On 03/11/2013 12:30 AM, Rex Dieter wrote:
alekcejk@xxxxxxxxxxxxxx wrote:
Last KDE nightly composes failed because of error
DEBUG util.py:264: Error creating Live CD : Failed to build transaction :
MySQL-libs conflicts with mariadb-libs-5.5.29-7.fc19.x86_64
http://koji.fedoraproject.org/koji/taskinfo?taskID=5100301
http://koji.fedoraproject.org/koji/taskinfo?taskID=5100302
Previous composes before importing new MySQL package was fine,
so last composes problem related with new MySQL.
There is also other problem with missing MySQL component in bugzilla,
looks like for bugzilla MySQL=mysql, it finds mysql bugs
for MySQL package.
In addition to there not being any MySQL bugzilla component,
1. Shouldn't the 'mysql' component be retired and blocked now?
It's done now.
2. Should the "best practice" be to use
Requires: mariadb, BuildRequires: mariadb-devel
instead of
Requires: mysql, BuildRequires: mysql-devel
now?
mysql and mysql-devel should be still OK, since only mariadb packages
currently provide these symbols (they became only virtual names).
3. And long-term, what to do about MySQL-libs getting pulled into live
image composes? can it be blacklisted or something?
It won't be problem only in case of live image composes, but in
requiring libmysqlclient.so.18 generally, because RPM simply chooses one
of the packages providing the same library. It is officially
un-deterministic (usually it is the shorter one), but we cannot depend
on it right now.
So to solve the live image composes issue for now I adjusted the major
soname number to libmysqlclient.so.1018. I know it doesn't solve all the
issues, though.
Honza
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel
