On 01/11/2024 10:12, Emilio Pozuelo Monfort wrote:
On 30/10/2024 17:44, Bastien Roucariès wrote:
Le mercredi 23 octobre 2024, 12:03:21 UTC Emilio Pozuelo Monfort a écrit :
Hi,

Control: tags -1 confirmed

On 20/10/2024 11:04, Bastien Roucariès wrote:
Le jeudi 17 octobre 2024, 20:02:56 UTC Johannes Schauer Marin Rodrigues a écrit :
Hi,

On Tue, 24 Sep 2024 12:58:48 +0000 Bastien =?ISO-8859-1?Q?Roucari=E8s?= <ro...@debian.org> wrote:
Le mardi 20 août 2024, 07:11:13 UTC Emilio Pozuelo Monfort a écrit :
On 28/07/2024 20:56, Bastien Roucariès wrote:
control: tags -1 - moreinfo

Hi,

Last reverse deps of lib magick pipeline is not really bad
https://salsa.debian.org/debian/imagemagick/-/pipelines/708187

A lot of failure are due to broken package or does not use pkgconfig

I suppose we could go to experimental

Yes, uploading to experimental would be the first step, as I said on my previous email. Then we would need bug reports for packages that fail to build against imagemagick 7. Make those bugs block this one and use some usertag to ease tracking.

If you want this to be done for trixie, we need to move fast.

Could we began with transition ?

I have fixed a few package and I think it is ok.

I am working with java team to fix jamagick

to support Bastien's request, I created a wiki page as an overview of the
current status:

https://wiki.debian.org/ImageMagick7

This page includes those source packages which FTBFS in the salsa-ci pipeline
that Bastien set up to test reverse dependencies of imagemagick:

https://salsa.debian.org/debian/imagemagick/-/pipelines/746158/builds

Nice pipeline, and thanks for the wiki page.

Most of the remaining failures are packages which FTBFS for other reasons than
imagemagick 7. I have linked the respective FTBFS bugs behind the "other FTBFS"
links. The remaining packages fall into either of the following categories:

   - not in main
   - popcon below 10
   - has a bug filed blocking this bug
   - dune-* (i believe those are false positives as i cannot reproduce them
     locally)

There are only two exceptions with which I need Bastien's help:

   - src:dx has not seen a new upstream release in 16 years and can be built
     without imagemagick by running ./configure --without-magick
Fixed in last salsa
   - src:digikam FTBFS with "MagickCore/magick-config.h:25:10: fatal error:
     MagickCore/magick-baseconfig.h: No such file or directory" -- is that a
     missing dependency of src:digikam on libmagickcore-dev (which pulls in
     libmagickcore-7-arch-config) or should libmagick++-dev gain a dependency on
     libmagickcore-dev?

fixed in last salsa



Should I be looking into the FTBFS issues in contrib and non-free as well?

Bastien, I also created a MR against imagemagick's salsa repo which refactors
your rdep building construct, turning it from a manual selection of rdeps into
an automatic computation using build-rdeps from devscripts. You can have a look
if you like but it needs a bit more work and unfortunately, salsa does still
seem to be stuck, so my work on that is stalled:

https://salsa.debian.org/debian/imagemagick/-/merge_requests/1

Thanks!

Things are looking good. Let's go ahead.

Go ahead two day ago forget to mention here

I scheduled the binNMUs. Unfortunately there's many packages that build-depend on versioned imagemagick 6 -dev packages and so the resulting binNMUs still build against the old imagemagick 6. This worries me as many of this will likely FTBFS against imagemagick 7 and so the transition is probably in a worse state than I thought.

See php-imagick, pfstools, odr-padenc, kxstitch, gnustep-gui, drawtiming, wmaker, photoqt

Please file bugs for those.

I filed bugs for: jhead drawtiming dx kxstitch libopenshot odr-padenc pfstools php-imagick synfig wand wmaker photoqt.

Emilio

Reply via email to