This may not be the best place to ask... But what exactly is the problem with mariadb shipping with static libs? On 14-May-2014 8:53 PM, "Doug Newgard" <scimmia@archlinux.info> wrote:
On 2014-05-14 06:51, Christian Hesse wrote:
Antonio Rojas <nqn1976list@gmail.com> on Wed, 2014/05/14 11:51:
Christian Hesse wrote:
I think gcc, glibc, llvm and friends are ok. But zlib, mupdf, mysql/mariadb and some others should go away.
There was a to-do list to cleanup all static libs, so all remaining ones are there for a reason. Check the changelogs for the specific reasons for each package, e.g.
https://projects.archlinux.org/svntogit/packages.git/ commit/trunk?h=packages/zlib&id=83d05088a1cb1b56561b9ebe365d18d033752c72
Is it possible to fix binutils testsuite?
Remember the security flaws in zlib? Does anybody know what package has been built against static zlib?
https://projects.archlinux.org/svntogit/community.git/
commit/trunk?h=packages/mupdf&id=c34f53eeb8efd6b4b033c2fdc58d0a 329efdeeef
This brings the static libraries back, but there is no reason.
libmariadbclient ships with static libraries because a package from AUR (neko) requires it. I think anybody should fix neko, but shipping official packages with static libraries in this situation is just stupid.
Removing static libraries (and keeping them away!) should be treated more strict.
That's completely up to the maintainer. If they decide to ship static libs for any reason, that's their choice to make. There are very few "strict" rules.