A few months ago, Alexandre started including lists of fixed bugs in his release announcements. See e.g. http://www.winehq.org/announce/1.1.15, which says
Bugs fixed in 1.1.15: 5694 Lionhead Black & White 2 demo crashes 7014 Unhandled page fault when exiting Commandos - BEL 7297 MIDI in/out fails, ports recognized ...
I rather like those lists. He generates them by simply querying for open but fixed bugs right when he does the release; he then closes the bugs. They're good for PR if nothing else.
For this to work, we have to let Alexandre be the only person to close fixed bugs. (Of course, it's ok for other people to mark bugs fixed, and it's ok for anyone to close invalid or dup bugs.) If anybody else closes a fixed bug, it won't show up in the list of fixed bugs for any release.
A corner case is when a bug is marked fixed sometime after release. I think it's ok to leave it for Alexandre to close; that way it shows up in at least one release announcement. Even if it shows up a few releases late, that's better than not showing up at all.
Sound reasonable?
I ask because Vitaliy closed http://bugs.winehq.org/show_bug.cgi?id=10923 even after I tried to explain this to him, so I figured it needed a wider discussion. - Dan
Dan Kegel wrote:
A few months ago, Alexandre started including lists of fixed bugs in his release announcements. See e.g. http://www.winehq.org/announce/1.1.15, which says
Bugs fixed in 1.1.15: 5694 Lionhead Black & White 2 demo crashes 7014 Unhandled page fault when exiting Commandos - BEL 7297 MIDI in/out fails, ports recognized ...
I rather like those lists. He generates them by simply querying for open but fixed bugs right when he does the release; he then closes the bugs. They're good for PR if nothing else.
IMHO there are always enough bugs fixed for each release in buzilla. Agreed meaning for "closed" status was "the fix present in Wine release".
If bug reporter wasn't fast enough to report bug fixed in the 5 minutes elapsed between AJ pushing new release tag and wholesale closing of all fixed bugs in bugzilla does that mean we have to wait for 2 more weeks to close such bugs? Same applies to people closing their bugs themselves after verifying the fix in latest git. If they not supposed to do that, then prohibit anyone except AJ from closing bugs.
Vitaliy.
On Sat, Feb 14, 2009 at 5:59 PM, Vitaliy Margolen wine-devel@kievinfo.com wrote:
Agreed meaning for "closed" status was "the fix present in Wine release".
Right, agreed.
If bug reporter wasn't fast enough to report bug fixed in the 5 minutes elapsed between AJ pushing new release tag and wholesale closing of all fixed bugs in bugzilla...
? Not sure what you're saying there. Nobody has to act unreasonably fast. They just have to mark bugs fixed when they're fixed in git.
does that mean we have to wait for 2 more weeks to close such bugs?
Exactly, I'm proposing letting them sit there, fixed but open, until the next release, so Alexandre doesn't have to change the simple way he generates his list of fixed bugs.
[Aren't people allowed to close] their bugs themselves after verifying the fix in latest git?
They're encouraged to mark them fixed once the fix is in git. As you pointed out, 'closed' is for when the fix is in a release.
If they not supposed to do that, then prohibit anyone except AJ from closing bugs.
It seems overkill to modify bugzilla for this, since there's no reason to prohibit closing of invalid or duplicate bugs, and changing bugzilla is hardish.
I just kind of like the simple model of "we fix 'em, Alexandre closes 'em", and was hoping we could agree to follow it. - Dan
Dan Kegel wrote:
On Sat, Feb 14, 2009 at 5:59 PM, Vitaliy Margolen wine-devel@kievinfo.com wrote:
If they not supposed to do that, then prohibit anyone except AJ from closing bugs.
It seems overkill to modify bugzilla for this, since there's no reason to prohibit closing of invalid or duplicate bugs, and changing bugzilla is hardish.
+1, anyone should be able to close invalid, duplicate or otherwise erroneous bug reports. Only AJ should be closing fixed bug reports. It should be permissible to mark a Fixed bug as Resolved, however.
James McKenzie
On Saturday 14 February 2009, Dan Kegel wrote:
A few months ago, Alexandre started including lists of fixed bugs in his release announcements. See e.g. http://www.winehq.org/announce/1.1.15, which says
Bugs fixed in 1.1.15: 5694 Lionhead Black & White 2 demo crashes 7014 Unhandled page fault when exiting Commandos - BEL 7297 MIDI in/out fails, ports recognized ...
I rather like those lists. He generates them by simply querying for open but fixed bugs right when he does the release; he then closes the bugs. They're good for PR if nothing else.
For this to work, we have to let Alexandre be the only person to close fixed bugs. (Of course, it's ok for other people to mark bugs fixed, and it's ok for anyone to close invalid or dup bugs.) If anybody else closes a fixed bug, it won't show up in the list of fixed bugs for any release.
A corner case is when a bug is marked fixed sometime after release. I think it's ok to leave it for Alexandre to close; that way it shows up in at least one release announcement. Even if it shows up a few releases late, that's better than not showing up at all.
Sound reasonable?
I ask because Vitaliy closed http://bugs.winehq.org/show_bug.cgi?id=10923 even after I tried to explain this to him, so I figured it needed a wider discussion.
- Dan
Couldn't you just query Bugzilla for all the bugs closed between $release and $lastrelease?
- Neil
-
Dan Kegel -
James McKenzie -
Neil Skrypuch -
Vitaliy Margolen