+1 maintenance reports

[Olivier Gayot]

Hi,

I was on +1 maintenance this week (Jan 30th to Feb 3rd). I worked on the 
following items:

git-annex:
---------
   * attempted a merge from Debian of 10.20221003-3 to fix FTBFS. Sadly 
it fails to link from source with a relocation error. I could not 
reproduce the error when building from Debian.
   * Build logs: 
https://launchpadlibrarian.net/648774267/buildlog_ubuntu-lunar-amd64.git-annex_10.20221003-3ubuntu1~ppa1_BUILDING.txt.gz
   * There is an existing bug report for Bryce Harrington: 
https://bugs.launchpad.net/ubuntu/+source/git-annex/+bug/2003918
   * if someone who speaks Haskell wants to have a look.. :)

numpy transition:
----------------
   * bcolz:
     * filed a removal bug since it FTBFS against numpy 1.24 and 
upstream declared the project unmaintained.
     * https://bugs.launchpad.net/ubuntu/+source/bcolz/+bug/2004256
   * tpot:
     * Debian merged a fix but has not yet released it.
     * 
https://salsa.debian.org/science-team/tpot/-/commit/b7989129ccecae0b9cff9a3fa5b863a0d551815e
     * filed https://bugs.launchpad.net/ubuntu/+source/tpot/+bug/2004166
     * should be resolved by a manual sync when Debian releases 
0.11.7+dfsg-6.

python3-defaults transition:
---------------------------
   * kitty ftbfs on ppc64el
     * the crypto test calls mlock(2) multiple times
     * on our ppc64el builders, pagesize = RLIMIT_MEMLOCK = 64k so only 
a single call to mlock(2) can succeed
     * discussed the issue briefly in #launchpad - turned out 
RLIMIT_MEMLOCK should be set to a more sensible value once builders get 
upgraded to 22.04.
     * uploaded a new version where the crypto test is skipped if the 
RLIMIT_MEMLOCK is too low
     * still need to open a PR upstream, but upstream is willing to take 
a patch
     * filed https://bugs.launchpad.net/ubuntu/+source/kitty/+bug/2004435
     * upload was sponsored

django-rich ftbfs:
-----------------
   * the build-time test suite failed because python-rich disables color 
and style when TERM=unkonwn
   * sent a patch upstream which was applied upstream with modifications
   * filed 
https://bugs.launchpad.net/ubuntu/+source/django-rich/+bug/2004553 and 
attached a debdiff for sponsorship

mailman3 failing autopkgtest:
----------------------------
   * autopkgtest failed because mailman3 takes to long to start in the 
test-bed.
   * the service consumes > 750 MiB of memory at startup, bringing down 
qemu testbed that use the default --ram-size
   * filed https://bugs.launchpad.net/ubuntu/+source/mailman3/+bug/2004658
   * attached a debdiff giving the service more time to start
   * also added more verbose output to the failing test
   * if the patch does not help, we should consider adding mailman3 to 
big_packages

misc:
----
asked for a bunch of no-change-rebuilds to complete some unfinished 
transitions:
   * igraph, zxing-cpp, xen, openimageio


Olivier