Drop support for 32-bit builds
Owners, do not implement this work until the FESCo vote has explicitly ended. The Fedora Program Manager will create a tracking bug in Bugzilla for this Change, which is your indication to proceed. See the FESCo ticket policy and the Changes policy for more information.
REMINDER: This ticket is for FESCo members to vote on the proposal. Further discussion should happen in the devel list thread linked above.
Metadata Update from @amoloney: - Issue set to the milestone: Fedora Linux 41 - Issue tagged with: system wide change
+1
Aren't there still multiarch libraries that need php at build time? For example, libvpx build-requires php-cli.
libvpx
php-cli
I don't see anything in libvpx sources code using PHP Dropping this BuildRequires seems OK (local test build)
P.S. https://src.fedoraproject.org/rpms/libvpx/pull-request/7
OK, that was just one example -- have you looked for others? Something like:
$ dnf repoquery --release rawhide '--disablerepo=*' '--enablerepo=fedora-source' --arch=src \ --whatrequires=php{,-cli,-devel} -q | grep -v ^php composer-0:2.7.2-1.fc41.src flamegraph-0:1.0-17.20231107gitcd9ee4c.fc40.src libguestfs-1:1.52.0-7.fc40.src libkolabxml-0:1.2.0-22.fc40.src libvpx-0:1.14.0-1.fc40.src libvpx8-0:1.13.1-1.fc40.src lsp-plugins-0:1.2.15-1.fc41.src mapserver-0:8.0.1-13.fc40.src mediawiki-0:1.40.0-3.fc40.src mlt-0:7.22.0-4.fc40.src owfs-0:3.2p4-6.fc40.src roundcubemail-0:1.6.6-2.fc40.src scip-0:9.0.0-1.fc41.src soplex-0:7.0.0-1.fc41.src swig-0:4.2.1-2.fc41.src uwsgi-0:2.0.24-1.fc40.src wordpress-0:6.4.3-1.fc40.src
Some of those are noarch, which is fine, and some like libguestfs already dropped i686 themselves. swig is far-reaching, but looks like it only uses php for tests.
noarch
libguestfs
i686
swig
php
swig is used to generate library bindings, php support is optional, and already conditional (so easy to disable on 32-bit)
some libraries provide php bindings (mlt, libkolabxml...), so the subpackage may be disabled.
For uwsgi, the php plugin is already conditional and may be disabled on 32-bit
To be discussed at mondays meeting.
@remi if you could make it that would be great.
Metadata Update from @kevin: - Issue tagged with: meeting
This was discussed during the meeting, where we realized that the ticket was already approved here according to the policy.
APPROVED (+3, 0, 0)
Metadata Update from @zbyszek: - Issue untagged with: meeting - Issue close_status updated to: Accepted - Issue status updated to: Closed (was: Open)