in accordance with the announced Fedora feature[1][2], we (the Boost
maintainers) plan to rebase Boost to 1.47.0 really soon now. Boost
1.47.0 has been released recently and Denis Arnaud kindly did the
packaging, so it is ready for scratch builds, smoke testing, and related
fun.
I'll do that in the next few days and if all comes out green-ish, I'll
push the package into Fedora 16 and write a follow-up to this e-mail
with guidelines on overcoming the obstacles, if any.
Further plans: originally, we hoped that 1.48.0 would be about out by
now, and we would manage to get beta into Fedora 16, much like we did
with 1.46.0 for Fedora 15. But the Boost schedule slipped, and made any
such plans impossible to realize[2]. At the same time, the sentiment of
Boost upstream seems to be to (gradually) get back to the original
schedule, so we may end up with 1.50.0 in Fedora 17. Handling +3 bump
might end up being more interesting than is desirable, so to alleviate
this, we will most probably want to do at least one larger rebase
mid-Rawhide, either to 1.48.0, or 1.49.0. I'll write more when I know
more.
Don't hesitate to ping me on irc (_petr) with any concerns that you
have.
Thanks,
Petr Machata
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
07-19-2011, 07:02 AM
Peter Robinson
boost 1.47.0
On Mon, Jul 18, 2011 at 9:35 PM, Petr Machata <pmachata@redhat.com> wrote:
Hi there,
in accordance with the announced Fedora feature[1][2], we (the Boost
maintainers) plan to rebase Boost to 1.47.0 really soon now. *Boost
1.47.0 has been released recently and Denis Arnaud kindly did the
packaging, so it is ready for scratch builds, smoke testing, and related
fun.
I'll do that in the next few days and if all comes out green-ish, I'll
push the package into Fedora 16 and write a follow-up to this e-mail
with guidelines on overcoming the obstacles, if any.
Further plans: originally, we hoped that 1.48.0 would be about out by
now, and we would manage to get beta into Fedora 16, much like we did
with 1.46.0 for Fedora 15. *But the Boost schedule slipped, and made any
such plans impossible to realize[2]. *At the same time, the sentiment of
Boost upstream seems to be to (gradually) get back to the original
schedule, so we may end up with 1.50.0 in Fedora 17. *Handling +3 bump
might end up being more interesting than is desirable, so to alleviate
this, we will most probably want to do at least one larger rebase
mid-Rawhide, either to 1.48.0, or 1.49.0. *I'll write more when I know
more.
Don't hesitate to ping me on irc (_petr) with any concerns that you
have.
The only concern I have is with jumping to a release later than 1.47 in F-16 post alpha which is in fact the time when features should be complete. So in fact it should have already landed. Post alpha the release will branch from rawhide at which point please feel free to push > 1.48 to F-17 rawhide with appropriate heads up to people.
Peter
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
07-19-2011, 07:15 AM
Kalev Lember
boost 1.47.0
On 07/18/2011 11:35 PM, Petr Machata wrote:
> Hi there,
>
> in accordance with the announced Fedora feature[1][2], we (the Boost
> maintainers) plan to rebase Boost to 1.47.0 really soon now. Boost
> 1.47.0 has been released recently and Denis Arnaud kindly did the
> packaging, so it is ready for scratch builds, smoke testing, and related
> fun.
>
> I'll do that in the next few days and if all comes out green-ish, I'll
> push the package into Fedora 16 and write a follow-up to this e-mail
> with guidelines on overcoming the obstacles, if any.
How are you planning to handle the rebuilds for all the 176 affected
source packages? Is there going to be a koji side tag for rebuilds?
--
Kalev
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
07-19-2011, 10:27 AM
Petr Machata
boost 1.47.0
Kalev Lember <kalevlember@gmail.com> writes:
> On 07/18/2011 11:35 PM, Petr Machata wrote:
>> in accordance with the announced Fedora feature[1][2], we (the Boost
>> maintainers) plan to rebase Boost to 1.47.0 really soon now. Boost
>> 1.47.0 has been released recently and Denis Arnaud kindly did the
>> packaging, so it is ready for scratch builds, smoke testing, and related
>> fun.
>>
>> I'll do that in the next few days and if all comes out green-ish, I'll
>> push the package into Fedora 16 and write a follow-up to this e-mail
>> with guidelines on overcoming the obstacles, if any.
>
> How are you planning to handle the rebuilds for all the 176 affected
> source packages? Is there going to be a koji side tag for rebuilds?
The last time around maintainers were largely able to respin their
packages themselves when given a notice. What I did in the past, and
plan an doing this time again, is that I go through the list, rebuilding
the packages in mock. If bugs occur, I fix them either in Boost, or in
the package itself, in which case I open FTBFS with a patch attached. I
don't have privileges to bump, rebuild and patch the packages myself.
Thanks,
PM
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
07-19-2011, 10:43 AM
Petr Machata
boost 1.47.0
Peter Robinson <pbrobinson@gmail.com> writes:
> On Mon, Jul 18, 2011 at 9:35 PM, Petr Machata <[1]pmachata@redhat.com>
> wrote:
>
> in accordance with the announced Fedora feature[1][2], we (the Boost
> maintainers) plan to rebase Boost to 1.47.0 really soon now. Boost
> 1.47.0 has been released recently and Denis Arnaud kindly did the
> packaging, so it is ready for scratch builds, smoke testing, and related
> fun.
>
> Don't hesitate to ping me on irc (_petr) with any concerns that you
> have.
>
> The only concern I have is with jumping to a release later than 1.47 in
> F-16 post alpha which is in fact the time when features should be
> complete. So in fact it should have already landed. Post alpha the release
> will branch from rawhide at which point please feel free to push > 1.48 to
> F-17 rawhide with appropriate heads up to people.
We definitely won't bump again in F16 cycle. What is likely to happen
is series of isolated patches that we backport from future releases, as
requested in bug reports or otherwise. We might consider pushing 1.47.1
if it turns up, but that largely depends on whether it would be humanly
possible to review the patch-set for ABI breakages.
PM
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
07-20-2011, 11:09 AM
Petr Machata
boost 1.47.0
Petr Machata <pmachata@redhat.com> writes:
> in accordance with the announced Fedora feature[1][2], we (the Boost
> maintainers) plan to rebase Boost to 1.47.0 really soon now. [...]
>
> I'll do that in the next few days and if all comes out green-ish, I'll
> push the package into Fedora 16 and write a follow-up to this e-mail
> with guidelines on overcoming the obstacles, if any.
It all seems good, I rebuilt 25 packages and hit no boost-related
problems. That means that boost::filesystem v2 has still not been
obsoleted. According to the documentation, this will happen in version
1.48.0, which would thus make a good candidate for early rebase in
Fedora 17 rawhide.
I'm now building Boost 1.47.0 in rawhide for Fedora 16 rawhide. This
involves soname bump, as usual, and so maintainers of dependent packages
need to rebuild. The full list has been posted by Kalev Lember to this
very thread.
Thanks,
PM
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
07-21-2011, 12:51 PM
"Jon Ciesla"
boost 1.47.0
> Hi there,
>
> in accordance with the announced Fedora feature[1][2], we (the Boost
> maintainers) plan to rebase Boost to 1.47.0 really soon now. Boost
> 1.47.0 has been released recently and Denis Arnaud kindly did the
> packaging, so it is ready for scratch builds, smoke testing, and related
> fun.
>
> I'll do that in the next few days and if all comes out green-ish, I'll
> push the package into Fedora 16 and write a follow-up to this e-mail
> with guidelines on overcoming the obstacles, if any.
>
> Further plans: originally, we hoped that 1.48.0 would be about out by
> now, and we would manage to get beta into Fedora 16, much like we did
> with 1.46.0 for Fedora 15. But the Boost schedule slipped, and made any
> such plans impossible to realize[2]. At the same time, the sentiment of
> Boost upstream seems to be to (gradually) get back to the original
> schedule, so we may end up with 1.50.0 in Fedora 17. Handling +3 bump
> might end up being more interesting than is desirable, so to alleviate
> this, we will most probably want to do at least one larger rebase
> mid-Rawhide, either to 1.48.0, or 1.49.0. I'll write more when I know
> more.
>
> Don't hesitate to ping me on irc (_petr) with any concerns that you
> have.
>
> [1] https://fedoraproject.org/wiki/Features/F16Boost147
> [2] https://bugzilla.redhat.com/show_bug.cgi?id=711845
Wesnoth seems not to like the new Boost all that well:
Is this a common sort of error? My Boost-fu is weak, so there might be
something obvious I'm missing.
-J
> Thanks,
> Petr Machata
> --
> 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
07-21-2011, 07:48 PM
Petr Machata
boost 1.47.0
"Jon Ciesla" <limb@jcomserv.net> writes:
> Wesnoth seems not to like the new Boost all that well:
>
> http://koji.fedoraproject.org/koji/getfile?taskID=3218720&name=build.log
>
> Is this a common sort of error? My Boost-fu is weak, so there might be
> something obvious I'm missing.
Yeah, all boost errors tend to spew lines and lines of error messages
like that. That's the nature of them templetes. I'm looking into this
one, and it seems to be some interplay between BOOST_FOREACH,
boost:tr_vector and GCC version. Not sure what exactly yet.
It can be worked around by replacing the "foreach" calls in
wesnoth-1.8.6/src/gui/widgets/tree_view_node.cpp in this manner:
- foreach(const ttree_view_node& node, children_) {
+ for (boost:tr_vector<ttree_view_node>::const_iterato r it
+ = children_.begin (); it != children_.end (); ++it) {
+ const ttree_view_node& node = *it;
Thanks,
PM
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
07-21-2011, 08:13 PM
Petr Machata
boost 1.47.0
Petr Machata <pmachata@redhat.com> writes:
> "Jon Ciesla" <limb@jcomserv.net> writes:
>
>> Wesnoth seems not to like the new Boost all that well:
>>
>> http://koji.fedoraproject.org/koji/getfile?taskID=3218720&name=build.log
>>
>> Is this a common sort of error? My Boost-fu is weak, so there might be
>> something obvious I'm missing.
>
> Yeah, all boost errors tend to spew lines and lines of error messages
> like that. That's the nature of them templetes. I'm looking into this
> one, and it seems to be some interplay between BOOST_FOREACH,
> boost:tr_vector and GCC version. Not sure what exactly yet.
... and boost::noncopyable. Forgot about that one.
PM
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel
07-21-2011, 09:55 PM
Petr Machata
boost 1.47.0
Petr Machata <pmachata@redhat.com> writes:
> "Jon Ciesla" <limb@jcomserv.net> writes:
>
>> Wesnoth seems not to like the new Boost all that well:
>>
>> http://koji.fedoraproject.org/koji/getfile?taskID=3218720&name=build.log
>>
>> Is this a common sort of error? My Boost-fu is weak, so there might be
>> something obvious I'm missing.
>
> I'm looking into this one, and it seems to be some interplay between
> BOOST_FOREACH, boost:tr_vector and GCC version. Not sure what
> exactly yet.
Opened a bugzilla to track this FTBFS. With GCC 4.6+, BOOST_FOREACH
cannot be used to iterate over noncopyable collections.
boost 1.47.0
tr_vector and GCC version. Not sure what exactly yet.
