Status

Differences between revisions 1 and 115 (spanning 114 versions)
⇤ ← Revision 1 as of 2011-11-10 12:06:48 →
Size: 1120
Editor: 82-69-40-219
Comment:
← Revision 115 as of 2020-06-19 19:52:33 → ⇥
Size: 8711
Editor: ahasenack
Comment:
Deletions are marked like this. Additions are marked like this.
Line 3: Line 3:
== Daily build status == == Proposed Migration ==
Line 5: Line 5:
This is a scratchpad for known showstopper issues in current daily builds. This doesn't mean problems we would refuse to release Ubuntu with; this means problems that prevent anyone using the daily builds and which force the +1 maintenance team and/or the developer responsible for the breakage to make fixing this their number one task. This will amount to something like:  * https://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html
Line 7: Line 7:
 * build fails
 * principal metapackages (ubuntu-minimal, ubuntu-standard, ubuntu-desktop) uninstallable
 * automated tests can't run (and not a test bug)
 * can't get to a desktop or run the server		  * https://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt
Line 12: Line 9:
No known issues at this time. However, daily builds are currently suspended while we switch the master build system to a new server.  * Build failures: http://qa.ubuntuwire.com/ftbfs/
Line 14: Line 11:
== Other ==  * Last test rebuild (may be out of date): https://people.canonical.com/~doko/ftbfs-report/test-rebuild-20200327-focal-focal.html
Line 16: Line 13:
Other things the +1 maintenance team is working on or tracking. Please don't add things to this list unless you're on the team. == Known issues ==
Line 18: Line 15:
 * bzr/armel FTBFS (referred to jelmer, another bzr core developer is working on it)
 * sshfs uninstallable, breaking ltsp-client (correct fix is to merge fuse; cjwatson part-way through this)		 == In progress ==

Items that are being actively worked on. Please always mention the name / irc nick, who is working on that.

 * [rbalint] (needs review) retry-autopkgtest-regressions improvements (retry based on log regex, don't retry queued/running)
 [[https://code.launchpad.net/~rbalint/ubuntu-archive-tools/retry-intermittent/+merge/384468|review]]
 * [rbalint] (needs review) autopkgtest-cloud: retry on some frequent intermittent failures [[https://code.launchpad.net/~rbalint/autopkgtest-cloud/+git/autopkgtest-cloud/+merge/384609|review]]
 * [ahasenack]:
   * --(golang-testify/golang-go.uber-zap migration)--
      * --(golang-go.uber-zap migrated)--
      * --(retriggered golang-testify tests with fixed golang-go.uber-zap)--
      * --(uploaded [[https://launchpad.net/ubuntu/+source/golang-go.uber-zap/1.14.1-1ubuntu1|golang-go.uber-zap 1.14.1-1ubuntu1]])--
   * [[https://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#golang-yaml.v2|golang-yaml.v2]] is [[https://bugs.launchpad.net/ubuntu/+source/golang-yaml.v2/+bug/1883770|bug #1883770]] and --([[https://github.com/go-yaml/yaml/issues/629|issue #629]])-- upstream
    * this looks like a golang bug, as it does not happen in bionic or eoan (golang 1.10 and 1.12), just focal (1.13) and groovy (1.14). Upstream golang bug: https://github.com/golang/go/issues/39651, already assigned to someone from IBM
    * indeed, golang issue. Here is the proposed fix: https://go-review.googlesource.com/c/go/+/238628/
    * if we really need to get golang-yaml.v2 migrated, we can apply this emergency patch: https://bugs.launchpad.net/ubuntu/+source/golang-yaml.v2/+bug/1883770/comments/4
  * --(retried golang-github-xeipuuv-gojsonschema with the correct trigger for golang-github-opencontainers-image-spec/1.0.1-4. Already passed on s390x, should be green for the others soon.)--
  * --(Untangle golang-github-opencontainers-image-spec, golang-google-cloud, golang-google-api:)--
    * --(retried golang-google-cloud 0.56.0-1 with golang-google-genproto 0.0~git20200413.b5235f6-1 on amd64. Previous run was with golang-google-cloud/0.49.0-1)--
    * --(retried golang-google-cloud 0.56.0-1 with golang-google-api/0.21.0-1. Previous amd64 red run was with golang-google-cloud/0.49.0-1)--
  * --([[https://bugs.launchpad.net/ubuntu/+source/golang-github-fsouza-go-dockerclient/+bug/1883960|#1883960 FTBFS golang-github-fsouza-go-dockerclient/1.6.5-1]], blocking golang-logrus.)--
    * --(fixed via https://code.launchpad.net/~ahasenack/ubuntu/+source/golang-github-fsouza-go-dockerclient/+git/golang-github-fsouza-go-dockerclient/+merge/385957
    * DONE, it migrated (no idea why "strikethrough" wiki syntax isn't working here) Retriggered golang-logrus with the new golang-github-fsouza-go-dockerclient/1.6.5-1ubuntu1 for all arches
  * --(Looking at golang-github-go-ldap-ldap/2.5.1-4 armhf failure blocking migration of [[https://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#golang-gopkg-asn1-ber.v1|golang-gopkg-asn1-ber.v1]])--
    * --(This [[https://code.launchpad.net/~ahasenack/ubuntu/+source/golang-github-go-ldap-ldap/+git/golang-github-go-ldap-ldap/+merge/386025|MP]] has a fix, but we should be able to drop this package once syncthing is updated to at least 1.6.1, where it switched the dependency on golang-github-go-ldap-ldap to golang-gopkg-ldap.v3.)--
  * Looking at golang-github-miekg-dns, which is blocking golang-github-hashicorp-memberlist
    * integer overflow on armhf: https://github.com/miekg/dns/issues/1129, https://bugs.launchpad.net/ubuntu/+source/golang-github-miekg-dns/+bug/1884152
  * yara and golang-github-hiller-go-yara: looks like it just needs a trigger. Trying s390x first.
  * investigating golang-golang-x-tools next, version ...a31 never passed in Ubuntu nor in Debian
    * Couldn't fix it yet, but filed a bug with more details: https://bugs.launchpad.net/ubuntu/+source/golang-golang-x-tools/+bug/1884314 and https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=963166
 * [oSoMoN] rocksdb FTBFS ([[https://bugs.launchpad.net/ubuntu/+source/rocksdb/+bug/1884072|LP: #1884072]]), [[https://launchpadlibrarian.net/484839820/rocksdb.debdiff|debdiff]] attached to bug report and ready for sponsoring (but low priority because the new version in experimental builds fine)
 * [oSoMoN] dh-python autopkgtest failures ([[https://bugs.launchpad.net/ubuntu/+source/dh-python/+bug/1883297|LP: #1883297]], fixed with a combination of cherry-picking [[https://salsa.debian.org/python-team/tools/dh-python/-/commit/e289c25c861ae65ae9e7b52ebc09cf1f56061fa7|an upstream commit]] that hasn't been released yet, and disabling Python2 tests (as was done in version 4.20191017ubuntu4 in focal). Needs sponsoring ([[https://launchpadlibrarian.net/484997506/dh-python.debdiff|debdiff]] attached to the bug report)

== Things to pick from ==

These are issues that are not yet done, after somebody's rotation/shift ended.
Or someone has spotted them but had no chance to work on it.
If someone trying to pick them realized that it was resolved by time/someone-else please be so kind and remove the related entries.

 * Ongoing transitions
  * there are plenty of rust packages stuck in -proposed. Unfortunately the rust packages we sync from Debian are not generally in a consistent state allowing migration (to Debian testing or to the Ubuntu release). If someone takes a look at these, they will probably want to recommend removals of various leaf packages blocking transitions.
  * the r-base transition has now started
  * vcmi ftbfs, pinged xnox about it, boost 1.71 related
  * ubuntu-app-launcher, url-dispatcher, unity fallout
    Who to approach to remove those? xnox will look into them
  * rust-* some test and many "unsatisfiable depends", one needs to check how this was meant to work ...
    * the rust packages only build if uploaded in a very precise order, which they were to debian but the debian import freeze messes this up for us
    * it's not clear that any of these packages help users of Ubuntu at all, so putting effort into these packages has a low ROI IMO (I == mwhudson)
  * there is a mess of go packages that needs unpicking. It's the sort of thing where we/Debian has updated one package to a given version, but some of the rdeps upstream are still pinning an older version.
    * this is blocking some potentially useful packages
  * autopkgtest for r-cran-* is in excuses quite often, this needs to be checked and probably custom test triggers (there also is a whopping 1.8k r-cran-* tests in http://autopkgtest.ubuntu.com/running atm)
  * perl micro update 5.30.3 - On Monday there was another upload of 5.30.3-4 and someone overzealously removed all perl tests. Cpaelzer/xnox/Laney re-triggered the tests the results ahve to be re-checked.
  * [cpaelzer] php-horde* needs [[https://bugs.launchpad.net/ubuntu/+source/php-horde/+bug/1868281|removal]] and sync block to be processed by an Archive Admin.


 * Not ready yet
  * haskell-* seems like mass-FTBFS
    * [cpaelzer] this is due to a change in haddock-interface, per coordination with Debian the following need to be resolved first. I've asked to be pinged once things are ready to re-trigger builds/tests on Ubuntu as needed. That will be visible on [[https://lists.debian.org/debian-haskell/2020/06/msg00003.html|their ML as reply]] and I'll ping the current +1 duty people.
      * Because of the haddock-interface change, every single package needs a sourceful upload. This isn't relevant for Ubuntu.
      * All packages need to be updated to the version in the package-plan, which generally corresponds to Stackage LTS 15.13.
      * Both packages which do not need a version update and packages which do need to be upgraded also need patching because we do not automatically apply Hackage metadata updates which help
        keep Stackage consistent.
      * All new packages pulled in as dependencies of the new versions
        need to be packaged, and get processed through NEW. Some of
        these packages have been stuck in NEW for weeks, but most of
        them have not been packaged at all.
      * Build/test failures need to be addressed.

+1 Maintenance Status

Proposed Migration

Known issues

In progress

Items that are being actively worked on. Please always mention the name / irc nick, who is working on that.

Things to pick from

These are issues that are not yet done, after somebody's rotation/shift ended. Or someone has spotted them but had no chance to work on it. If someone trying to pick them realized that it was resolved by time/someone-else please be so kind and remove the related entries.

  • Ongoing transitions
    • there are plenty of rust packages stuck in -proposed. Unfortunately the rust packages we sync from Debian are not generally in a consistent state allowing migration (to Debian testing or to the Ubuntu release). If someone takes a look at these, they will probably want to recommend removals of various leaf packages blocking transitions.
    • the r-base transition has now started
    • vcmi ftbfs, pinged xnox about it, boost 1.71 related
    • ubuntu-app-launcher, url-dispatcher, unity fallout
      • Who to approach to remove those? xnox will look into them
    • rust-* some test and many "unsatisfiable depends", one needs to check how this was meant to work ...
      • the rust packages only build if uploaded in a very precise order, which they were to debian but the debian import freeze messes this up for us
      • it's not clear that any of these packages help users of Ubuntu at all, so putting effort into these packages has a low ROI IMO (I == mwhudson)
    • there is a mess of go packages that needs unpicking. It's the sort of thing where we/Debian has updated one package to a given version, but some of the rdeps upstream are still pinning an older version.
      • this is blocking some potentially useful packages

    • autopkgtest for r-cran-* is in excuses quite often, this needs to be checked and probably custom test triggers (there also is a whopping 1.8k r-cran-* tests in http://autopkgtest.ubuntu.com/running atm)

    • perl micro update 5.30.3 - On Monday there was another upload of 5.30.3-4 and someone overzealously removed all perl tests. Cpaelzer/xnox/Laney re-triggered the tests the results ahve to be re-checked.

    • [cpaelzer] php-horde* needs removal and sync block to be processed by an Archive Admin.

  • Not ready yet
    • haskell-* seems like mass-FTBFS

      • [cpaelzer] this is due to a change in haddock-interface, per coordination with Debian the following need to be resolved first. I've asked to be pinged once things are ready to re-trigger builds/tests on Ubuntu as needed. That will be visible on their ML as reply and I'll ping the current +1 duty people.

        • Because of the haddock-interface change, every single package needs a sourceful upload. This isn't relevant for Ubuntu.
        • All packages need to be updated to the version in the package-plan, which generally corresponds to Stackage LTS 15.13.
        • Both packages which do not need a version update and packages which do need to be upgraded also need patching because we do not automatically apply Hackage metadata updates which help
          • keep Stackage consistent.
        • All new packages pulled in as dependencies of the new versions
          • need to be packaged, and get processed through NEW. Some of these packages have been stuck in NEW for weeks, but most of them have not been packaged at all.
        • Build/test failures need to be addressed.

PlusOneMaintenanceTeam/Status (last edited 2020-10-21 18:41:40 by rbalint)