Re: [gentoo-dev] bug wrangler queue is large...

2010-05-26 Thread Panagiotis Christopoulos
On 05:30 Wed 26 May , Jeroen Roovers wrote:
> To follow up on that, we could add some language to the point that if
> it takes more than X days (say a week) then assignment to maintainers
> proper should go through. This could be especially important when bugs
> do seem relevant but should be looked at by someone (or a freakin' whole
> "herd") with somewhat more intimate knowledge of the package in
> question.
> 
  Another idea crossed my mind when I was bug-wrangling the other day,
that could help, until Robin's "automatic assignment computation of
bugs"[1] gets implemented and becomes reality. What if we create a
partial-bug-wranglers mail alias, write a cron job to calculate
statistics from from the bug-wranglers queue and send mails to that
alias according to various rules eg. to send mail when the queue passed
a limit of bugs eg. 100 or 200 unassigned, or statistics about how many
bugs that were reported before more than a week are still in the queue.
  This way, people like me, who are willing to do partial bug-wrangling,
they just add themselves in the alias and get activated only when the
queue is huge or there are many old bugs around. The mails could work
well as a reminder and we avoid the huge bug-wranlgers mail traffic when we
monitor them in bugzie. I don't have time now, to implement any of this,
I don't know even if it's a good idea, but if you like it, I could try
to write a working script in some weeks (thanks to Jeremy(darkside) for showing
me the pybugz magic and Christian(idl0r) for the curl magic).


[1] http://thread.gmane.org/gmane.linux.gentoo.devel/66279

-- 
Panagiotis Christopoulos ( pchrist )
( Gentoo Lisp Project )


pgp54AaVjRZQK.pgp
Description: PGP signature


Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Graham Murray
Mike Frysinger  writes:

> the bug reporter can open their own bugs.  gentoo developers can open any 
> bug.  
> that's about it.

Which can be a pain for other users who suffered the same bug (and are
probably on the CC list), the maintainer says to re-open if the problem
is not fixed, the user finds the problem is still there but the bug
reporter does not re-open the bug. All you can do is add a comment and
hope that a developer sees it and re-opens the bug.



Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Mike Frysinger
On Tuesday 25 May 2010 22:58:43 Jeroen Roovers wrote:
> On Tue, 25 May 2010 14:24:05 -0400 Mike Frysinger wrote:
> > On Tuesday 18 May 2010 02:02:01 Andreas K. Huettel wrote:
> > > could you please help the poor bug wranglers a bit?! The queue has
> > > reached 170 unassigned bugs...
> > 
> > people dont seem to realize that bug-wranglers isnt just for
> > re-assigning to the proper maintainer.  they are supposed to be doing
> > basic triage, user feedback, as well as cleaning up the bug.  i
> > shouldnt be seeing bugs assigned to maintainers that have
> > "${PN}-1.ebuild" as their subject, nor bugs that lack basic things
> > like `emerge --info` or build logs.
> 
> Without looking at the rest of this whole thread, that's lucky lucky
> wishful thinking. It doesn't match reality. Hope you're OK with that.

reality matches what i said.  bug wranglers *are supposed to be*.  if reality 
matched what i wanted, then i wouldnt have written an e-mail of "supposed to 
be" and "shouldnt be" statements.
-mike


signature.asc
Description: This is a digitally signed message part.


Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Jeroen Roovers
On Wed, 26 May 2010 05:02:10 +0200
Jeroen Roovers  wrote:

> http://www.gentoo.org/proj/en/qa/bug-wranglers/index.xml
> 
> If you find that you can't do that, then we should be sorry for the
> maintainers. Bug wrangling is the most unthankful job you can
> voluntarily perform, so there.

To follow up on that, we could add some language to the point that if
it takes more than X days (say a week) then assignment to maintainers
proper should go through. This could be especially important when bugs
do seem relevant but should be looked at by someone (or a freakin' whole
"herd") with somewhat more intimate knowledge of the package in
question.


Regards,
 jer



Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Jeroen Roovers
On Tue, 25 May 2010 23:40:44 +0200
Harald van Dijk  wrote:

> Yes, people like myself who don't normally wrangle bugs but try to
> help out occasionally. I'm not really interested in receiving all bug
> wrangler e-mails.

Nobody should be required to read all that crap. :)


  jer



Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Jeroen Roovers
On Tue, 25 May 2010 22:08:55 +0200
Harald van Dijk  wrote:

> On Tue, May 25, 2010 at 03:33:33PM -0400, Mike Frysinger wrote:
> > On Tuesday 25 May 2010 14:46:01 Matti Bickel wrote:
> > > I wrangle bugs when there's a need and I'd
> > > like to hear what maintainers want to see on a bug assigned to
> > > them. If info is missing I usually ask for it and assign the bug
> > > anyway. If that's not wanted, let me know.
> > 
> > i dont feel like this should go to the maintainer yet.  if a report
> > is missing something that the maintainer needs, it isnt ready for
> > them to look at.  so wranglers ask for it, leave it assigned to
> > bug-wranglers, and close as NEEDINFO.  when (if) things become
> > available, it can then be re-opened and moved to the maintainer.
> 
> No, don't close as NEEDINFO, mark as ASSIGNED. NEEDINFO bugs cannot be
> reopened by other users, even if they provide the requested
> information. NEEDINFO bugs are also easily forgotten about when the
> reporter forgets to reopen the bug him/herself. Plus, it's in the
> docs anyway.

If you have to CLOSE a bug because it's blatantly lacking the proper
info, language, spelling etc, then CC yourself. It's the least you can
do to make sure there's some kind of follow-up.


Regards,
  jer



Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Jeroen Roovers
On Tue, 25 May 2010 20:46:01 +0200
Matti Bickel  wrote:

> On 05/25/2010 08:24 PM, Mike Frysinger wrote:
> > they are supposed to be doing basic triage, user feedback
> 
> Can you be more specific? I wrangle bugs when there's a need and I'd
> like to hear what maintainers want to see on a bug assigned to them.
> If info is missing I usually ask for it and assign the bug anyway. If
> that's not wanted, let me know.
> 

http://www.gentoo.org/proj/en/qa/bug-wranglers/index.xml

If you find that you can't do that, then we should be sorry for the
maintainers. Bug wrangling is the most unthankful job you can
voluntarily perform, so there.


Regards an happy wrangling,
 jer



Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Jeroen Roovers
On Tue, 25 May 2010 14:24:05 -0400
Mike Frysinger  wrote:

> On Tuesday 18 May 2010 02:02:01 Andreas K. Huettel wrote:
> > could you please help the poor bug wranglers a bit?! The queue has
> > reached 170 unassigned bugs...
> 
> people dont seem to realize that bug-wranglers isnt just for
> re-assigning to the proper maintainer.  they are supposed to be doing
> basic triage, user feedback, as well as cleaning up the bug.  i
> shouldnt be seeing bugs assigned to maintainers that have
> "${PN}-1.ebuild" as their subject, nor bugs that lack basic things
> like `emerge --info` or build logs. -mike

Without looking at the rest of this whole thread, that's lucky lucky
wishful thinking. It doesn't match reality. Hope you're OK with that.


Regards,
 jer



Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Mike Frysinger
On Tuesday 25 May 2010 17:06:39 Matti Bickel wrote:
> On 05/25/2010 10:08 PM, Harald van Dijk wrote:
> > NEEDINFO bugs cannot be reopened by other users, even if they provide
> > the requested information.
> 
> I utterly fail at finding documentation on that. I've recently hit a
> problem where a user couldn't reopen a RESO FIXED bug, too. Are bugzi
> permissions documented somewhere?

the bug reporter can open their own bugs.  gentoo developers can open any bug.  
that's about it.

(there are other rules minor, but they arent really relevant)
-mike


signature.asc
Description: This is a digitally signed message part.


Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Harald van Dijk
On Tue, May 25, 2010 at 04:25:20PM -0400, Mike Frysinger wrote:
> and people on the wrangler alias see that traffic, so the state doesnt 
> matter.  
> but i guess you're trying to cater to people who only scan the assigned list 
> rather than watching the e-mails sent to it.

Yes, people like myself who don't normally wrangle bugs but try to help out
occasionally. I'm not really interested in receiving all bug wrangler
e-mails.



Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Matti Bickel
On 05/25/2010 10:08 PM, Harald van Dijk wrote:
> NEEDINFO bugs cannot be reopened by other users, even if they provide
> the requested information.

I utterly fail at finding documentation on that. I've recently hit a
problem where a user couldn't reopen a RESO FIXED bug, too. Are bugzi
permissions documented somewhere?



signature.asc
Description: OpenPGP digital signature


Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Mike Frysinger
On Tuesday 25 May 2010 16:08:55 Harald van Dijk wrote:
> On Tue, May 25, 2010 at 03:33:33PM -0400, Mike Frysinger wrote:
> > On Tuesday 25 May 2010 14:46:01 Matti Bickel wrote:
> > > I wrangle bugs when there's a need and I'd
> > > like to hear what maintainers want to see on a bug assigned to them.
> > > If info is missing I usually ask for it and assign the bug anyway. If
> > > that's not wanted, let me know.
> > 
> > i dont feel like this should go to the maintainer yet.  if a report is
> > missing something that the maintainer needs, it isnt ready for them to
> > look at.  so wranglers ask for it, leave it assigned to bug-wranglers,
> > and close as NEEDINFO.  when (if) things become available, it can then
> > be re-opened and moved to the maintainer.
> 
> No, don't close as NEEDINFO, mark as ASSIGNED. NEEDINFO bugs cannot be
> reopened by other users, even if they provide the requested information.
> NEEDINFO bugs are also easily forgotten about when the reporter forgets
> to reopen the bug him/herself.

and people on the wrangler alias see that traffic, so the state doesnt matter.  
but i guess you're trying to cater to people who only scan the assigned list 
rather than watching the e-mails sent to it.

> Plus, it's in the docs anyway.

then i guess i'll point this out to those who wrangle these things to 
maintainers w/out following the docs.  thanks.
-mike


signature.asc
Description: This is a digitally signed message part.


Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Matti Bickel
On 05/25/2010 10:08 PM, Harald van Dijk wrote:
>   http://www.gentoo.org/proj/en/qa/bug-wranglers/index.xml

Cool, I clearly not up to date here, I've never thought this to be a
project. Thanks for the link.

Wrt mentioning metadata.xml for herd lookup in there: I've found
willikins' meta -v (in a query, mind you) saves me a few keystrokes.



signature.asc
Description: OpenPGP digital signature


Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Harald van Dijk
On Tue, May 25, 2010 at 03:33:33PM -0400, Mike Frysinger wrote:
> On Tuesday 25 May 2010 14:46:01 Matti Bickel wrote:
> > I wrangle bugs when there's a need and I'd
> > like to hear what maintainers want to see on a bug assigned to them.
> > If info is missing I usually ask for it and assign the bug anyway. If
> > that's not wanted, let me know.
> 
> i dont feel like this should go to the maintainer yet.  if a report is 
> missing 
> something that the maintainer needs, it isnt ready for them to look at.  so 
> wranglers ask for it, leave it assigned to bug-wranglers, and close as 
> NEEDINFO.  when (if) things become available, it can then be re-opened and 
> moved to the maintainer.

No, don't close as NEEDINFO, mark as ASSIGNED. NEEDINFO bugs cannot be
reopened by other users, even if they provide the requested information.
NEEDINFO bugs are also easily forgotten about when the reporter forgets
to reopen the bug him/herself. Plus, it's in the docs anyway.

"Do not assume that the reporter ought to know how to report bugs. An
 omitted `emerge --info' does not call for a public flogging, it simply
 calls for the missing `emerge --info'. Even experienced reporters make
 mistakes, so simply request the information, mark the bug as ASSIGNED
 and wait for the information you requested."

  http://www.gentoo.org/proj/en/qa/bug-wranglers/index.xml



Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Matti Bickel
On 05/25/2010 09:33 PM, Mike Frysinger wrote:
> i posted some specific examples already ...

Sure enough. Just wanted to know if there's more to it than build.log
and emerge --info. I'll try to extract something more than that next
time. Goes w/o saying that bug cleanup should be done prior
to assignments.

> so wranglers ask for it, leave it assigned to bug-wranglers, and
> close as NEEDINFO.  when (if) things become available, it can then be
> re-opened and moved to the maintainer. -mike

Good enough for me. The intention behind the immediate assignments was
the hope that maintainers could figure out the error from the already
provided info ("does not work" won't get assigned), netting quicker
reaction times.



signature.asc
Description: OpenPGP digital signature


Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Mike Frysinger
On Tuesday 25 May 2010 14:46:01 Matti Bickel wrote:
> On 05/25/2010 08:24 PM, Mike Frysinger wrote:
> > they are supposed to be doing basic triage, user feedback
> 
> Can you be more specific?

i posted some specific examples already ... a little double checking of 
reports and suggestion of basic debugging steps (strace/whatever) would be 
nice, but i can let that slide since this is a bit more package specific.

> I wrangle bugs when there's a need and I'd
> like to hear what maintainers want to see on a bug assigned to them.
> If info is missing I usually ask for it and assign the bug anyway. If
> that's not wanted, let me know.

i dont feel like this should go to the maintainer yet.  if a report is missing 
something that the maintainer needs, it isnt ready for them to look at.  so 
wranglers ask for it, leave it assigned to bug-wranglers, and close as 
NEEDINFO.  when (if) things become available, it can then be re-opened and 
moved to the maintainer.
-mike


signature.asc
Description: This is a digitally signed message part.


Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Richard Freeman

On 05/25/2010 02:24 PM, Mike Frysinger wrote:

On Tuesday 18 May 2010 02:02:01 Andreas K. Huettel wrote:

could you please help the poor bug wranglers a bit?! The queue has reached
170 unassigned bugs...


people dont seem to realize that bug-wranglers isnt just for re-assigning to
the proper maintainer.  they are supposed to be doing basic triage, user
feedback, as well as cleaning up the bug.  i shouldnt be seeing bugs assigned
to maintainers that have "${PN}-1.ebuild" as their subject, nor bugs that lack
basic things like `emerge --info` or build logs.


As long as the status quo works I don't think we need to change it. 
However if we are running into issues with keeping up it might make 
sense to just have wranglers do assignments to maintainers and let the 
maintainers deal with the rest.


The reason is that the maintainer might be able to spot dups much more 
readily, or spot obvious solutions to bugs, where a wrangler might be 
hunting around.


By all means wranglers should do what they can when they can, but keep 
in mind that if you yell at a wrangler any time they "do it wrong" a 
natural response of devs will be to not bother with bug wrangling or 
just looking for their own bugs in the list.


I'm not necessarily proposing any changes here, but in general we need 
to be careful about barriers to entry in projects that are undermanned.


Rich



Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Matti Bickel
On 05/25/2010 08:24 PM, Mike Frysinger wrote:
> they are supposed to be doing basic triage, user feedback

Can you be more specific? I wrangle bugs when there's a need and I'd
like to hear what maintainers want to see on a bug assigned to them.
If info is missing I usually ask for it and assign the bug anyway. If
that's not wanted, let me know.



signature.asc
Description: OpenPGP digital signature


Re: [gentoo-dev] bug wrangler queue is large...

2010-05-25 Thread Mike Frysinger
On Tuesday 18 May 2010 02:02:01 Andreas K. Huettel wrote:
> could you please help the poor bug wranglers a bit?! The queue has reached
> 170 unassigned bugs...

people dont seem to realize that bug-wranglers isnt just for re-assigning to 
the proper maintainer.  they are supposed to be doing basic triage, user 
feedback, as well as cleaning up the bug.  i shouldnt be seeing bugs assigned 
to maintainers that have "${PN}-1.ebuild" as their subject, nor bugs that lack 
basic things like `emerge --info` or build logs.
-mike


signature.asc
Description: This is a digitally signed message part.


[gentoo-dev] bug wrangler queue is large...

2010-05-17 Thread Andreas K. Huettel

Hi everyone, 

could you please help the poor bug wranglers a bit?! The queue has reached 170 
unassigned bugs...

Thanks a lot, 
dilfridge

-- 
Dr. Andreas K. Huettel
Institute for Experimental and Applied Physics
University of Regensburg
D-93040 Regensburg
Germany

tel. +49 151 241 67748 (mobile)
e-mail m...@akhuettel.de
http://www.akhuettel.de/research/



signature.asc
Description: This is a digitally signed message part.