Antoine, as a reminder: mails sent to n...@bugs.debian.org are *not* cc-ed to the submitter. I have bounced your answer to Dominick.
Antoine Le Gonidec via Pkg-a11y-devel, le dim. 18 sept. 2022 20:38:35 +0200, a ecrit: > On Sun, 18 Sep 2022 09:59:01 +0200 Dominick Grift > <dominick.gr...@defensec.nl> wrote: > > libatk1.0-0 (2.46.0-1) pulls in at-spi2-core (and > > gsettings-desktop-schemas). It would be really nice if we could > > opt out of at-spi2-core. at-spi2-core is very intrusive and by itself it > > does not add any value. > > at-spi2-core is only a recommended dependency of libatk1.0-0, not a hard > dependency: > > $ apt depends libatk1.0-0 > libatk1.0-0 > Depends: libc6 (>= 2.4) > Depends: libglib2.0-0 (>= 2.62) > Recommends: at-spi2-core > at-spi2-core:i386 > > You should be able to remove at-spi2-core, or prevent its installation, and > keep libatk1.0-0 installed at the same time. Indeed. That being said, it's true that since recommends are pulled by default, this can be pulling by default more than desired. The reason for the recoommends was only that the atk translated messages are now in at-spi2-core.mo files, shipped by at-spi2-core. I have now uploaded a 2.46.0-2 version that adds a new package, at-spi2-common, to contain them, which atk can thus depend on without trouble. That will have to go through NEW, though. I have increased the severity of this bug to prevent the migration to testing in the meanwhile. Thanks for the report, Samuel
