#include <hallo.h>
* Jonathan Nieder [Mon, Nov 21 2011, 03:56:47PM]:
> retitle 649522 xz-utils/changelog: note on lzma_code@Base compatibility
> symbol is misleading
> severity 649522 important
> quit
>
> Eduard Bloch wrote:
>
> > Looking at the last changelog entry, it seems like it decided to
> > exterminate lzma_code in the public ABI.
>
> No, not true at all. I do not want to keep the unversioned
> compatibility symbol lzma_code@Base in liblzma5, because all programs
> should either:
>
> A. have been built against liblzma5, in which case they would be
> using the versioned symbol lzma_code@XZ_5.0, or
Seriously? I am just trying to build my package (apt-cacher-ng) against
current liblzma-dev, and it fails. Now I have also tried a random
package using liblzma* (fsarchiver) and it's FTBFSing too.
dpkg-shlibdeps: warning: can't parse dependency liblzma_private_symbols
dpkg-shlibdeps: error: invalid dependency got generated: liblzma5 (>=
5.1.1alpha+20110809), libcomerr2 (>= 1.01), liblzma_private_symbols,
libgpg-error0 (>= 1.10), libbz2-1.0, libgcrypt11 (>= 1.4.5), libc6 (>= 2.3.2),
liblzo2-2, libuuid1 (>= 2.16), e2fslibs (>= 1.41.99), zlib1g (>= 1:1.1.4),
libblkid1 (>= 2.16)
> Please do not panic. Your programs will continue to work without
> any action on your part.
Apparently not. And this failure message of dpkg-shlibdeps is
totally misleading, the first ideas were to look for a bug in dpkg-dev
or something caused by the current Perl transition.
Regards,
Eduard.
--
Gott zieht uns an der Hand, der einen,
der Teufel zieht an beiden Beinen.
-- Wilhelm Busch
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
