Re: [FWD] Re: [Jackit-devel] parallel installs to die
Hi all,
|--==> Free Ekanayaka writes:
FE> Hi Junichi,
|--==> Junichi Uekawa writes:
JU> Hi,
>>>http://svn.debian.org/wsvn/demudi/jack-audio-connection-kit/trunk/debian/?rev=0&sc=0
>>>
>>>and tested the new packages on amd64. Everything seems to work fine,
>>>and upgrades from the previous packages are possible through the dummy
>>>packages I've added in debian/control.
>>>
>>>If everything is fine for you I'd upload this new revision to
>>>unstable (as Etch is now in freeze I think it's safe to do that).
JU> I'd rather wait until etch release, since this transition will require
JU> rebuild of every single package that depends on libjack, and it
JU> usually takes around two months to settle. I don't want unstable to
JU> be in a too-broken state while etch is preparing to release.
FE> Waiting is fine for me, but actually this new package does not break
FE> any other package depending on the current jack library package.
FE> As far package dependences are concerned they are satisfied by a dummy
FE> libjack0.100.0-0 package which in turns depends on the new libjack0
FE> package, while the shared library file /usr/lib/libjack-0.100.0.so.0
FE> shipped by the current package is replaced in the new package by a
FE> symlink to libjack.0.0.23, making the linker happy.
FE> All the packages I've tested (hydrogen, ardour, zynaddsubfx, ..) work
FE> fine with the new packges.
I've doubled checked the new packages, both on amd64 and i386, and I
confirm that the work fine, without breaking other packages (this
0.102.20 upstream release was also included in the recent 1.0 64
Studio release, and no bug reports have come so far).
So if everybody is fine with it I'm going to upload the new package
within this week.
I'm also thinking to file bug reports against those packages that
build depend on libjack0.100.0-dev, and which should from now on use
simply libjack-dev, although a dummy libjack0.100.0-dev package
depending on libjack-dev is still provide to avoid FTBFSs.
Ciao!
Free
Reply to: