About bug submission

[Mario Torre]

Il giorno dom, 25/11/2007 alle 21.05 -0800, David Herron ha scritto:

[--- Long mail follow ---]

Hello David, Phil,

> Mario, FWIW I've forwarded the bug information to the quality team and 
> I'm hoping to find out why we missed that one (since you found it using 
> one of our demo cases)

Thanks for the quick and kind replies. 

Actually I did not find the bug, only tracked down the problem, but I
would like to point that this mail was not intended to try to resolve
this bug or to push a quick reference in the Sun bug database (though of
course I'm happy this was the result :), just to try to address a need
that I have found and, talking to other hackers in #classpath, it's felt
as a general problem.

Also, I'm really happy (and though I missed the blog post from Mark
Reinhold I had no doubt about that) to see that this will be addressed
at some time in the not so distant future.

> Of course there's no inter-bug-tracking-system-protocol which can 
> automatically propogate bugs from one system to an upstream or 
> downstream system.  Of course that's a desired thing.  Of course there 
> are dozens of hurdles to jump to get such a thing implemented, and they 
> aren't all technical hurdles either.  For example for a non-Sun person 
> to automatically enter a bug into Sun's bug system, well, there's 
> several IT policies which would prevent us from doing so.  IT policies 
> which were decided prior to the age of widespread Open Source at Sun.

I would not change that necessarily, of course easing the procedure is
always good, but I understands the needs behind this, and I'm not
against waiting for few days (or up to few weeks, FWIW) to have a valid
id number.

I only think that we could setup a procedure that easy the way bugs are
submitted.

> What sort of automatic system would you be looking for?

As this will require time, talking about this procedure now may sound a
bit premature, but anyway, I want to share with you my experience.

As you probably know, we use eclipse (if not emacs or vi) to develop
classpath, and eclipse has a nice plugin named mylyn that connects to
the various bug databases and let you perform queries, and follow
progresses of bugs.

Though I prefer the web interface in this case, it's handy as you can
track changes directly in the IDE.

Now, we could just have some system that simply fetch bug data from a
source and copy all the relevant fields into the main bug database, like
mylyn does. This could work in both directions
(Sun<->IcedTea<->Distributors).

This way, to submit a bug you always have to follow whatever procedure
is needed (say, have a valid account, wait for a first review etc...),
but filing the bug data itself is automatic, and ensures that once done
and accepted, the bug is linked and we can track the lifetime of this
bug in any bug system.

Example test case: ----------

Guido Rossi, developer of John Doe Linux distribution finds a bug in
IcedTea, and submit it to our bugzilla.

Laura Palmer, fellow Classpath contributor, read the bug entry and find
it's an upstream problem. She then goes to file it upstream. After log
in and having filled the preliminary data (What module it belongs to,
what version of the jdk was tested etc..), then she points the bug id of
the IcedTea bugzilla, instead of copying all the data from this
database. The system then copies all the relevant data (actually it
could link it, so changes in one system are also added to the other).

After a couple of days, Laura gets a mail from Bob at Sun, kindly
telling the new bug id in the upstream bug system.

Laura the can link back the bug from upstream, so that the bug can be
tracked easily.

End of example test case ----------

If we do that in public, we can also share some code needed to adapt a
similar procedure to other bug systems/projects. I would also volunteer
for that, if there is interest, and given someone helps me with the
details.

The end result would be to also increase chances to fix the bugs in
question, because tracking downstream an upstream bug let a casual
developer that finds the bug in his distribution bugzilla and has the
skill to code the fix, to actually fix the bug and contribute the fix
back.

Of course, just an idea, feel free to mark is a "dumb" one :)

Thanks,
Mario
-- 
Lima Software - http://www.limasoftware.net/
GNU Classpath Developer - http://www.classpath.org/
Fedora Ambassador - http://fedoraproject.org/wiki/MarioTorre
Jabber: neugens at jabber.org
pgp key: http://subkeys.pgp.net/ PGP Key ID: 80F240CF
Fingerprint: BA39 9666 94EC 8B73 27FA  FC7C 4086 63E3 80F2 40CF

Please, support open standards:
http://opendocumentfellowship.org/petition/
http://www.nosoftwarepatents.com/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Questa ? una parte del messaggio	firmata digitalmente
URL: <http://mail.openjdk.java.net/pipermail/discuss/attachments/20071126/a054cf6c/signature.asc>