Quickly getting the status of multiple tickets

[Jason L Tibbitts III]

I maintain the code that spits out the review status reports at
http://fedoraproject.org/PackageReviewStatus/

In order to generate these reports, I run a few queries to get the list
of tickets in the Package Review component and process them a bit.  One
thing I need to know is whether any tickets in the Blocked By field are
still open, so I make a list of them and pass them to getbugssimple().
Unfortunately with the new bugzilla instance currently running, doing
this for just 500 tickets takes so long that everything aborts with
proxy errors, so I have to split the list and make multiple queries.

Is there any better way to do this?  I know bugzilla needs to know this
often, in order to strike through ticket dependencies which have been
closed.  Is there any way to simply get the status of dependencies out
of the result of a query without doing a separate lookup?

- J<

[Cole Robinson]

On 05/24/2013 11:10 AM, Jason L Tibbitts III wrote:
> I maintain the code that spits out the review status reports at
> http://fedoraproject.org/PackageReviewStatus/
>
> In order to generate these reports, I run a few queries to get the list
> of tickets in the Package Review component and process them a bit.  One
> thing I need to know is whether any tickets in the Blocked By field are
> still open, so I make a list of them and pass them to getbugssimple().
> Unfortunately with the new bugzilla instance currently running, doing
> this for just 500 tickets takes so long that everything aborts with
> proxy errors, so I have to split the list and make multiple queries.
>
> Is there any better way to do this?  I know bugzilla needs to know this
> often, in order to strike through ticket dependencies which have been
> closed.  Is there any way to simply get the status of dependencies out
> of the result of a query without doing a separate lookup?

You will still have the query them separately, but you can do it in such a way
that it should return minimal info and be much quicker. something like (untested):

bugs = bz.query(bz.build_query(bug_id=[list, of, 500, bug_ids, ...],
                               include_fields=["status"])

Will return a list of bug objects that are only filled with bug_id and status
values.

You could also do build_query(..., status=["NEW", "ASSIGNED", ...]) and it
will return only the bugs in those states, which will speed it up.

- Cole

[Jason L Tibbitts III]

> > > > > "CR" == Cole Robinson  writes:

CR> You will still have the query them separately, but you can do it in
CR> such a way that it should return minimal info and be much
CR> quicker.

Thanks, I hadn't thought of doing a query for just the ones I want.
I'll do some testing and see if I can't get the time down.  (What used
to run in under 30 seconds I'm now lucky to run in 500 after recent
bugzilla changes, so I'm really trying to optimize things.)

- J<

[Cole Robinson]

On 05/24/2013 11:54 AM, Jason L Tibbitts III wrote:
> > > > > > "CR" == Cole Robinson  writes:
>
> CR> You will still have the query them separately, but you can do it in
> CR> such a way that it should return minimal info and be much
> CR> quicker.
>
> Thanks, I hadn't thought of doing a query for just the ones I want.
> I'll do some testing and see if I can't get the time down.  (What used
> to run in under 30 seconds I'm now lucky to run in 500 after recent
> bugzilla changes, so I'm really trying to optimize things.)

Yeah something is funky with new bugzilla no doubt, XMLRPC queries are timing
out for a lot of people. I know the rhbz guys are working on it at least.

- Cole