Hi again,
besides my suggested solution to split up emboss-lib again (and when
doing so make the package emboss-lib a metapackage depending from single
packages to match all its rdepends) I wonder whether we should provide
EMBOSS for 64 bit architectures only. While we probably need to file a
lot of removal requests to 32 bit packages of its rdepends it somehow
fits the reality that these days nobody seriously runs emboss on 32bit
architectures. As explained in the according wiki page[1] other binary
distros are dropping 32-bit at all and Debian rather cares for
automotive, IOT, TVs, routers, plant control, building
monitoring/control, cheap Android phones - nothing that is using EMBOSS.
I would really like to get some input from *our users* here on the
Debian Med list since I need your response to draw a sensible decision.
Kind regards
Andreas.
[1] https://wiki.debian.org/ReleaseGoals/64bit-time
Am Fri, Jan 26, 2024 at 10:44:09AM +0100 schrieb Andreas Tille:
> Hi Charles,
>
> I wonder how we can properly solve this bug. In the early stage of
> Emboss packaging obviously the packages
>
> libajax6,
> libajax6-dev,
> libnucleus6,
> libnucleus6-dev
>
> existed (thus the remaining Conflicts/Replaces on emboss-lib which can
> probably be removed right now). I wonder whether we should restore those
> single library package per shared library + devel package, merge
> static+shared library in one package or even merge
>
> libacd 6 emboss-lib (>= 6.6.0+dfsg)
> libajax 6 emboss-lib (>= 6.6.0+dfsg)
> libajaxdb 6 emboss-lib (>= 6.6.0+dfsg)
> libajaxg 6 emboss-lib (>= 6.6.0+dfsg)
> libensembl 6 emboss-lib (>= 6.6.0+dfsg)
> libnucleus 6 emboss-lib (>= 6.6.0+dfsg)
>
> in libemboss6
>
> libepcre 7 emboss-lib (>= 6.6.0+dfsg)
>
> in libemboss-pcre7 (ugly since its a code copy of pcre3 but well,
> that's another battle field) and
>
> libeplplot 3 emboss-lib (>= 6.6.0+dfsg)
>
> in libemboss-plplot3 to express the proper sonames inside the library
> package names. Any more sensible suggestion is pretty welcome.
>
> Kind regards
> Andreas.
>
> --
> http://fam-tille.de
--
http://fam-tille.de
