teward | vorlon: that be in Groovy yes | 00:00 |
---|---|---|
teward | vorlon: do we need to wait for the apt maintainer first? | 00:00 |
teward | cjwatson: well it's failing multiple studio builds so it's on the radar for "this is broken hard" | 00:00 |
teward | as virtually all studio packages Eickmeyer's trying to get out of proposed and NEW dep on it | 00:01 |
teward | so it kind of is breaking Studio stuff :P | 00:01 |
teward | haven't noticed if there's any OTHER packages of mine affected (triggering PPA builds in Groovy of xca, vmfs6-tools, nginx "soon") | 00:01 |
teward | but...... | 00:02 |
teward | not sure if you should rollback until someone fixes it or not | 00:02 |
teward | ODDLY ENOUGH the fails are only in amd64 and risc, somehow the builds for mcpdist in other archs succeeded | 00:04 |
vorlon | teward: is this happening with the -proposed version of apt? | 00:18 |
vorlon | since apt hasn't updated in the release pocket since July 9 | 00:19 |
vorlon | if so, I think we should just bounce 2.1.9 out of -proposed | 00:19 |
vorlon | and let juliank look it over at his leisure | 00:19 |
vorlon | teward: amd64 and riscv64 have commonality in terms of which network the builders run on, vs the other architectures | 00:24 |
wgrant | teward, vorlon: Specifically, amd64 and riscv64 are talking to Apache, while the others are usually talking to Squid. | 00:27 |
cjwatson | vorlon: The reports I've seen have been with apt from groovy-proposed. I haven't done any kind of exhaustive analysis | 00:27 |
cjwatson | But I think rolling back is probably appropriate here | 00:27 |
wgrant | The changes in groovy-proposed look very very suspicious, and it's clearly only that version that is showing trouble. | 00:27 |
wgrant | So yeah, backing 2.1.9 out of groovy-proposed seems the right path. | 00:27 |
vorlon | ok, removing | 00:27 |
vorlon | juliank: ^^ apt 2.1.9 removed from proposed due to the regressions it causes for package builds | 00:27 |
cjwatson | Thanks. I can do some kind of bulk retry tomorrow if nobody beats me to it before then | 00:28 |
-queuebot:#ubuntu-release- New: accepted bslizr [amd64] (groovy-proposed) [1.2.8-0ubuntu1] | 00:29 | |
-queuebot:#ubuntu-release- New: accepted bslizr [armhf] (groovy-proposed) [1.2.8-0ubuntu1] | 00:29 | |
-queuebot:#ubuntu-release- New: accepted bslizr [s390x] (groovy-proposed) [1.2.8-0ubuntu1] | 00:29 | |
-queuebot:#ubuntu-release- New: accepted bslizr [arm64] (groovy-proposed) [1.2.8-0ubuntu1] | 00:29 | |
-queuebot:#ubuntu-release- New: accepted bslizr [ppc64el] (groovy-proposed) [1.2.8-0ubuntu1] | 00:29 | |
wgrant | I'm doing it. | 00:31 |
wgrant | Well, for the primary archive. | 00:31 |
cjwatson | Thanks | 00:31 |
teward | vorlon: all these packages are in proposed | 00:58 |
teward | so if groovy-proposed builds with -proposed enabled and available I'd say "yes" to it being apt in proposed | 00:59 |
teward | let me check | 00:59 |
teward | 'cause i just got spammed | 00:59 |
teward | apt 2.1.9 | 01:00 |
teward | ah good so you just yanked it | 01:01 |
teward | i'm slow :) | 01:01 |
teward | wgrant: so can i retrigger the mspdisp errors in Studio's builds then or is that on your radar too | 01:04 |
vorlon | teward: you wouldn't have to wait on him to do it, in any case; but I've just retriggered those fwiw | 01:08 |
-queuebot:#ubuntu-release- New binary: mcpdisp [amd64] (groovy-proposed/universe) [0.1.2-0ubuntu1] (no packageset) | 01:11 | |
-queuebot:#ubuntu-release- New binary: mcpdisp [riscv64] (groovy-proposed/universe) [0.1.2-0ubuntu1] (no packageset) | 01:24 | |
teward | vorlon: tyvm | 01:29 |
teward | is apt source tracked somewhere curious what changes were made that broke this so hard | 01:30 |
teward | esp. between 2.1.7 and 2.1.9 (where'd 2.1.8 go xD) | 01:30 |
teward | not a pro but i want to know heh | 01:30 |
juliank | vorlon: ooh failures. I need to reproduce | 06:04 |
juliank | Because I removed the code masking them, as it caused other bugs | 06:04 |
juliank | And now need to fix the bugs | 06:04 |
juliank | cjwatson: wgrant I might actually need shell access to an instance that can trigger the data left in buffer bug. Also am I hiding actual error message? I think I am! Because data left in buffer is a server side connection closure that was not handled properly | 06:09 |
juliank | I can't reproduce any bugs locally really, these are all very connection specific. | 06:09 |
juliank | Arguably they might depend on MTU, package size, and transfer speeds | 06:10 |
juliank | frustrating! | 07:10 |
-queuebot:#ubuntu-release- Unapproved: muffin (focal-proposed/universe) [4.4.3-1 => 4.4.3-1ubuntu0.1] (no packageset) | 09:44 | |
cjwatson | juliank: I don't have a way to give you shell access to any of our builders | 09:48 |
juliank | ack | 09:48 |
cjwatson | juliank: Hopefully it can be reproduced from something else in the Canonical DC that has direct access to ftpmaster.internal:80 (may not be required, but it's possible) | 09:48 |
juliank | cjwatson: The funny thing is I have another reproducer on debian, it fails like 10% | 09:56 |
juliank | cjwatson: The bug only happens if headers and content are in the same packet :/ | 09:56 |
juliank | At least the one I see | 09:56 |
juliank | I've made good progress | 09:56 |
Laney | juliank: Use autopkgtest's access to scalingstack if you want, that might be close enough | 10:01 |
juliank | That makes sense | 10:02 |
juliank | I also need a failing build | 10:02 |
cjwatson | That sounds promising. So effectively only if your internet connection is too fast? :) | 10:02 |
cjwatson | nautilus/groovy-proposed was failing I think | 10:02 |
cjwatson | It's been retried now, but you could try rebuilding the same thing | 10:03 |
juliank | yeah | 10:03 |
juliank | I bet there's a ton more bugs | 10:03 |
juliank | this fixed some https://salsa.debian.org/apt-team/apt/-/merge_requests/130/diffs?commit_id=d3ce6323d450c80bbcc8daee944271f1c0af9691 | 10:03 |
juliank | also this one https://salsa.debian.org/apt-team/apt/-/merge_requests/130/diffs?commit_id=4b910d12a51f812b802951cf011f5abeaf410299 | 10:03 |
juliank | I think I broke pipelining | 10:04 |
juliank | because I stopped exiting successfully on In.IsLimit() == true | 10:04 |
juliank | And this was all papered over before because apt silently retried | 10:04 |
juliank | with a brand new connection | 10:04 |
juliank | Still need to improve error handling a bit | 10:05 |
juliank | or rather, reporting | 10:05 |
seb128 | cjwatson, juliank , that was not source specific, random packages were failing to get some of the debs and different ones at each retry | 10:07 |
juliank | seb128: it is | 10:07 |
seb128 | so I think just apt build-dep for some non trivial package should be enough to trigger | 10:07 |
juliank | seb128: always specific | 10:07 |
juliank | it just does not appear so | 10:08 |
juliank | but it depends on the pipeline and response order, and length of packets and such | 10:08 |
seb128 | k, well it happened like on half the packages I uploaded/synced yesterday | 10:08 |
seb128 | I would say it's just random and if you download enough debs from the archive you would end up hitting it | 10:09 |
seb128 | I'm unsure how it could be specific to a source, the source code isn't involved at this point | 10:09 |
juliank | I think you currently don't see Error reading from server. Remote end closed connection, but I can't test it really, my connection is too fast | 10:28 |
LocutusOfBorg | vorlon, would it be possible to hint gfs2-utils on ppc64el that has regressed in release? https://autopkgtest.ubuntu.com/packages/g/gfs2-utils/groovy/ppc64el | 10:37 |
juliank | I'm fairly optimistic that this works now, I messed up the success recovery paths in the connection closure path | 11:51 |
juliank | I've also rewritten the entire thing to do the same in readable | 11:51 |
juliank | which meant doing code path analysis | 11:52 |
* juliank wonders if there's a tool for that | 11:52 | |
juliank | (like, tell it which boolean variables it is allowed to toggle and give me the outputs in a table) | 11:52 |
juliank | I can reproduce | 11:59 |
juliank | on autopkgtest | 11:59 |
Laney | ✅ | 12:01 |
juliank | I installed apt build-deps and I can no longer reproduce with my test, so I now download all of plasma-desktop | 12:07 |
juliank | because I need to download enough files for the server to close the connection on me | 12:07 |
juliank | All I can say is do not maintain 22 year old http clients | 12:11 |
juliank | Fix seems to work | 12:13 |
juliank | new apt landing, please stay attentive and ping me immediately if you see download errors https://launchpad.net/ubuntu/+source/apt/2.1.10 | 12:40 |
teward | juliank: so you fixed the 2.1.9 regression issue? they rolled back because it broke every build on select archs :P | 14:17 |
juliank | teward: yup | 14:18 |
teward | juliank: cool as long as it stops erroring heh | 14:39 |
-queuebot:#ubuntu-release- Unapproved: accepted apport [source] (focal-proposed) [2.20.11-0ubuntu27.8] | 15:27 | |
-queuebot:#ubuntu-release- Unapproved: rejected rpcbind [source] (xenial-proposed) [0.2.3-0.2ubuntu0.16.04.1] | 15:32 | |
-queuebot:#ubuntu-release- Unapproved: adobe-flashplugin (bionic-proposed/partner) [1:20200609.1-0ubuntu0.18.04.1 => 1:20200811.1-0ubuntu0.18.04.1] (no packageset) | 15:39 | |
-queuebot:#ubuntu-release- Unapproved: adobe-flashplugin (xenial-proposed/partner) [1:20200609.1-0ubuntu0.16.04.1 => 1:20200811.1-0ubuntu0.16.04.1] (no packageset) | 15:39 | |
-queuebot:#ubuntu-release- Unapproved: adobe-flashplugin (focal-proposed/partner) [1:20200609.1-0ubuntu0.20.04.1 => 1:20200811.1-0ubuntu0.20.04.1] (no packageset) | 15:39 | |
xnox | argh | 16:04 |
xnox | isn't it dead?! =) | 16:04 |
xnox | i thought 20202 was the year it will die | 16:05 |
-queuebot:#ubuntu-release- Unapproved: accepted adobe-flashplugin [source] (bionic-proposed) [1:20200811.1-0ubuntu0.18.04.1] | 16:05 | |
-queuebot:#ubuntu-release- Unapproved: accepted adobe-flashplugin [source] (focal-proposed) [1:20200811.1-0ubuntu0.20.04.1] | 16:05 | |
-queuebot:#ubuntu-release- Unapproved: accepted adobe-flashplugin [source] (xenial-proposed) [1:20200811.1-0ubuntu0.16.04.1] | 16:06 | |
Odd_Bloke | xnox: End of the year, I think? | 16:15 |
cjwatson | 20202> perhaps that typo is more true than you think | 16:18 |
xnox | hahahhahahahhahahhahahhahahahhaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaahahhhhhhhhhhhhhhhhhhhhhhahahahhahahahahahhahahhahaha | 16:21 |
patriciadomin | hello. I see 16.04.7rc ISO ready for testing on arm64 and amd64. Are you planning to build it for ppc64le and s390x? | 16:36 |
-queuebot:#ubuntu-release- New: accepted mcpdisp [amd64] (groovy-proposed) [0.1.2-0ubuntu1] | 16:38 | |
-queuebot:#ubuntu-release- New: accepted mcpdisp [armhf] (groovy-proposed) [0.1.2-0ubuntu1] | 16:38 | |
-queuebot:#ubuntu-release- New: accepted mcpdisp [riscv64] (groovy-proposed) [0.1.2-0ubuntu1] | 16:38 | |
-queuebot:#ubuntu-release- New: accepted mcpdisp [arm64] (groovy-proposed) [0.1.2-0ubuntu1] | 16:38 | |
-queuebot:#ubuntu-release- New: accepted mcpdisp [s390x] (groovy-proposed) [0.1.2-0ubuntu1] | 16:38 | |
-queuebot:#ubuntu-release- New: accepted mcpdisp [ppc64el] (groovy-proposed) [0.1.2-0ubuntu1] | 16:38 | |
vorlon | LocutusOfBorg: gfs2-utils/ppc64el hinted, thanks | 16:40 |
xnox | patriciadomin: no, we do not. | 16:41 |
xnox | patriciadomin: the BootHole does not affect ppc64le nor s390x. It only affect UEFI platforms. | 16:41 |
vorlon | cpaelzer: fyi after yesterday's conversation (and britney nagging me about a number of packages that I "uploaded" being stuck in -proposed too long ;), I am going ahead and rolling forward the various haskell packages that had been rolled back, after confirming they're not entangled with any current transitions. (e.g. the haskell-gi-* stuff can't be since it was all removed from release pocket) | 16:53 |
vorlon | WHY DOES REVERSE-DEPENDS OUTPUT TO STDERR | 16:55 |
patriciadomin | ok thanks xnox | 17:07 |
-queuebot:#ubuntu-release- New binary: libnfc [s390x] (groovy-proposed/universe) [1.8.0-1] (no packageset) | 17:24 | |
-queuebot:#ubuntu-release- New binary: aflplusplus [amd64] (groovy-proposed/universe) [2.66c-1] (no packageset) | 17:25 | |
-queuebot:#ubuntu-release- New binary: libzorpll [s390x] (groovy-proposed/universe) [7.0.4.0-1] (no packageset) | 17:25 | |
-queuebot:#ubuntu-release- New binary: libnfc [amd64] (groovy-proposed/universe) [1.8.0-1] (no packageset) | 17:25 | |
-queuebot:#ubuntu-release- New binary: libnfc [ppc64el] (groovy-proposed/universe) [1.8.0-1] (no packageset) | 17:27 | |
-queuebot:#ubuntu-release- New binary: libzorpll [amd64] (groovy-proposed/universe) [7.0.4.0-1] (no packageset) | 17:27 | |
-queuebot:#ubuntu-release- New binary: libzorpll [ppc64el] (groovy-proposed/universe) [7.0.4.0-1] (no packageset) | 17:28 | |
-queuebot:#ubuntu-release- New binary: wtdbg2 [amd64] (groovy-proposed/universe) [2.5-4] (no packageset) | 17:29 | |
-queuebot:#ubuntu-release- New binary: libnfc [armhf] (groovy-proposed/universe) [1.8.0-1] (no packageset) | 17:29 | |
-queuebot:#ubuntu-release- New binary: libnfc [arm64] (groovy-proposed/universe) [1.8.0-1] (no packageset) | 17:31 | |
-queuebot:#ubuntu-release- New binary: libzorpll [armhf] (groovy-proposed/universe) [7.0.4.0-1] (no packageset) | 17:31 | |
vorlon | not rolling forward haskell-ansi-terminal because pandoc. not rolling forward haskell-tasty / haskell-tasty-hunit, haskell-hledger-lib, haskell-hledger, because intertwined with haskell-ansi-terminal. not rolling forward haskell-quickcheck, because pandoc. | 17:31 |
vorlon | all the rest of my rollbacks have been rolled forward again | 17:31 |
-queuebot:#ubuntu-release- New binary: libzorpll [arm64] (groovy-proposed/universe) [7.0.4.0-1] (no packageset) | 17:32 | |
-queuebot:#ubuntu-release- New binary: libnfc [riscv64] (groovy-proposed/universe) [1.8.0-1] (no packageset) | 17:44 | |
-queuebot:#ubuntu-release- New binary: libzorpll [riscv64] (groovy-proposed/universe) [7.0.4.0-1] (no packageset) | 17:59 | |
-queuebot:#ubuntu-release- New: accepted libnfc [amd64] (groovy-proposed) [1.8.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libnfc [armhf] (groovy-proposed) [1.8.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libnfc [riscv64] (groovy-proposed) [1.8.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libzorpll [amd64] (groovy-proposed) [7.0.4.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libzorpll [armhf] (groovy-proposed) [7.0.4.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libzorpll [riscv64] (groovy-proposed) [7.0.4.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted wtdbg2 [amd64] (groovy-proposed) [2.5-4] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libnfc [arm64] (groovy-proposed) [1.8.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libnfc [s390x] (groovy-proposed) [1.8.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libzorpll [ppc64el] (groovy-proposed) [7.0.4.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libnfc [ppc64el] (groovy-proposed) [1.8.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libzorpll [s390x] (groovy-proposed) [7.0.4.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted libzorpll [arm64] (groovy-proposed) [7.0.4.0-1] | 18:17 | |
-queuebot:#ubuntu-release- New: accepted aflplusplus [amd64] (groovy-proposed) [2.66c-1] | 18:29 | |
-queuebot:#ubuntu-release- Unapproved: accepted umockdev [source] (focal-proposed) [0.14.1-1ubuntu0.1] | 18:31 | |
-queuebot:#ubuntu-release- Unapproved: accepted umockdev [source] (bionic-proposed) [0.11.1-1ubuntu0.1] | 18:34 | |
-queuebot:#ubuntu-release- Unapproved: accepted gnome-boxes [source] (focal-proposed) [3.36.5-0ubuntu2] | 18:38 | |
sergiodj | hello again, archive-admin :-). just a friendly ping to see if anyone can review telegraf, which is currently sitting at the NEW queue. TIA | 19:22 |
vorlon | node-sha.js - what incredibly terrible testsuite output | 20:39 |
vorlon | I'll just look for the "should be equal" test, shall I | 20:40 |
LocutusOfBorg | vorlon, ocaml-ctypes needs lwt on i386... can you please restore it? | 21:04 |
vorlon | why is anything ocaml part of the i386 set? | 21:08 |
vorlon | llvm-toolchain-10 build-depends on it? :P | 21:08 |
vorlon | ok I'll get it in | 21:09 |
LocutusOfBorg | :) | 21:19 |
seb128 | vorlon, did you see https://irclogs.ubuntu.com/2020/08/07/%23ubuntu-devel.html#t12:21 ? (just mentioning because you said you were looking at the package earlier) | 21:33 |
vorlon | seb128: had not seen it, thanks for the link | 21:47 |
seb128 | vorlon, np! | 21:53 |
seb128 | I see that bryce also replied on the list with pointers to the upstream bug and that they commit a fix today | 21:56 |
seb128 | (Olivier is off this week also so don't wait on him) | 21:56 |
xnox | vorlon: cleaned up more udebs, from unused seeds, which components-missmatches looks at. Hopefully next run of components missmatches will allow more things to be demoted or removed. | 22:35 |
RikMills | vorlon: hi, would you consider skiptesting kio and kconfig in proposed to get most of the KDE things through? the libreoffice armhf mess/fails are not their fault | 23:38 |
-queuebot:#ubuntu-release- Packageset: 55 entries have been added or removed | 23:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!