source: wine version: 10.0~repack-12+b1 severity: important blocks: #1122866 Quack, You changed the path of the wine binaries in #1063506 and that's nice but it would be nicer in the future if you communicated such important changes more directly so affected packages could prepare. Anyway I made the changes for dxvk now but I'm hitting another problem. You also decided to remove the /usr/bin/wineserver binary but it is needed in the dxvk autopkgtest to cleanup the lingering wineserver (wineserver --kill --wait) or autopkgtest is not very happy. Could you please reconsider? Regards. \_o<
Quack, I'm using a little script to help me create prefixes easily and I realized it's also using the same method. Sometimes an application misbehave or is frozen and the only easy way to be sure to kill all the processes and start over clean is to shutdown the wineserver. Regards. \_o<
Not the maintainer, but I remember the NEWS said that the binaries have been relocated to support multi-arch installs for wine 10.x. /usr/lib/x86_64-linux-gnu/wine/wineserver appears to exist. HTH, Chris
Quack, Thanks Chris, I did not realize this was moved. Now with this I would need to find the triplet. Seems like using `gcc -dumpmachine` is the easiest way to find it and I can make sure gcc is installed for the autopkgtest but if you need it in a script like my wine launcher then gcc might not be installed. It would be nice to have a wrapper/symlink/alternative for the native arch. \_o<
On i386 this returns i686...
Since this is only for a test, you could add a test dependency on
dpkg-dev and then do something like
DEB_HOST_MULTIARCH=$(dpkg-architecture -qDEB_HOST_MULTIARCH)
/usr/lib/${DEB_HOST_MULTIARCH}/wine/wineserver --kill --wait
cu
Adrian
Quack,
That is… unfortunate.
I was looking at my wrapper script in Python and it seems that would do
the trick:
sysconfig.get_config_var('MULTIARCH')
but now I wonder if that's not going to hit the same problem although I
have no way to test.
Would you happen to be able to test it for me please?
I would prefer to keep runtime tests reflecting the real situation but
I'm fine adding this as a *temporary* solution to unblock wine in
testing.
Thanks Adrian.