| === liushuyu1 is now known as liushuyu | ||
| tjaalton | he's off until jan | 06:19 |
|---|---|---|
| adrien | Hey, there have been many build failures due to an issue with gnupg2 during the week-end; I think it's now solved but dozens of builds probably failed. The common error message includes "devscripts : Depends: gnupg but it is not installable or". Is it possible to do large-scale retries of builds? Is there anyone to contact to do that? | 09:19 |
| adrien | To be clear, I'm not saying to do it right now, I'm just asking at the moment. However, could someone retry this build? https://launchpad.net/ubuntu/+source/vtk9/9.3.0+dfsg1-1.1build2/+build/29342140 We'll see if the issue is effectively solved. Thanks. :) | 09:20 |
| RikMills | adrien: retried | 09:22 |
| adrien | RikMills: thanks! | 09:22 |
| jbicha | athos: good morning. What are your plans for php version for Ubuntu 25.04? Debian just began their php 8.4 transition | 12:20 |
| athos | jbicha: I will follow Debian. I wish I could be more active in the transition and help unblocking things on their end (the plan was to perform the transition here right away as well). But I am currently working on other stuff ATM, so this is on hold. Cc jchittum | 12:36 |
| athos | jbicha: is anything blocking you ATM? | 12:36 |
| mateus-morais | Hi! Could somebody retry the failed builds for https://launchpad.net/ubuntu/+source/newt/0.52.24-4ubuntu1 please? The ones that failed have no buildlogs | 13:12 |
| sudip | mateus-morais: done | 13:18 |
| mateus-morais | sudip: thank you! | 13:18 |
| jbicha | athos: it's not blocking me, but Debian php-* packages are autosyncing and are unbuildable until we begin the transition | 13:25 |
| athos | ack | 13:58 |
| adrien | I'm trying to push the openmpi transition through and due to issues with gnupg2 during the week-end, a number of related packages have failed building on !amd64 ; according to the transition tracker, the first candidate is ruby-mpi: can someone retry the builds of https://launchpad.net/ubuntu/+source/ruby-mpi/0.3.2-4ubuntu2 ? Thanks | 14:21 |
| adrien | (and yes, I'll prepare my core dev application whenever I have some time) | 14:21 |
| adrien | note that the transition tracker lists openmpi at level 1 due to armhf and i386 but that's due to it dropping 32-bit support; as for genomicsdb, the build issue is unrelated to the one that occured during the week-end | 14:23 |
| ginggs | adrien: . | 14:23 |
| adrien | thanks! | 14:24 |
| ginggs | adrien: note that debian has ruby-mpi 0.4.0-1, can you check if it can be synced or merged? | 14:24 |
| adrien | sure | 14:26 |
| adrien | after some confusion, the conclusion is that we cannot: there's an added entry in d/rules that hardcodes "3.1.0" as the ruby version while we're with ruby 3.3 now | 14:50 |
| ginggs | adrien: thanks! | 14:52 |
| seb128 | @pilot in | 15:02 |
| === ChanServ changed the topic of #ubuntu-devel to: Archive: Open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Oracular | Patch Pilots: seb128 | ||
| adrien | and can someone retry the builds at https://launchpad.net/ubuntu/+source/hypre/2.32.0-2build1 ? the failures are related to libxml2 complaining about building and running with a different library version AFAIU (combined with the fact that some tools/steps/rules look at the presence of anything on stderr and assume failure) | 15:35 |
| adrien | hmmm, no, that's too recent: 5 hours ago; I doubt things have changed in the meantime, so I'm missing something | 15:35 |
| adrien | but we'll need to do something to ruby-mpi anyway because of https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1076063 which is fixed in 0.4.0-1 (FTBFS with mpich, so 32-bit arches) | 16:53 |
| -ubottu:#ubuntu-devel- Debian bug 1076063 in src:ruby-mpi "ruby-mpi: FTBFS with mpich as default MPI implementation on 32 bit architectures: mpi.c:64:13: error: ‘MPI_2COMPLEX’ undeclared (first use in this function); did you mean ‘MPI_COMPLEX’?" [Serious, Fixed] | 16:53 | |
| adrien | => tomorrow | 16:54 |
| ginggs | adrien: ruby-mpi was already removed on armhf in oracular, so it won't block this transition | 17:16 |
| sudip | rpuby-mpi is ftbfs on riscv64 | 17:17 |
| sudip | s/rpuby-mpi/ruby-mpi/ | 17:17 |
| ginggs | sudip: that looks the the gnupg2 problem adrien mentioned earlier | 17:20 |
| sudip | yeah. but isn't that fixed? | 17:20 |
| ginggs | i thought it was, but maybe something got another upload | 17:21 |
| sudip | retried again. the riscv64 build of gnupg2 finished just an hour before the last retry. | 17:28 |
| adrien | ginggs: oh, I wasn't aware; I had assumed that it was kept along since it could work with mpich, thanks (and thanks sudip for retrying the build too :) ) | 17:29 |
| sudip | the ftbfs of abyss is something else | 17:31 |
| ginggs | tsimonq2 dropped the delta when he sync'd abyss | 17:32 |
| sudip | I can add it back if you have not already started | 17:35 |
| adrien | yay, I spammed the php8.4 test on ppc64el using various triggers just so one would eventually pass; I feel slightly ashamed but the arch is problematic :'( | 17:35 |
| adrien | (hopefully libxml2 will migrate and we'll properly get rid of the version difference warning messages) | 17:36 |
| ginggs | sudip: i have not started, but i'm actually more interested why abyss builds in debian without the explicit -latomic | 18:07 |
| sudip | ginggs: might be because of '-Bsymbolic-functions' which is in Ubuntu builds but not in Debian | 18:15 |
| seb128 | @pilot out | 19:04 |
| === ChanServ changed the topic of #ubuntu-devel to: Archive: Open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Oracular | Patch Pilots: N/A | ||
| adrien | libxml2 is migrating; can someone retry builds in https://launchpad.net/ubuntu/+source/hypre/2.32.0-2build1 ? the architectures that built did so because the build-time tests are disabled there | 20:04 |
| tsimonq2 | sudip, ginggs: Happy to sponsor or fix something (abyss) if I broke it via syncpackage -f | 20:04 |
| adrien | s/is migrating/has migrated/ (I wrote that two hours ago and was waiting for it to actually be in the apt repository :) ) | 20:04 |
| tsimonq2 | adrien: looking | 20:05 |
| tsimonq2 | adrien: . :) | 20:05 |
| adrien | thanks! | 20:07 |
| sudip | tsimonq2: I have abyss ready to be uploaded with the delta, but just trying to find out first why its not failing in Debian. | 20:13 |
| tsimonq2 | sudip: I'll be around for a few hours+, no rush. | 20:34 |
| adrien | hypre failed building again and still has the same libxml2 warnings; I guess it's a dependency that built with another libxml2 version and that is actually consistent with libxml2-dev not being installed during the build so maybe it makes sense to find how it determines that version thing, download all the dependencies, and find out which ones are concerned | 20:43 |
| adrien | (tomorrow) | 20:43 |
| vpa1977 | adrien: I'll try to handle it after standup | 20:44 |
| adrien | vpa1977: thanks, I remember you opened bugs on that; note that I'm not 100% sure the warning is the cause of the build failure (but let me check something) | 20:45 |
| vpa1977 | https://bugs.launchpad.net/ubuntu/+source/libxml-libxml-perl/+bug/2089307 - just for the reference if we need to do any other no change rebuilds | 20:45 |
| -ubottu:#ubuntu-devel- Launchpad bug 2089307 in libxslt (Ubuntu) "llibxml-libxml-perl autopkgtest fails due to non-whitelisted output vs libxml 2.12.7+dfsg+really2.9.14-0.2build1" [Undecided, Fix Released] | 20:45 | |
| tsimonq2 | vpa1977: I don't remember off the top of my head if you have upload perms but I'll offer to sponsor/retry if you need it :) | 20:45 |
| tsimonq2 | adrien: ditto | 20:45 |
| vpa1977 | tsimonq2: Thank you =) | 20:46 |
| adrien | I just checked hypre's testsuite and it indeed checks if there are any non-empty ".err" files and uses that as a criteria to determine success or failure; which means that libxml2's warning is indeed going to make it consider that a failure | 20:50 |
| adrien | (but again, which one of its dependencies... lots of fun) | 20:50 |
| vpa1977 | I have uploaded ncr for hwloc, will retrigger hypre once it builds | 22:31 |
| tsimonq2 | Is http://sponsoring-reports.ubuntu.com/ down for everyone, or just me? | 22:44 |
| arraybolt3 | Am I the only one who can't sbuild things properly anymore since the latest updates? Trying to build something with `sbuild -d noble-amd64-shm` results in `error: Unmet build dependencies: abc def ghi` (obviously those are placeholder package names). Why is sbuild looking on my host for build dependencies that belong in the schroot? | 22:45 |
| arraybolt3 | for avoidance of doubt, I'm on Plucky | 22:46 |
| arraybolt3 | or at least my build infra is | 22:46 |
| arraybolt3 | ahem, I'm being a bit vague I see. What I usually do when I build a package is: | 22:48 |
| arraybolt3 | * cd to the unpacked source package tree | 22:48 |
| arraybolt3 | * sbuild -d whatever-amd64-shm | 22:48 |
| arraybolt3 | * let sbuild make the source package for me and build it | 22:48 |
| tsimonq2 | (I usually do -c instead of -d so it doesn't set the actual distribution to noble-amd64-shm.) | 22:49 |
| arraybolt3 | But now I'm getting the error above when I do that. To work around it, I have to cd to the unpacked source package tree, run `debuild -S -d us -uc` to make the source package, and then `sbuild -d whatever-amd64-shm ../my-source-package.dsc`. | 22:49 |
| tsimonq2 | arraybolt3: Try purging your schroot and re-making it? | 22:49 |
| arraybolt3 | That's a lot of extra typing, cannot tolerate, fix plz kthxbye :P | 22:49 |
| tsimonq2 | XD | 22:50 |
| arraybolt3 | tsimonq2: it's not the chroot's fault, it's sbuild looking for build deps on my host. | 22:50 |
| arraybolt3 | If I use the aforementioned workaround, everything's peachy. | 22:50 |
| tsimonq2 | arraybolt3: show logs | 22:50 |
| * tsimonq2 runs `ssh arraybolt3 cat logs/*` | 22:50 | |
| arraybolt3 | please hold | 22:51 |
| arraybolt3 | and this is a local build silly :P | 22:51 |
| arraybolt3 | https://dpaste.com/AL9WKKYBD <-- behold, the logs | 22:52 |
| arraybolt3 | (this is for a package I'm building for my job, but this same thing has happened with other packages so this impacts my Ubuntu work too) | 22:52 |
| tsimonq2 | When you figure this one out, let me know. Sometimes there's checks done in an earlier stage of debuild which look for build dependencies locally. I *think* it's a packaging/rules thing | 22:52 |
| arraybolt3 | ugh, and don't use the workaround I use, it put the build output in the wrong directory and now my source tree and binary artifacts are nicely mixed together | 22:53 |
| tsimonq2 | XD | 22:53 |
| arraybolt3 | (put everything in a box and shake vigorously, why don't you sbuild?!) | 22:53 |
| arraybolt3 | anyway it wasn't doing this just the other day so I'm sure it's a recent change, probably worth a bug report | 22:54 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!