#3034 link bugzilla query in http://fedoraproject.org/PackageReviewStatus/NEEDSPONSOR.html
Closed: Fixed None Opened 12 years ago by till.

I would like to be able to easier find potential sponsorees. Can you please link the names of future maintainers on http://fedoraproject.org/PackageReviewStatus/NEEDSPONSOR.html with a link to a bugzilla query like:
https://bugzilla.redhat.com/query.cgi?component=Package%20Review&emailcc1=1&emaillongdesc1=1&emailreporter1=1&emailtype1=exact&product=Fedora&query_format=advanced&email1=EMAIL_OF_FUTURE_MAINTAINER

This will help to easier identify informal reviews performed by the future maintainer.


To avoid to specify the real e-mail address in the query URL, a regex could be used instead:
Then the @ can be substitude by a dot. For example with my e-mail address:
https://bugzilla.redhat.com/buglist.cgi?emailcc1=1&emailreporter1=1&emaillongdesc1=1&emailtype1=regexp&query_format=advanced&component=Package%20Review&product=Fedora&email1=opensource.till.name

I do not have access to the email address there since the script cannot log in to bugzilla.

Really, the scripts are pretty close to the limit of what I can do with the bugzilla interface. If I had access to more information I would have exposed it.

In general the script could login, but you need an account for it. I created one for my upstream reelease monitoring script, therefore nothing should be wrong with creating one for the report script.

Also note that once logged in (which happens during testing if I have a personal cookie file) I lose the real names and for whatever reason only get email addresses, which is kind of annoying and in my view less useful. And I still don't know how to make the script itself log in, so more detailed pointers would be useful.

Replying to [comment:4 tibbs]:

Also note that once logged in (which happens during testing if I have a personal cookie file) I lose the real names and for whatever reason only get email addresses, which is kind of annoying and in my view less useful. And I still don't know how to make the script itself log in, so more detailed pointers would be useful.

If the script is logged in, the reporter's name could be extracted from bug.longdescs, which contains all comments. The needinfo information can be found in bug.flag_types. Logging in works with {{{bz.login(user="my_user", password="my_password")}}} if bz is the Bugzilla instance.

If the source of the scripts is available, I can try to provide patches btw.

Of course the source is available; it's in the infrastructure git repo.

And I'm not really keen on sticking the login info directly in the script. Not really a fan of additionally complicating things but we'll see.

A test whether there is a needinfo flag set for the reporter seems to be the following:

{{{filter(lambda flag: flag["name"] == "needinfo", bug.flag_types)[0]["flags"][0]['requestee_id'] == bug.reporter_id}}}

The password could be stored in an external config file. I will see what I can come up with within the next days.

This is now implemented upstream. As before, this can't be pushed to Fedora's infrastructure until after the freeze and conversion of the other services to the new python-bugzilla.

If you want to see it running, though, I occasionally update http://www.math.uh.edu/~tibbs/rev-a/ while I'm testing. This isn't updated live, though, and may occasionally be broken as I'm implementing other things.

Login to comment on this ticket.

Metadata