[Mono-dev] dependency conflicts on rpm based distros
Martin Vogt
vogt at itwm.fraunhofer.de
Thu Jul 23 06:21:34 EDT 2009
Andrew Jorgensen wrote:
> Hi Martin,
>
> If mono-find-requires says that 2.0.3.0 is required that's really what is required and 2.0.3.2 will not work.
>
Hello Andrew,
Should banshee-1 crash?
Despite the fact that banshee-1 may not the appropiate example,
because it almost does not run at all for me (with taglib-sharp-2.0.3.0
and taglib-sharp-2.0.3.2) I can start it without a crash.
And at least it opens the dll:
>vogt[~]>lsof -p 6106 | grep taglib-sharp
>banshee-1 6106 vogt mem REG 253,1 287744 1134919
>/usr/lib/mono/gac/taglib-sharp/2.0.3.2__db62eba44689b5b0/taglib-sharp.dll
6106 is the pid of banshee.
I looked an the configure.ac for taglib-sharp:
2.0.3.0:
--------
ASSEMBLY_NAME="taglib-sharp"
VERSION="2.0.3.0"
ASSEMBLY_VERSION="$VERSION"
POLICY_2_0_VERSIONS="2.0.0.0-$ASSEMBLY_VERSION"
AC_SUBST(POLICY_2_0_VERSIONS)
2.0.3.2:
--------
AC_INIT([taglib-sharp], [2.0.3.2])
AM_INIT_AUTOMAKE([1.9 dist-bzip2 tar-ustar])
AM_MAINTAINER_MODE
ASSEMBLY_NAME="$PACKAGE"
ASSEMBLY_VERSION="$VERSION"
POLICY_2_0_VERSIONS="2.0.0.0-$ASSEMBLY_VERSION"
AC_SUBST(POLICY_2_0_VERSIONS)
AC_SUBST(VERSION)
AC_SUBST(ASSEMBLY_VERSION)
AC_SUBST(ASSEMBLY_NAME)
But the question is, does the find-requires script follows the
recommended quidelines for mono library versioning or not?
I dont know much about mono and its internal libs/versions, but the
find-requires script return "mono(taglib-sharp) = 2.0.3.0"
Is this meant to be the <major>.<minor>.<build>.<revision> from a
mono lib?
If it is, then the common practice is, that a lib is compatible
if only "<build>.<revision>" changes.
RPM cannot handle this.
regards,
Martin
> There are certainly problems with the mono-find-* scripts but that is not one of them.
> You could verify what I'm saying by manually installing banshee
> using the --nodeps option passed to rpm. You will find that banshee
> will not run properly without the correct version of taglib-sharp.
>
> Andrew
>
More information about the Mono-devel-list
mailing list