The bugzilla story

The bugzilla story

[Alexander Nyberg]

Hi

As some of you may have noticed I've been doing a run on bugzilla for
the last few days to close old bugs and upgrade those who still are real
bugs.

I have a few things that I would like to enforce on bugzilla to makes it
more maintanable which it clearly hasn't been for a while - there are
lots and lots of bugs open older than 2.6.0-final.

I think the alternative trees section should be dropped. This is
especially a matter for -mm which has most reports of the alternative
trees in bugzilla. -mm changes way too rapidly to keep track of at
bugzilla ending up with open bugs that are fixed long ago.

I also think this goes for any alternative tree, that problems should be
reported directly to the maintainer/LKML of the tree. Only if a problem
can be reproduced with the mainline kernel should the bug be reported at
bugzilla. 

New bugzilla reports against other trees than mainline should be
rejected and ask the submitter to report directly to the
maintainer/LKML.

Andrew, what do you think about bug reports against -mm on bugzilla?

Does anyone see a problem with this? 

Thanks
Alexander

Re: The bugzilla story

[Zwane Mwaikambo]

On Mon, 6 Dec 2004, Alexander Nyberg wrote:

> I think the alternative trees section should be dropped. This is
> especially a matter for -mm which has most reports of the alternative
> trees in bugzilla. -mm changes way too rapidly to keep track of at
> bugzilla ending up with open bugs that are fixed long ago.
> 
> I also think this goes for any alternative tree, that problems should be
> reported directly to the maintainer/LKML of the tree. Only if a problem
> can be reproduced with the mainline kernel should the bug be reported at
> bugzilla. 

Hi Alexander,
	Yes i agree, -mm has so many bugs fixed and introduced inbetween 
it's short releases that it does often lead to extremely short lived bug 
reports. Thanks for wading through them all.

	Zwane

Re: The bugzilla story

[Martin J. Bligh]

> As some of you may have noticed I've been doing a run on bugzilla for
> the last few days to close old bugs and upgrade those who still are real
> bugs.
> 
> I have a few things that I would like to enforce on bugzilla to makes it
> more maintanable which it clearly hasn't been for a while - there are
> lots and lots of bugs open older than 2.6.0-final.
> 
> I think the alternative trees section should be dropped. This is
> especially a matter for -mm which has most reports of the alternative
> trees in bugzilla. -mm changes way too rapidly to keep track of at
> bugzilla ending up with open bugs that are fixed long ago.
> 
> I also think this goes for any alternative tree, that problems should be
> reported directly to the maintainer/LKML of the tree. Only if a problem
> can be reproduced with the mainline kernel should the bug be reported at
> bugzilla. 
> 
> New bugzilla reports against other trees than mainline should be
> rejected and ask the submitter to report directly to the
> maintainer/LKML.
> 
> Andrew, what do you think about bug reports against -mm on bugzilla?
> 
> Does anyone see a problem with this? 

The subtrees section was created for exactly that reason - to isolate the
bugs in alternate trees, and keep the out of everyone else's hair. Now if
one was to argue that bugs in -mm should be cleaned up much more agressively,
that'd make a lot of sense ... but that should be fairly easy to do as 
they're in a separate category you can search by.

The main issue is that there haven't been many people doing what you're doing
now ... going through and housekeeping the bugs.

M.

Re: The bugzilla story

[Alan Cox]

On Llu, 2004-12-06 at 16:26, Alexander Nyberg wrote:
> The thing is that -mm changes so fast that a bug reported can be solved
> an hour later, leaving a stale bug report for a few days (or years).
> Whoever wants to pick up the bug quite much has to mail either the
> bug-submitter asking if the bug has been resolved, mail the maintainer
> of whatever area the bug concerns or mail akpm. 
> 
> This leads me to thinking that bugzilla doesn't serve any functionality
> for at least -mm.

Sometimes they do - for looking back and finding when a problem came in,
or for spotting common patterns. They are less useful but not of no use

Re: The bugzilla story

[Alexander Nyberg]

On Mon, 2004-12-06 at 07:28 -0800, Martin J. Bligh wrote:
> > As some of you may have noticed I've been doing a run on bugzilla for
> > the last few days to close old bugs and upgrade those who still are real
> > bugs.
> > 
> > I have a few things that I would like to enforce on bugzilla to makes it
> > more maintanable which it clearly hasn't been for a while - there are
> > lots and lots of bugs open older than 2.6.0-final.
> > 
> > I think the alternative trees section should be dropped. This is
> > especially a matter for -mm which has most reports of the alternative
> > trees in bugzilla. -mm changes way too rapidly to keep track of at
> > bugzilla ending up with open bugs that are fixed long ago.
> > 
> > I also think this goes for any alternative tree, that problems should be
> > reported directly to the maintainer/LKML of the tree. Only if a problem
> > can be reproduced with the mainline kernel should the bug be reported at
> > bugzilla. 
> > 
> > New bugzilla reports against other trees than mainline should be
> > rejected and ask the submitter to report directly to the
> > maintainer/LKML.
> > 
> > Andrew, what do you think about bug reports against -mm on bugzilla?
> > 
> > Does anyone see a problem with this? 
> 
> The subtrees section was created for exactly that reason - to isolate the
> bugs in alternate trees, and keep the out of everyone else's hair. Now if
> one was to argue that bugs in -mm should be cleaned up much more agressively,
> that'd make a lot of sense ... but that should be fairly easy to do as 
> they're in a separate category you can search by.
> 
> The main issue is that there haven't been many people doing what you're doing
> now ... going through and housekeeping the bugs.

The thing is that -mm changes so fast that a bug reported can be solved
an hour later, leaving a stale bug report for a few days (or years).
Whoever wants to pick up the bug quite much has to mail either the
bug-submitter asking if the bug has been resolved, mail the maintainer
of whatever area the bug concerns or mail akpm. 

This leads me to thinking that bugzilla doesn't serve any functionality
for at least -mm.

Alexander

Re: The bugzilla story

[Rafael J. Wysocki]

On Monday 06 of December 2004 17:14, Alan Cox wrote:
> On Llu, 2004-12-06 at 16:26, Alexander Nyberg wrote:
> > The thing is that -mm changes so fast that a bug reported can be solved
> > an hour later, leaving a stale bug report for a few days (or years).
> > Whoever wants to pick up the bug quite much has to mail either the
> > bug-submitter asking if the bug has been resolved, mail the maintainer
> > of whatever area the bug concerns or mail akpm.

I have stopped reporting the -mm issues using bugizlla exactly for this 
reason.  Apparently, the reports get more attention when they are sent to 
LKML. ;-)

> > This leads me to thinking that bugzilla doesn't serve any functionality
> > for at least -mm.
> 
> Sometimes they do - for looking back and finding when a problem came in,
> or for spotting common patterns. They are less useful but not of no use

Yes, but many things are not reported there anyway.  IMVHO the bugzilla is an 
overkill for -mm bug reporting.  The -mm kernels are released too often, they 
change too much, and the issues are often too simple to be reported this way.

Greets,
RJW

-- 
- Would you tell me, please, which way I ought to go from here?
- That depends a good deal on where you want to get to.
		-- Lewis Carroll "Alice's Adventures in Wonderland"

Re: The bugzilla story

[Jeff Garzik]

Alexander Nyberg wrote:
> Hi
> 
> As some of you may have noticed I've been doing a run on bugzilla for
> the last few days to close old bugs and upgrade those who still are real
> bugs.
> 
> I have a few things that I would like to enforce on bugzilla to makes it
> more maintanable which it clearly hasn't been for a while - there are
> lots and lots of bugs open older than 2.6.0-final.

Thanks for running through these...

	Jeff