[arch-dev-public] problem with community repo db files
Paul Mattal
paul at mattal.com
Tue May 29 16:41:28 EDT 2007
Eric Belanger wrote:
> On Sat, 26 May 2007, eliott wrote:
>
>> There is some problem with the script that updates the community db file.
>>
>> I couldn't figure out exactly what the breakage is..other than the script
>> itself not working.
>>
>> As a temporary measure, I manually regenerated the i686 and x86_64 repo db's.
>>
>
> I went through the logs and here's the errors I saw. Maybe you are
> already aware of there error messages. In the log for the
> x86_64 community repo (/home/aur/tupkgupdate64.log), all packages that I
> had uploaded had this error message:
>
> ==> ERROR: could not find xv-3.10a-8-i686.pkg.tar.gz - aborting
> Updatesync upd returned an error!
> cp --preserve=timestamps '/home/aur/packages64/full/xv-3.10a-8.pkg.tar.gz' '/home/ftp/community/os/x86_64/xv-3.10a-8.pkg.tar.gz'
> rm '/home/ftp/community/os/x86_64/xv-3.10a-7.pkg.tar.gz'
> rm '/home/aur/packages64/full/xv-3.10a-8.pkg.tar.gz'
> fakeroot updatesync upd '/home/ftp/community/os/x86_64/community.db.tar.gz' '/home/aur/cvs64/multimedia/xv/PKGBUILD' '/home/ftp/community/os/x86_64'
>
> Why is there a reference to i686 in the first line? It looks like the
> script for community64 is trying to use the wrong package name.
>
>
> In the log for the i686 community repo, the error message for each
> package is:
>
> Traceback (most recent call last):
> File "/home/aur/aur/tupkg/update/tupkgupdate", line 578, in <module>
> retval = runRepoAdd(repo_dir, package.new.file)
> File "/home/aur/aur/tupkg/update/tupkgupdate", line 311, in runRepoAdd
> destfile = os.path.join(repo, os.path.basename(package.new.file))
> AttributeError: 'str' object has no attribute 'new'
>
> I don't know what it means. I hope this will help in figuring out what's
> the problem.
Thanks, Eric, for the help in diagnosis. I believe I've now actually
fixed this issue in the i686 repo. Can someone confirm with a
package they added/updated in the last day or so?
If this is the right fix, I will then try to apply this fix to the
x86_64 update script.
- P
More information about the arch-dev-public
mailing list