26 12:34:25 <jirutka> no, at least I don’t know about any software for Linux doing this
2018-04-26 12:34:29 <agowa338> For me it looks like you're having the assumption "linux software does not install a root certificate silently"
2018-04-26 12:34:37 <jirutka> and definitely no such software is in any Linux distribution repository
2018-04-26 12:35:23 <agowa338> yes, because software doing this ships with a binary only installer that bypasses the package manager...
2018-04-26 12:35:33 <jirutka> isn’t there some problem with musl compatibility on WSL?
2018-04-26 12:35:40 <agowa338> musl?
2018-04-26 12:35:44 <jirutka> musl libc
2018-04-26 12:35:48 <jirutka> libc used in Alpine
2018-04-26 12:35:59 <agowa338> have an example command to check?
2018-04-26 12:36:28 <jirutka> well, if someone really runs some binary installed downloaded from internet on his/her system, then (s)he should be pwned…
2018-04-26 12:36:37 <agowa338> jirutka: Should say, that I'm on "Skip ahead" aka, bleading edge windows
2018-04-26 12:36:38 <fcolista2> jirutka, have one sec?
2018-04-26 12:36:45 <jirutka> fcolista2: yeah
2018-04-26 12:37:07 <fcolista2> about lxc
2018-04-26 12:37:15 <jirutka> agowa338: well, there’s no easy way how to check it, just try to install and run some packages and you will see if it works
2018-04-26 12:37:35 <fcolista2> https://dpaste.de/r1GR/raw
2018-04-26 12:38:19 <fcolista2> in this way i don't get any error on init
2018-04-26 12:38:29 <fcolista2> I've already pushed this and it works fine
2018-04-26 12:38:30 <jirutka> agowa338: there’s the one Alpine related issues on WSL: https://github.com/Microsoft/WSL/issues/2372
2018-04-26 12:38:36 <fcolista2> but I'm sure there's a better way to do that
2018-04-26 12:39:36 <jirutka> fcolista2: uh, it quite doesn’t make sense to run cut on awk output which supposed to parse the file…
2018-04-26 12:39:44 <jirutka> fcolista2: just modify awk accordingly
2018-04-26 12:39:55 <jirutka> fcolista2: I don’t know what they changed in the config format
2018-04-26 12:40:48 <fcolista2> output is dir:/var/lib/lxc/container/rootfs
2018-04-26 12:41:22 <jirutka> aha, they added prefix dir:
2018-04-26 12:41:33 <fcolista2> yes
2018-04-26 12:41:44 <fcolista2> this get rid of "dir:"
2018-04-26 12:43:03 <jirutka> fcolista2: `rootpath=$(lxc_get_var lxc.rootfs.path); rootpath="${rootpath#*:}"` would be cleaner
2018-04-26 12:43:30 <jirutka> fcolista2: hm, actually rootpath=${rootpath#dir:} would be better
2018-04-26 12:43:47 <jirutka> not sure what other prefixes may be there and if dir: prefix is mandatory
2018-04-26 12:43:53 <agowa338> jirutka: init is statically linked:
2018-04-26 12:43:54 <agowa338> # ldd /init
2018-04-26 12:43:54 <agowa338> ldd: /init: Not a valid dynamic program
2018-04-26 12:44:09 <jirutka> yeah, they already fixed this, that’s good
2018-04-26 12:44:23 <fcolista2> jirutka, when/if they change dir: in fancyprefix: this will break
2018-04-26 12:44:54 <fcolista2> but for now, yeah, that makes sense
2018-04-26 12:44:56 <agowa338> It was fixed before september 19 2017
2018-04-26 12:45:07 <jirutka> fcolista2: yes… and if they allow to use e.g. file:/path/to/image.tar, then the affected logic in init would break anyway…
2018-04-26 12:45:25 <jirutka> fcolista2: check the docs what prefixes are allowed there
2018-04-26 12:45:55 <fcolista2> k
2018-04-26 12:46:04 <fcolista2> perhaps wait for 3.0.0?
2018-04-26 12:46:20 <fcolista2> gotta check doc for 3.0.0 too..
2018-04-26 12:46:48 <jirutka> fcolista2: 3.0.0 has been already released
2018-04-26 12:46:51 <agowa338> If anyone knows something that "should/could" not work in WSL, please provide an example, I was not able to find any kernel related issues, I already crosscompiled the whole linux kernel for arm using wsl (but with debian userland)
2018-04-26 12:47:07 <agowa338> except loading kernel modles
2018-04-26 12:47:09 <fcolista2> yes, I mean wait to bring it in alpine
2018-04-26 12:51:04 <jirutka> agowa338: I don’t know about anything particular that may not work
2018-04-26 12:52:12 <jirutka> agowa338: how will it work? can Windows user choose what Alpine release (stable such as v3.7 or edge) want to install and what mirror to use?
2018-04-26 12:53:00 <agowa338> jirutka: I would have followed what debian and so do, just pull one version by default and drop the user into the userland. From there you can perfomr inplace up/downgrades as you like.
2018-04-26 12:53:27 <agowa338> Every user gets a root shell (that is running within the context of a normal windows user, no administrative rights required)
2018-04-26 12:53:36 <jirutka> agowa338: okay, so the latest stable (currently v3.7)?
2018-04-26 12:54:06 <agowa338> jirutka: currently, yes.
2018-04-26 12:55:54 <agowa338> jirutka: As WSL is targeted to developers and not endusers, what would you expect from such an developer environment? Running some nodejs stuff?
2018-04-26 12:56:50 <agowa338> or building Alpine packages? Think I'll check the later one...
2018-04-26 12:56:55 <jirutka> agowa338: I dunno tbh
2018-04-26 13:00:12 <agowa338> jirutka: Your skript alpine-make-rootfs runs on Alpine I assume? Are there any other dependencies than noted in the Readme?
2018-04-26 13:00:42 <jirutka> agowa338: yes, but not just on Alpine, it should work on any Linux system
2018-04-26 13:01:12 <agowa338> jirutka: just wanted to make sure, so now I'm testing it on WSL Alpine ;-)
2018-04-26 13:01:51 <jirutka> agowa338: I’m quite curious if https://github.com/alpinelinux/alpine-make-vm-image works on WSL, it uses nbd
2018-04-26 13:04:40 <agowa338> jirutka: Executing alpine-make-rootfs already exposed the first error, windows does not have a tty "-ash: can't set tty process group (Operation not permitted)"
2018-04-26 13:05:51 <agowa338> but screen should worke for that, lmc
2018-04-26 13:07:11 <agowa338> ok it does not
2018-04-26 13:08:57 <jirutka> hm, tty is problematic even in some other environments
2018-04-26 13:09:11 <jirutka> but I don’t remember what I did with such problem
2018-04-26 13:10:38 <agowa338> ok, found the bug, it is something busybox wsl related "/bin/echo | sha1sum -c" throws the error...
2018-04-26 13:11:03 <jirutka> uh
2018-04-26 13:11:07 <agowa338> If I use bash it works
2018-04-26 13:11:07 <jirutka> that’s weird
2018-04-26 13:11:51 <jirutka> does ZSH works on it?
2018-04-26 13:12:22 <agowa338> I'm on a crappy wifi, installing takes a moment, ...
2018-04-26 13:14:18 <agowa338> zsh also works
2018-04-26 13:15:48 <jirutka> then i’d set zsh as default shell instead of ash, if ash doesn’t work on WSL
2018-04-26 13:16:30 <agowa338> jirutka: now the rally strange stuff, if I call "busybox echo '714a764b96442f551084fc50941fd1352e15ca6d  alpine-make-rootfs' | sha1sum -c" instead it also works.
2018-04-26 13:17:50 <agowa338> did busybox implement "echo" in two different ways?
2018-04-26 13:17:51 <jirutka> echo may be ash’s builtin
2018-04-26 13:18:19 <agowa338> but ls -la /bin/ash
2018-04-26 13:18:19 <agowa338> lrwxrwxrwx    1 root     root            12 Nov 30 22:50 /bin/ash -> /bin/busybox
2018-04-26 13:18:31 <jirutka> yes; I’m not sure about it, but most likely echo is builtin, so `busybox echo` runs different code than `echo` in ash
2018-04-26 13:20:46 <agowa338> also `ash -c "echo '714a764b96442f551084fc50941fd1352e15ca6d  alpine-make-rootfs'" | sha1sum -c` works...
2018-04-26 13:26:20 <agowa338> If someone wants to have a look here is the ash strace https://pastebin.com/uJMdNLHp
2018-04-26 13:39:08 <agowa338> and here is the bug report: https://github.com/Microsoft/WSL/issues/3129
2018-04-26 13:39:31 <kpcyrd> I filed a few PRs on aports, is there anything I can do to improve them? https://github.com/alpinelinux/aports/pulls/kpcyrd
2018-04-26 13:46:37 <agowa338> jirutka: Are you a Alpine maintainer? Aka. Are you the right person to speak to for releasing my build to the windows store?
2018-04-26 13:47:10 <jirutka> agowa338: yes, I’m one of Alpine core devs
2018-04-26 13:47:16 <agowa338> It requires some verification, currently I don't know how it is done.
2018-04-26 13:48:42 <agowa338> jirutka: I've to leave, I'm going to contact you again, when it's time for that ;-)
2018-04-26 13:48:53 <jirutka> agowa338: okay, thanks!
2018-04-26 14:39:04 <danieli> you want your build released as an official build?
2018-04-26 17:00:11 <liwakura> clandmeter1: feature request for the boot.alpinelinux.org images: virtio networking
2018-04-27 00:57:40 <jirutka> I wrote man pages for apk-autoupdate, please take a look if it’s readable and coherent https://github.com/jirutka/apk-autoupdate/tree/master/man (cc ncopa kaniini)
2018-04-27 08:39:08 <liwakura> jirutka: thats some documentation i like to read!
2018-04-27 08:39:17 <liwakura> also kudos for the mmap check thing
2018-04-27 17:34:35 <ncopa> tmh1999: looks like sudo testsuite fails. do you have time to have a look at it? looks like it fails to get correct process start time or something like that
2018-04-27 17:34:48 <ncopa> fails on s390x only
2018-04-27 17:35:24 <tmh1999> ncopa: I'm just chilling at the moment to wait for server.. I will try to reproduce it on the server I'm having (not container).
2018-04-27 17:39:15 <ncopa> its a blocker for getting the v3.8 build up and run
2018-04-27 18:03:52 <tmh1999> I see
2018-04-27 18:06:46 <tmh1999> ncopa: it passes on the native server
2018-04-27 18:07:48 <tmh1999> if people don't reply back early next week, I will call them
2018-04-27 18:08:31 <tmh1999> 
ACTION sounds like a boss lol 

2018-04-27 18:09:35 <tmh1999> but really.
2018-04-27 19:15:48 <danieli> it is friday my dudes
2018-04-27 19:16:48 <maxice8> yes
2018-04-27 19:34:59 <liwakura> danieli: aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
2018-04-27 19:35:05 <danieli> aaaAAAAAAAAAAAAAAaa
2018-04-27 19:36:33 <_ikke_> ..
2018-04-27 20:52:26 <duncan^> err
2018-04-27 22:03:05 <andypost> needs help to backport this commit and secfix to older releases https://github.com/alpinelinux/aports/pull/4118
2018-04-27 22:06:28 <andypost> and this one to 3.5 https://github.com/alpinelinux/aports/pull/4140
2018-04-28 07:04:21 <mikedld|w> hi there. where can I find the roadmap (if there is one) for upcoming release(s)? there's some link to redmine but web browsers complain about bad cert and don't open it
2018-04-28 07:06:23 <awilfox> https://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases maybe you want this?
2018-04-28 07:25:19 <mikedld|w> thanks but no. I'm looking for a schedule for the upcoming release(s). I see that each release takes a few months, but e.g. there's no indication for when 3.8 is planned to come out
2018-04-28 07:25:48 <awilfox> I'm not sure if that is in redmine either
2018-04-28 07:26:34 <mikedld|w> okay, thanks ;)
2018-04-28 15:01:34 <andypost> jirutka, please commit secfixes to older branches for php and drupal7
2018-04-28 15:31:23 <jirutka> andypost: done
2018-04-28 15:33:21 <liwakura> jirutka: nice!
2018-04-28 15:33:33 <liwakura> i always wonder, how does the build infrastructure look like?
2018-04-28 15:34:57 <jirutka> liwakura: hm, to be honest, the most accurate description of our current build infra is… something like a pile of shit :/
2018-04-28 15:35:17 <liwakura> like, a set of servers?
2018-04-28 15:35:27 <jirutka> liwakura: aha, servers, HW is fine
2018-04-28 15:35:47 <liwakura> alternatively i could setup sth on my own
2018-04-28 15:36:00 <jirutka> liwakura: what exactly do you wanna know?
2018-04-28 15:36:04 <liwakura> but thats only like, one additional repo url
2018-04-28 15:36:36 <liwakura> uhm, dont know what to ask for. i only see your commits on one side, and later, there is new stuff on the dl-cdn magically
2018-04-28 15:37:50 <liwakura> so is there some server polling git an then running abuild -r for every APKBUILD?
2018-04-28 15:38:52 <jirutka> basically, yes
2018-04-28 15:38:59 <jirutka> but it’s not just one server, but multiple servers
2018-04-28 15:39:09 <liwakura> one per arch?
2018-04-28 15:40:28 <jirutka> not exactly, some arches share one server (e.g. x86 and x86_64), but runs in different LXC containers
2018-04-28 15:40:40 <jirutka> andypost: do you know why this one test fails? https://travis-ci.org/alpinelinux/aports/builds/370351756#L3614
2018-04-28 15:41:00 <liwakura> one trouble point is how different builders are able to update the same repo
2018-04-28 15:41:21 <liwakura> (in terms of, i dont understand how that would be managed)
2018-04-28 15:43:19 <andypost> jirutka, thanks! btw that looks like some travis issue, as I reported it works locally & on sw-arm8
2018-04-28 15:43:25 <andypost> jirutka, https://github.com/alpinelinux/aports/pull/4066#issuecomment-383760193
2018-04-28 15:44:22 <andypost> I used to dig test but not found nothing wrong with it
2018-04-28 15:46:41 <jirutka> andypost: it fails on my machine too
2018-04-28 15:48:02 <jirutka> hm, I think that I know the cause
2018-04-28 15:48:17 <andypost> ipv6?!
2018-04-28 15:50:00 <jirutka> no, I don’t know, 127.0.0.2 works
2018-04-28 15:50:35 <jirutka> both builders and me use LXC
2018-04-28 15:53:46 <andypost> jirutka, any idea why s390 build is not listed at pkgs.a.o? http://build.alpinelinux.org/buildlogs/build-edge-s390x/community/php7/
2018-04-28 15:54:38 <jirutka> andypost: dunno
2018-04-28 15:55:53 <jirutka> # This test requires superuser privileges and CAP_NET_ADMIN.; rm "$srcdir"/nginx-tests-*/proxy_bind_transparent.t
2018-04-28 15:56:15 <jirutka> they have probably split proxy_bind_transparent to two tests
2018-04-28 15:56:28 <jirutka> the failing one is proxy_bind_transparent_capability
2018-04-28 15:56:36 <jirutka> and that needs to run under (real) root
2018-04-28 15:57:00 <andypost> it means lxc has no this cap?
2018-04-28 15:58:04 <jirutka> in the configuration we use, no, it’s disallowed
2018-04-28 16:57:34 <jirutka> andypost: njs module is tricky, it actually provides two modules
2018-04-28 17:51:27 <andypost> jirutka, that's where I stuck cos never used it
2018-04-28 17:51:43 <jirutka> andypost: I’ve already fixed it, now trying to run njs tests
2018-04-28 18:23:09 <andypost> jirutka, looks build went fine!
2018-04-28 18:23:21 <jirutka> yes
2018-04-28 19:54:01 <andypost> jirutka, looks php went fine in edge - please spread to other releases
2018-04-28 21:01:09 <jirutka> I still don’t understand why the github-pr-closer is not closing PRs made by awilfox :( I don’t see anything special in the webhook payloud
2018-04-28 21:03:27 <andypost[m]> jirutka: I guess because some builds are not finished yet
2018-04-28 21:03:43 <jirutka> nope, that’s irrelevant for it
2018-04-28 21:09:59 <[[sroracle]]> Is it gpg signing?
2018-04-28 21:10:52 <jirutka> that’s what I thought, but I don’t see any difference in webhook payloud
2018-04-29 00:26:14 <agowa338> jirutka: I just submited a request to the Windows Store to get the required permissions for accessing the required APIs. I'll get response in about 6 business days, hopefully they have nothing to complain. After that is done, I'll be able to publish the app for access by url, so that it is not yet visible in the store.
2018-04-29 11:45:38 <clandmeter> liwakura: can you try latest boot.a.o
2018-04-29 12:06:30 <liwakura> clandmeter: alright, i ran it, the virtio module is in the initrd
2018-04-29 12:06:43 <liwakura> im still investigating why it fails anyways
2018-04-29 12:24:30 <liwakura> clandmeter: it looks fine from your side
2018-04-29 12:24:59 <liwakura> the "virtio" feature from mkinitfs is lacking the virtio_pci module, which is a requirement for virtio disk and virtio network
2018-04-29 12:28:23 <liwakura> when using mkinitfs when doing an regular disk install, those modules get included still
2018-04-29 12:37:01 <terra> can't we include "service" command into openrc package again?
2018-04-29 15:27:50 <clandmeter> liwakura, what is the modules exact name?
2018-04-29 15:36:27 <liwakura> clandmeter: virtio_pci
2018-04-29 15:36:55 <clandmeter> liwakura, please try again
2018-04-29 15:37:01 <clandmeter> or wait 1 min
2018-04-29 15:37:03 <liwakura> what did you do?
2018-04-29 15:37:06 <liwakura> its not urgent
2018-04-29 15:37:17 <liwakura> IMHO it should be added to /etc/mkinitfs/features.d/virtio.modules
2018-04-29 15:37:21 <clandmeter> please try if you can
2018-04-29 15:37:46 <clandmeter> no thats full virtio support
2018-04-29 15:37:52 <clandmeter> we dont need that with netboot
2018-04-29 15:38:17 <liwakura> netboot needs virtio_pci, since the network goes via virtio pci
2018-04-29 15:38:22 <liwakura> for the modlook
2018-04-29 15:38:24 <liwakura> *modloop
2018-04-29 15:38:30 <clandmeter> no
2018-04-29 15:38:42 <clandmeter> modloop has nothing to do with it
2018-04-29 15:38:50 <clandmeter> we need it in initramfs
2018-04-29 15:38:59 <clandmeter> oh ok
2018-04-29 15:39:03 <clandmeter> i read it wrong
2018-04-29 15:39:08 <clandmeter> yes to load the modloop
2018-04-29 15:39:27 <clandmeter> but i dont want all of virtio in initramfs
2018-04-29 15:39:56 <clandmeter> we need network features to be inside initramfs
2018-04-29 15:40:01 <clandmeter> and it was lacking virtio
2018-04-29 15:40:26 <liwakura> boots with network working, here is lsmod: https://w1r3.net/KWdBJt.txt
2018-04-29 15:41:28 <clandmeter> ok so boot is ok now?
2018-04-29 15:43:13 <liwakura> checking the other usecase...
2018-04-29 15:43:49 <clandmeter> i wonder why virtio_net doesnt depend on virtio_pci
2018-04-29 15:46:45 <liwakura> yeah, the cloud stuff also works
2018-04-29 15:47:41 <liwakura> FYI one of my use-cases is one-shot starting an VPS and it ends up replacing itself with an alpine install
2018-04-29 15:47:50 <liwakura> optionally with an apkovl set up
2018-04-29 15:49:10 <liwakura> clandmeter: do you know how this module issue doen't happen if i do an regular install on a virtio machine?
2018-04-29 15:50:03 <clandmeter> liwakura, if you install you will generate your own initramfs
2018-04-29 15:50:11 <clandmeter> which will have all virtio included
2018-04-29 15:50:26 <liwakura> does mkinitfs detects that via lsmod?
2018-04-29 15:50:44 <clandmeter> it depends what iin mkinitfs.conf
2018-04-29 15:50:56 <clandmeter> maybe there is some autodetection
2018-04-29 15:51:53 <liwakura> mh, default mkinitfs.conf includes virtio
2018-04-29 15:52:06 <liwakura> but it ends up including virtio_pci
2018-04-29 16:00:16 <liwakura> clandmeter: what did you change in the generation scripts?
2018-04-29 16:11:22 <clandmeter> liwakura, nothing
2018-04-29 23:53:17 <terra> Guys, which script is both for new install and upgrade? $pkgname-pre-install?
2018-04-30 04:49:07 <tmh1999> jirutka: mind if I upgrade openblas 0.2.20. Looks like you're working on it a while ago and those issues have been resolved.
2018-04-30 13:22:50 <tmh1999> ncopa: I'm checking the linux-vanilla
2018-04-30 13:23:21 <tmh1999> there's some new  config for spectre
2018-04-30 14:47:20 <terra> How can I prevent to install runtime dependencies during package build?
2018-04-30 14:48:31 <terra> Due to this, I can't bootstrap some new package
2018-04-30 15:21:16 <terra> Ok, I figured it out. Just define depends locally in package()
2018-04-30 15:45:50 <fcolista2> so, we have gns3 broken in community due to aiohttp
2018-04-30 15:46:13 <fcolista2> I wonder to ship a py3-aiohttp-gns3 and set "conflicts" with py3-aiohttp
2018-04-30 15:46:23 <fcolista2> in order to use a different, compatible version
2018-04-30 15:46:49 <fcolista2> I should do the same with py3-aiohttp-*
2018-04-30 16:55:13 <fcolista2> any thought on that?
2018-04-30 17:08:31 <fcolista2> archlinux does the same
2018-04-30 17:09:21 <fcolista2> in aur they have a bounch of python packages with suffix -gns3
2018-04-30 17:12:27 <fcolista2> 6 pacakges
2018-04-30 17:12:35 <fcolista2> ncopa, what do you think?
2018-04-30 17:39:37 <fcolista2> jirutka, ^^
2018-04-30 19:02:59 <azarus> Do we have any package for wake on lan?
2018-04-30 19:03:05 <azarus> kinda missing wol(1)
2018-04-30 19:03:17 <azarus> I might just package it
2018-04-30 19:03:29 <duncan^> busybox has functionality to respond to magic packets
2018-04-30 19:03:54 <azarus> I mean, issuing wake on lan packets, not receiving.
2018-04-30 19:04:35 <azarus> ah, it's ether-wake
2018-04-30 19:04:52 <duncan^> ether-wake is to respond to them IIRC
2018-04-30 19:05:07 <duncan^> oh no
2018-04-30 19:05:12 <duncan^> yes ether-wake will send
2018-04-30 19:05:26 <azarus> ;)
2018-04-30 19:05:28 <duncan^> ether-wake was what I was thinking of, before
2018-04-30 19:06:56 <azarus> hmm, "ether-wake: socket(AF_PACKET,3,0): Operation not permitted
2018-04-30 19:07:17 <azarus> This shouldn't need root, right? :o
2018-04-30 19:07:47 <duncan^> Are you using grsec kernel?
2018-04-30 19:07:53 <azarus> no, linux-vanilla
2018-04-30 19:12:23 <azarus> Doesn't complain when executed as root, but it doesn't seem to work :/