--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
12-08-2010, 11:01 AM
Bastien Nocera
Proposed package blocking due to FTBFS
On Wed, 2010-12-08 at 00:50 -0800, Adam Williamson wrote:
> On Wed, 2010-12-08 at 01:05 +0000, Bastien Nocera wrote:
>
> > And I'll go back to fixing actual bugs encountered by people instead of
> > random bot-driven bugs.
>
> every abrt report, ever, is an actual bug encountered by an actual
> person. They have to be sufficiently narked about the app crashing (and
> it really must have crashed) to click through a rather convoluted
> process (the first time, anyway) to send in a report.
Given the time it takes triage them, compared to how long it takes to
file them, I'm not sure it's a win for us.
> so are all these bugs, for that matter: they're actual bugs encountered
> by Matt. The package failing to build is clearly a bug. Matt tried to
> build it and so encountered the bug. Where does it fail to meet your
> criteria?
It's a file'n'dump bug. There's no one that actually looked at the bugs
to try and analyse them, nobody to offer a reminder in the bugs (they
were filed and left untouched).
> I agree it's a bit questionable whether we should block packages for
> FTBFS, but the argument can clearly be made; being self-hosting is
> obviously important for an F/OSS project. At some point it devolves into
> Stallmanite wankery about whether you can flash your mouse, but where
> exactly we should draw the line isn't a slam-dunk
I'm not sure what this has to do with anything. The signal to noise
ratio in the RH bugzilla is far too low to be anything useful, and
piling another bug on top of other bugs, with no reminder apart from
this mail is rude.
FWIW, the bug I found in gnokii was really a bug in pcsc which just
removed a enum member without bumping soname, thus breaking API, and
possibly ABI. I should have received an automated mail about broken
dependencies, and not be having discussions about the quality of our
bugzilla.
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
12-08-2010, 12:22 PM
Matt Domsch
Proposed package blocking due to FTBFS
On Wed, Dec 08, 2010 at 12:01:39PM +0000, Bastien Nocera wrote:
> On Wed, 2010-12-08 at 00:50 -0800, Adam Williamson wrote:
> > On Wed, 2010-12-08 at 01:05 +0000, Bastien Nocera wrote:
> >
> > > And I'll go back to fixing actual bugs encountered by people instead of
> > > random bot-driven bugs.
> >
> > every abrt report, ever, is an actual bug encountered by an actual
> > person. They have to be sufficiently narked about the app crashing (and
> > it really must have crashed) to click through a rather convoluted
> > process (the first time, anyway) to send in a report.
>
> Given the time it takes triage them, compared to how long it takes to
> file them, I'm not sure it's a win for us.
>
> > so are all these bugs, for that matter: they're actual bugs encountered
> > by Matt. The package failing to build is clearly a bug. Matt tried to
> > build it and so encountered the bug. Where does it fail to meet your
> > criteria?
>
> It's a file'n'dump bug. There's no one that actually looked at the bugs
> to try and analyse them, nobody to offer a reminder in the bugs (they
> were filed and left untouched).
>
> > I agree it's a bit questionable whether we should block packages for
> > FTBFS, but the argument can clearly be made; being self-hosting is
> > obviously important for an F/OSS project. At some point it devolves into
> > Stallmanite wankery about whether you can flash your mouse, but where
> > exactly we should draw the line isn't a slam-dunk
>
> I'm not sure what this has to do with anything. The signal to noise
> ratio in the RH bugzilla is far too low to be anything useful, and
> piling another bug on top of other bugs, with no reminder apart from
> this mail is rude.
I'm confused. You want reminders filed in the bugs, but then you say
the S-to-N ratio is to low to be useful.
I could add automatic reminders in bugzilla, but I don't think that
solves your key concern.
The packages I posted last night were since F12 only. Personally, I'd
like to see all FTBFS since even F14 fixed (they all have bugs filed).
--
Matt Domsch
Technology Strategist
Dell | Office of the CTO
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
12-08-2010, 01:32 PM
Bruno Wolff III
Proposed package blocking due to FTBFS
On Wed, Dec 08, 2010 at 12:01:39 +0000,
Bastien Nocera <bnocera@redhat.com> wrote:
>
> It's a file'n'dump bug. There's no one that actually looked at the bugs
> to try and analyse them, nobody to offer a reminder in the bugs (they
> were filed and left untouched).
I went through a number of FTBFS bugs for other people's packages. A few
other volunteers did as well. We didn't get full coverage, but we did
help some of them get fixed.
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
12-08-2010, 02:52 PM
Jon Masters
Proposed package blocking due to FTBFS
On Tue, 2010-12-07 at 20:29 -0600, Matt Domsch wrote:
> My goal isn't to make life difficult for everyone. My goal is to keep
> the distribution in a form where it can actually build from the open
> source we provide.
Thanks Matt. What you're doing is vitally important for the
distribution, since it should build from source always. You do a lot of
great work in this area and I hope you continue for a long time!
Jon.
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
--
in your fear, seek only peace
in your fear, seek only love
-d. bowie
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
12-08-2010, 11:43 PM
Toshio Kuratomi
Proposed package blocking due to FTBFS
On Wed, Dec 08, 2010 at 11:48:26AM +0000, Peter Robinson wrote:
> On Wed, Dec 8, 2010 at 8:50 AM, Adam Williamson <awilliam@redhat.com> wrote:
> > so are all these bugs, for that matter: they're actual bugs encountered
> > by Matt. The package failing to build is clearly a bug. Matt tried to
> > build it and so encountered the bug. Where does it fail to meet your
> > criteria?
> >
> > I agree it's a bit questionable whether we should block packages for
> > FTBFS, but the argument can clearly be made; being self-hosting is
> > obviously important for an F/OSS project. At some point it devolves into
> > Stallmanite wankery about whether you can flash your mouse, but where
> > exactly we should draw the line isn't a slam-dunk
>
> I'm sitting on the fence on this one. There are packages built on F-12
> that work perfectly well on rawhide that don't build on rawhide. What
> about an instance where there's dependant packages. Do they
> automatically get blocked too or do we go through another route of
> FTBFS on those too?
>
Yes, they should get automatically blocked too.
> In the case of a leaf one it might be that by it
> not building currently doesn't affect anything and the maintainer is
> aware of the problem but needs the time to fix the issue properly when
> he gets time. In this case the maintainer then has to jump through the
> review process all over again to get it unblocked and then will likely
> just not be bothered.
They shouldn't have to go through a re-review unless they've let the package
sit in retirement for (I believe it's six months but someone else might have
the policy URL handy).
-Toshio
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
12-09-2010, 08:22 AM
Peter Robinson
Proposed package blocking due to FTBFS
On Thu, Dec 9, 2010 at 12:43 AM, Toshio Kuratomi <a.badger@gmail.com> wrote:
> On Wed, Dec 08, 2010 at 11:48:26AM +0000, Peter Robinson wrote:
>> On Wed, Dec 8, 2010 at 8:50 AM, Adam Williamson <awilliam@redhat.com> wrote:
>> > so are all these bugs, for that matter: they're actual bugs encountered
>> > by Matt. The package failing to build is clearly a bug. Matt tried to
>> > build it and so encountered the bug. Where does it fail to meet your
>> > criteria?
>> >
>> > I agree it's a bit questionable whether we should block packages for
>> > FTBFS, but the argument can clearly be made; being self-hosting is
>> > obviously important for an F/OSS project. At some point it devolves into
>> > Stallmanite wankery about whether you can flash your mouse, but where
>> > exactly we should draw the line isn't a slam-dunk
>>
>> I'm sitting on the fence on this one. There are packages built on F-12
>> that work perfectly well on rawhide that don't build on rawhide. What
>> about an instance where there's dependant packages. Do they
>> automatically get blocked too or do we go through another route of
>> FTBFS on those too?
>>
> Yes, they should get automatically blocked too.
>
>> In the case of a leaf one it might be that by it
>> not building currently doesn't affect anything and the maintainer is
>> aware of the problem but needs the time to fix the issue properly when
>> he gets time. In this case the maintainer then has to jump through the
>> review process all over again to get it unblocked and then will likely
>> just not be bothered.
>
> They shouldn't have to go through a re-review unless they've let the package
> sit in retirement for (I believe it's six months but someone else might have
> the policy URL handy).
My understanding was that if it was blocked it had to go through review again.
Peter
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
12-09-2010, 01:30 PM
"Jason L Tibbitts III"
Proposed package blocking due to FTBFS
>>>>> "PR" == Peter Robinson <pbrobinson@gmail.com> writes:
PR> My understanding was that if it was blocked it had to go through
PR> review again.
Depends on how long:
http://fedoraproject.org/wiki/Orphaned_package_that_need_new_maintainers
"
Re-review required for older packages
If a package was last updated more than three months ago (running git
log *.spec can show you this information), you will need to submit a
review request (a new bugzilla ticket) and have the package approved by
a reviewer as if it were new to Fedora. See the package review process
for more information. There are a couple of small changes though, be
sure to submit a 'update' request to the SCM, and before you will be
able to run the final 'make build' commands you will need to file a
ticket w/ release engineering to unblock your package
https://fedorahosted.org/rel-eng/newticket
"
I note that you'd probably never find that policy if you didn't already
know what to search for, and that I need to add some additional info
somewhere about how to make the SCM request.
- J<
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
12-10-2010, 11:44 PM
Kevin Kofler
Proposed package blocking due to FTBFS
Matt Domsch wrote:
> Last built on Fedora 12 (52):
Huh?
The right metric is not "when was this last built" but "when was this last
BUILDABLE". We don't randomly rebuild stuff which doesn't need to be
rebuilt.
E.g.:
> celestia-1.5.1-2.fc12 [u'631077 NEW'] (build/make) steve,mmahut
(the first one on that list) is in F14FTBFS, so it should be recorded as
last buildable in F13 / during early F14 development.
Kevin Kofler
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
Proposed package blocking due to FTBFS
