Please import SF bugs for GNU Mailman

This question was expired because it remained in the 'Open' state without activity for the last 15 days.

mwhudson indicated he might be able to help with this.

Graham would be better to handle this.

Oops, you're right. I misremember the offer. Thanks!

This question was expired because it remained in the 'Open' state without activity for the last 15 days.

This shouldn't be expired.

Graham, Bjorn. Did this question expire while gmb was assigned? If so I think we have a bug. Revisions to the question via Edit question are not logged in the history...that might be a nice feature to add.

This is still Open though, because it hasn't been addressed yet AFAIK.

Graham, ping?

https://bugs.demo.launchpad.net/mailman/

Graham (who btw, rocks) put this up yesterday. I sent the announcement to the Mailman team, so let's give it a few days. We can close the question when we've done the production import.

Thanks Curtis Hovey, that solved my question.

Actually it seems like a bunch of of issues didn't get imported. Graham should have some details about that.

2008/11/27 Barry Warsaw <email address hidden>:
> Question #30336 on Launchpad itself changed:
> https://answers.edge.launchpad.net/launchpad/+question/30336
>
> Status: Solved => Open
>
> Barry Warsaw is still having a problem:
> Actually it seems like a bunch of of issues didn't get imported. Graham
> should have some details about that.

I tried doing a local re-import today but am unable to reproduce the
problem (i.e. everything gets imported, AFAICS, and all the bug
nicknames are set correctly). I'll keep working on this, but I suspect
that I might need to try it on staging or demo.

Did you get the same set of imported bugs as before? If not, will a new import overwrite any existing bugs? Does it make sense to just run the real import again in case it picks up anything new?

I also wouldn't be surprised if something wacky on SF is causing missed imports.

2008/12/3 Barry Warsaw <email address hidden>:
> Did you get the same set of imported bugs as before? If not, will a new
> import overwrite any existing bugs? Does it make sense to just run the
> real import again in case it picks up anything new?
>

Well, the problem is that if we just re-run the import all the stuff
we've already imported will be imported again because the import
script isn't smart enough to check to see whether it's already been
done (the import scripts for other bug trackers are but the SF import
script lags behind somewhat in terms of functionality). At best, we'd
see OOPSes as it tried to assign bug nicknames that already exist and
it would fall over. At worst it would re-import all the bugs and we'd
end up with a whole bunch of dupes. It's worth trying on staging,
though, just to see what happens.

> I also wouldn't be surprised if something wacky on SF is causing missed
> imports.
>

Part of the problem we had with the mailman import is that the XML
that got produced by the screen-scraping export tool thingamy was full
of weird control characters that made expat puke all over my shoes
whenever I tried to run the import script against it. I don't know if
this is SourceForge's fault (i.e. the control characters may have been
there in SourceForge's output in the first place) or the screen
scraper's.

One thing we could do is write a script to convert the SourceForge bug
dump format into the Launchpad bug interchange format. That's more
long-winded but we would then be more assured that the data we're
trying to import are all valid (and we can be a bit smarter about
weeding out stuff we've already done.

  Graham Binns said on 2008-12-03:

>2008/12/3 Barry Warsaw <email address hidden>:
>> Did you get the same set of imported bugs as before? If not, will a new
>> import overwrite any existing bugs? Does it make sense to just run the
>> real import again in case it picks up anything new?
>>
>
>Well, the problem is that if we just re-run the import all the stuff
>we've already imported will be imported again because the import
>script isn't smart enough to check to see whether it's already been
>done (the import scripts for other bug trackers are but the SF import
>script lags behind somewhat in terms of functionality). At best, we'd
>see OOPSes as it tried to assign bug nicknames that already exist and
>it would fall over. At worst it would re-import all the bugs and we'd
>end up with a whole bunch of dupes. It's worth trying on staging,
>though, just to see what happens.

Please see my comments on the bug report at <https://bugs.launchpad.net/mailman/+bug/294223>.

I have determined what is missing and it appears that what happened during the original import is

1.All the Bugs were imported.

2.All the Feature Requests were imported.

3.Patches through 538983 (* 2002-04-03 22:52) were imported when *fill in the blank* happened and the import process died.

I don't know how difficult it would be to just restart the import beginning with patch 543943 (* 2002-04-15 04:22) and continuing through the end of the patches, but if that could be done, it would solve the problem.