Hi,
thanks for your effor Tomas, please see few comments below:
----- Original Message -----
Hi,
I fixed the SQL query, please comment before I push. I also have a
design-related question, if I may.
Currently, the baseurl in the .repo file points to the root of the
repo. However, if a build is done on multiple architectures, will that
work, or do I need a baseurl pointing to that architecture
subdirectory? Based on the answer, I will also implement a link from
the copr overview page to the respective .repo files.
Seth would probably know this best. Seth?
Thank you, TR
<snip>
+@coprs_ns.route('/detail/<username>/<coprname>/<chrootname>.repo',
methods =
['GET'])
- methods = ['GET'] is default, you don't need to specify that explicitly
- If I get it correctly, the .repo filename is only chroot_name? That would mean something
like "fedora-18-x86_64.repo"? If so, this would only allow you to have one copr
in your yum.repos.d, right?
<snip>
+ urls = [build.results for build in copr.builds if build.results
!= None]
- I don't really like this. This will take _all_ the builds and traverse and create a
list of possibly all of them (since you can assume that almost all builds are complete
after the copr has lived for some time). I'd suggest using a generator with the same
expression and use it to traverse builds - something along the lines of:
urls = (<the same that you have>)
for u in urls:
if u:
break
else:
# all traversed, but no finished build found => page_not_found
# if we get here, u is a finished build => render repofile
+ if not urls:
+ return page_not_found('Repository not initialized: No builds in
{0}/{1} have finished.'.format(username, coprname))
+
+ response =
flask.make_response(flask.render_template('coprs/coprchroot.repo',
copr=copr, url=urls[0]))
+ response.mimetype='text/plain'
+ return response
Could you please fix these nits before committing?
Thanks a lot,
Slavek.
--
Regards,
Bohuslav "Slavek" Kabrda.