Patch Pilot Hand-off 24.04

Community
7
  • tp-smapi #2026589: added a comment, question, and marked as incomplete
  • flash-kernel #2037407: I’m not familiar with this package. Added some comments about the devel-series patch, and subscribed @waveform who has uploaded this in the past. I also added proper ubuntu release tasks to the bug, given the attached debdiffs.
  • xdg-desktop-portal-gtk #2039187: discussion still going on in the upstream PR. I added a comment to the LP bug and unsubscribed sponsors, asking to subscribe it again once a decision has been reached.
  • wget merge #2039429: sponsored
  • iproute2 #2039280: useful patches for iproute2, which in ubuntu is behind debian. I agree a merge would be better, but I don’t know how long that would take, given we have been behind since at least lunar. I asked for information and some changes.
  • bluez #2036149: just unsubscribed sponsors, since it was uploaded yesterday
  • transmission #1973084: memleak in transmission in jammy and lunar. The reports claim that the attached patch fixes them, so I prepared a PPA for lunar and jammy and asked for testing. If all goes well, this is definitely worth an SRU.
2 Likes
8

Notable achievements

  • Reviewed GDB MP, left comments asking for small adjustments, but the bulk of the work went into reserving an ARM64 machine and testing the fix before sponsoring. Update: getting an ARM64 machine proved to be basically impossible, so I decided to go ahead and upload the package after having reviewed and agreed with the changes.
  • Reviewed and sponsored gnutls MP to disable TLS 1.0/1.1 via config file. Left comments suggesting better practices next time.
  • Reviewed tmux MP, left comments asking for more information about a specific thing, but otherwise the MP is ready to be sponsored, which I will do once @mdzurick provides the request info.
3 Likes
9

Notable Achievements:

  • LP: #2040352 [openjdk-8]: sync’d
  • LP: #2041724 [python-linecache2]: sync’d

Important Context

  • mp:e2fsprogs/454796 [e2fsprogs]: Left some review comments, this sort of feels like two fixes in one, and should maybe have two SRU bug reports.
  • LP: #2039460 [network-manager-applet]: Needs targeted to noble and bug ref added. Unsubbed sponsors for now.
  • LP: #2039873 lxc: Waiting for debdiff, requested by Robie. Unsubbed sponsors for now.

Questions

  • There are a number of [needs-packaging] entries in the queue, that are adding new packages to noble, presumably for HWE/kernel needs? Are these intended to be included in the patch pilot effort? If so, is there a documented process for how to handle them?

  • mp:secureboot-db/454748: This lacks a d/changelog entry, which I almost commented about, but then I’m not certain if this is actually intended to be a sponsorship request, or if Ubuntu Sponsors got auto-added erroneously? If the latter is the case, then it may be worthwhile to add an entry to the git-ubuntu patch pilot document[1], such as:

      "How do I remove the 'Ubuntu Sponsors' reviewer from a git-ubuntu MP?"

    Is there a best practice established for this by now? (I know a workaround is to claim the slot and then close it as Abstain. If there is not a better way, could/should we document that?)

1: https://canonical-git-ubuntu.readthedocs-hosted.com/en/latest/howto/patch-pilot.html

1 Like
10

I was on shift yesterday. Since we are at an internal event at Canonical, I didn’t do any active sponsoring work, but was still available on IRC, where no one came over to ask me questions.

1 Like
11

@tsimonq2 took care of some sync requests in the sponsoring queue (thanks!).

In this shift I mostly did reviews and mentoring.

Important context

2 Likes
12
  • LP: #1508290 - uploaded with block-proposed-jammy
  • LP: #2043494 - processed sync request
  • LP: #1097467 - uploaded, thanks @waveform80 for the attention on bugs that have been open for a very long time. (this also cleans up LP: #589496 !)
  • LP: #2036873 - reviewed, nearly convinced, start discussion on plan for the test suite

Nothing to report from IRC

1 Like
13
  • alacarte LP: #2037326 fix was auto-sync’d from Debian, asked in the bug about SRUs, unsubscribed Ubuntu Sponsors
  • evince LP: #1794064 reviewed and sponsored SRU for Jammy
  • gdb LP: #2040113 reviewed and sponsored SRU for Lunar

The request for the gdb SRU came from irc. There was another request just after I ended my shift for an openssl SRU, but it needs more than just a quick look.

1 Like
14

Notable achievements

  • MP #455564 and MP #455517: apache2 MPs for @mdzurick. Left comments asking for improvements.

  • Bug #1991909: the user addressed the comments made by @paride, so I uploaded the package and staged the SRU as per @ahasenack’s instructions.

  • MP #455006: curl merge from @danilogondolfo. Left minor comments asking for improvements. He addressed them, and I uploaded the package.

  • MP #454972: netcat-openbsd merge from @danilogondolfo. Left a question. Also explained more about how dpkg-buildpackage scans changelog entries to automatically close bugs.

  • Bug #2008594: getmail6 fix. Uploaded after doing a small adjustment to the provided debdiff.

  • Bug #2038834: fix for mesa. The Noble upload was done, but there’s no debdiff for Mantic. I left a comment asking for it.

General comments

  • I think we should unsubscribe ~ubuntu-sponsors from bugs whose debdiffs/packages have been uploaded. If there are any subsequent problems with the upload, the reporter can always flag us again. I had a brief conversation with @tsimonq2 and @ahasenack on IRC about it.
1 Like
15
  • #2026589: left a comment asking how this is supposed to be tested, with a suggestion.
  • #2039280: no update since my last shift on Oct 27th, unsubscribing sponsors.
  • #2039429: stuck in noble migration, due to armhf dep8 infrastructure problems. Since I sponsored this upload, I’ll keep an eye on it. For now, unsubscribing sponsors from the bug, as there is nothing else to sponsor here.
  • #1973084: got confirmation the ppa fixes the problem for some user(s). I prepared the SRU template and sponsored the uploads.
  • #2042059, #2042094: unsubscribed sponsors, as it was uploaded already
  • #2042394: marked as incomplete, added other ubuntu release tasks. Patch is just for jammy, others needed.
  • #2029930: wget crash when calculated download speeds are in the Tb/s range. Prepared a ppa, asked for confirmation on the fix for noble at least, and help with coming up with a test scenario for a possible SRU.
1 Like
16

I’ve been casually taking care of one or two a day. I haven’t been posting here because I don’t want to fill up the thread with my one or two each time, but feel free to mention me. :slight_smile:

Today I did these two:

I think, intuitively I take the LIFO approach to the queue, but in reality it should be FIFO. Would be nice to see some of those older items gone. :slight_smile:

5 Likes
17
  • LP: #2026589; left a comment requesting the updated test case first referenced by @ahasenack and unsubscribed ubuntu-sponsors until this is addressed
  • Reviewed the status of nemos-dev-key, optee-os-s32, u-boot-s32, and arm-trusted-firmware-s32
  • With nemos-dev-key now sponsored thanks to @tsimonq2 (LP: #2043448), optee-os-s32 was ready to go (LP: #2034648) so I sponsored that with minor changes
  • Reviewed u-boot-s32 (LP: #2034650) but I need to compare it to a fresh build of u-boot under noble and I’m now getting yet-another FTBFS due to cross-build compiler flags on that. I’ll dig into it again tomorrow, but ultimately I’m aiming to deal with this and the remaining -s32 packages.
2 Likes
18

I did more than my usual “one or two” packages today, so here’s a post.

According to the KPI metrics, the sponsorship queue was at 33, 24 hours ago. Right now, it sits at 24. Here are the uploads I personally did today:

  • Looks like kernelshark was accepted. Thanks for your help, @tjaalton!
  • Sponsored redshift-qt to both Debian and Ubuntu NEW queues for @arraybolt3. With my Lubuntu hat on, it’s a feature goal for 24.04 LTS and has a straightforward license, so it should be a pretty simple review.
  • Sponsored cryptsetup for @mkukri. We’re both going to keep a close eye on Britney, so we can address any potential issues with autopkgtests.
  • Some NemOS packages (I think all of these are?), which are:
  • openjdk was sponsored as well, I expressed concerns with the nearly 50 Lintian warnings, @vpa19771 promised me he’d look into them before the release. Not very happy about sponsoring a package that looks like that, but I’ll give them the benefit of the doubt on this one.

Other extremely important notes:

  • [19:28] (@eeickmeyer) kernelshark do do do do do do :shark:
  • Re: [Merge] ~mkukri/ubuntu/+source/cryptsetup:merge into ubuntu/+source/cryptsetup:debian/sid
    • Review: Approve
            .-""-.
           / .--. \
          / /    \ \
          | |    | |
          | |.-""-.|
         ///`.::::.`\
        ||| ::/  \:: ;
        ||; ::\__/:: ;
         \\\ '::::' /
          `=':-..-'`

Remember, community members can patch pilot too. If you’re in ~ubuntu-dev, take a seat in the cockpit! :airplane:

I’d also like to highlight @bdrung’s starter work on an ubuntu-sponsoring package, with some additional refinements to the code. I think the original package looks promising, but if anyone (~ubuntu-dev or not) is looking for some low-hanging fruit, I’m sure he’d appreciate the help with test cases, splitting data collection and rendering, etc.

Thanks.

2 Likes
19

I will say my singing was on-point.

3 Likes
20
1 Like
21

  • LP: #2007702: asked the reporter to fix the version numbers for the SRU candidates and pointed to the SRU policy.

  • LP: #2041171: removed ubuntu-sponsors, patch already uploaded.

  • LP: #2042902: asked the reporter to fix the version number for the mantic SRU candidates, pointed to the SRU policy, and asked them to adjust the trackers for each series.

2 Likes
22

  • LP: #2042902:

    • Continuation of the work of @paride 10 days ago and @athos-ribeiro 1 day ago.
    • I got an IRC ping as well even before finding it in the queue
    • Playing this as “please do X” ping pong would not complete anytime soon, but sadly the requester was not available for direct discussion. So as fallback I provided a case study and reasoning/material to study how it should be done to learn from.
    • uploaded to noble, once good there and updates to the SRU uploads have been made it can be sponsored there

  • Ping about zlib sponsoring: I didn’t have time to do all of it, but at least was able to spot and suggest some organizational details in the changelog that should help to not be in the way when the next sponsor comes by

2 Likes
23

I had the opportunity to do a little piloting today. Here are a few notes…

Uploads

Workflow Notes

  • Protip for sponsors: if you find yourself manually specifying -k, just set DEBSIGN_KEYID=<YOUR_KEYID_HERE> in ~/.devscripts - this will pick up your key every time.
  • Differences in quilt configuration usually are not blockers for sponsorship (in my opinion,) unless those differences result in incompatible patches. When sponsoring khmer for @ogayot, I noticed some differences. Given some further followup, I use some sane defaults from a guide, which removes that Index header. /etc/quilt.quiltrc has the default quiltrc, which enables the header and differences in the config. We should consider whether changing the defaults, or noting in the packaging guide, might be a good idea. Less diff lines are Usually better.
  • Canonical employees should note that sometimes, when using their @canonical.com email address, it creates a ghost Launchpad account. That karma belongs to you! Here’s an example that I discussed with @adrien today.

General Notes

  • Remember, community members can patch pilot too! If you’re in ~ubuntu-dev, take a seat in the cockpit! :airplane:
  • I’m seeing a lot of kernel-related and security-related sponsorship queue items. For example, four of the oldest items in the queue are security updates, and many of the newer items are kernel-related patches. Like some others, I tend to not mess with anything tangential to the kernel, and I don’t have the permissions to upload to the security pocket (despite being a proud member of MOTU SWAT.) If we could get representatives from both teams to knock out these items before the holidays, I’d greatly appreciate it. Past people I’ve pinged include @xnox and @alexmurray - perhaps either of you would be able to point us in the right direction.

“Extremely Important” Notes

1 Like
24

The sponsor queue report that the Patch Pilot team uses is the general report that excludes security and kernel uploads.

2 Likes
25

Makes sense, I didn’t notice that. Thanks.

That being said, KPI shows all items, and IMO the point still stands: red items should get appropriate attention.

1 Like
26

Well, we can fix the KPI to track the thing we’re actually trying to drive to zero!

The other items should get attention, yes. But kernel patches in particular are not really an Ubuntu community thing.

2 Likes