[Qemu-devel] Any QEMU 2.8 release blockers?

[Qemu-devel] Any QEMU 2.8 release blockers?

[Stefan Hajnoczi]

Hi folks,
How is QEMU 2.8.0-rc3 holding up?

I'm not aware of release blockers.  QEMU 2.8.0 will be released on
Tuesday, December 13th if nothing comes up.  It will be the -rc3
source tree.

Stefan

Re: [Qemu-devel] Any QEMU 2.8 release blockers?

[Stefan Hajnoczi]

On Fri, Dec 9, 2016 at 3:34 PM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
> Hi folks,
> How is QEMU 2.8.0-rc3 holding up?
>
> I'm not aware of release blockers.  QEMU 2.8.0 will be released on
> Tuesday, December 13th if nothing comes up.  It will be the -rc3
> source tree.

There have been no release blockers so I plan to tag QEMU 2.8.0 today
(December 13th at 19:00 UTC).

If something critical turns up please notify me on IRC (stefanha on
#qemu irc.oftc.net).

Thanks,
Stefan

Re: [Qemu-devel] Any QEMU 2.8 release blockers?

[Stefan Hajnoczi]

On Tue, Dec 13, 2016 at 10:01 AM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
> On Fri, Dec 9, 2016 at 3:34 PM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
>> Hi folks,
>> How is QEMU 2.8.0-rc3 holding up?
>>
>> I'm not aware of release blockers.  QEMU 2.8.0 will be released on
>> Tuesday, December 13th if nothing comes up.  It will be the -rc3
>> source tree.
>
> There have been no release blockers so I plan to tag QEMU 2.8.0 today
> (December 13th at 19:00 UTC).
>
> If something critical turns up please notify me on IRC (stefanha on
> #qemu irc.oftc.net).

Mike Roth discovered a migration compatibility bug introduced by
commit d1b4259f1ab18af24e6a297edb6a8f71691f3256 ("virtio-bus: Plug
devices after features are negotiated").  The release will be delayed
until this is resolved.

Please check the mailing list for discussion.  Mike is sending out a
mail summarizing the bug.

Stefan

Re: [Qemu-devel] Any QEMU 2.8 release blockers?

[Michael Roth]

Quoting Stefan Hajnoczi (2016-12-13 15:04:24)
> On Tue, Dec 13, 2016 at 10:01 AM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
> > On Fri, Dec 9, 2016 at 3:34 PM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
> >> Hi folks,
> >> How is QEMU 2.8.0-rc3 holding up?
> >>
> >> I'm not aware of release blockers.  QEMU 2.8.0 will be released on
> >> Tuesday, December 13th if nothing comes up.  It will be the -rc3
> >> source tree.
> >
> > There have been no release blockers so I plan to tag QEMU 2.8.0 today
> > (December 13th at 19:00 UTC).
> >
> > If something critical turns up please notify me on IRC (stefanha on
> > #qemu irc.oftc.net).
> 
> Mike Roth discovered a migration compatibility bug introduced by
> commit d1b4259f1ab18af24e6a297edb6a8f71691f3256 ("virtio-bus: Plug
> devices after features are negotiated").  The release will be delayed
> until this is resolved.
> 
> Please check the mailing list for discussion.  Mike is sending out a
> mail summarizing the bug.

I've posted a summary in the original patch thread:

  https://lists.gnu.org/archive/html/qemu-devel/2016-12/msg01862.html

> 
> Stefan
> 

Re: [Qemu-devel] Any QEMU 2.8 release blockers?

[Stefan Weil]

On 12/13/16 22:04, Stefan Hajnoczi wrote:
> On Tue, Dec 13, 2016 at 10:01 AM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
>> On Fri, Dec 9, 2016 at 3:34 PM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
>>> Hi folks,
>>> How is QEMU 2.8.0-rc3 holding up?
>>>
>>> I'm not aware of release blockers.  QEMU 2.8.0 will be released on
>>> Tuesday, December 13th if nothing comes up.  It will be the -rc3
>>> source tree.
>>
>> There have been no release blockers so I plan to tag QEMU 2.8.0 today
>> (December 13th at 19:00 UTC).
>>
>> If something critical turns up please notify me on IRC (stefanha on
>> #qemu irc.oftc.net).
>
> Mike Roth discovered a migration compatibility bug introduced by
> commit d1b4259f1ab18af24e6a297edb6a8f71691f3256 ("virtio-bus: Plug
> devices after features are negotiated").  The release will be delayed
> until this is resolved.
>
> Please check the mailing list for discussion.  Mike is sending out a
> mail summarizing the bug.
>
> Stefan
>

Maybe you can update the po files before tagging the release.
They currently no longer match the code (changed line numbers,
new string "_Copy").

Long way: make install
Short way: make -C po build

Stefan

Re: [Qemu-devel] Any QEMU 2.8 release blockers?

[Stefan Hajnoczi]

On Tue, Dec 13, 2016 at 9:39 PM, Stefan Weil <sw@weilnetz.de> wrote:
> On 12/13/16 22:04, Stefan Hajnoczi wrote:
>>
>> On Tue, Dec 13, 2016 at 10:01 AM, Stefan Hajnoczi <stefanha@gmail.com>
>> wrote:
>>>
>>> On Fri, Dec 9, 2016 at 3:34 PM, Stefan Hajnoczi <stefanha@gmail.com>
>>> wrote:
>>>>
>>>> Hi folks,
>>>> How is QEMU 2.8.0-rc3 holding up?
>>>>
>>>> I'm not aware of release blockers.  QEMU 2.8.0 will be released on
>>>> Tuesday, December 13th if nothing comes up.  It will be the -rc3
>>>> source tree.
>>>
>>>
>>> There have been no release blockers so I plan to tag QEMU 2.8.0 today
>>> (December 13th at 19:00 UTC).
>>>
>>> If something critical turns up please notify me on IRC (stefanha on
>>> #qemu irc.oftc.net).
>>
>>
>> Mike Roth discovered a migration compatibility bug introduced by
>> commit d1b4259f1ab18af24e6a297edb6a8f71691f3256 ("virtio-bus: Plug
>> devices after features are negotiated").  The release will be delayed
>> until this is resolved.
>>
>> Please check the mailing list for discussion.  Mike is sending out a
>> mail summarizing the bug.
>>
>> Stefan
>>
>
> Maybe you can update the po files before tagging the release.
> They currently no longer match the code (changed line numbers,
> new string "_Copy").
>
> Long way: make install
> Short way: make -C po build

I've sent a patch to the list and CCed you.  I ran make -C po update.

Weird that the line numbers are hardcoded in the .po file format.  An
addr2line type tool would have been better to avoid the constant
outdated line numbers.  That said, I don't know anything about how
gettext is designed...

Stefan

[Qemu-devel] QEMU po files / translations (was: Any QEMU 2.8 release blockers?)

[Stefan Weil]

On 12/13/16 22:52, Stefan Hajnoczi wrote:
> I've sent a patch to the list and CCed you.  I ran make -C po update.
>
> Weird that the line numbers are hardcoded in the .po file format.  An
> addr2line type tool would have been better to avoid the constant
> outdated line numbers.  That said, I don't know anything about how
> gettext is designed...

xgettext has an optional parameter --no-location which suppresses
the comments with filename:line.

All QEMU texts currently come from gtk.c, so that information
is not very useful. We could try running with --no-location
for the next release.

Even then the po files need a little bit more maintenance:
several texts don't have translations. And of course there
are more strings than those in gtk.c which could be translated,
e. g. the help texts, monitor output, error messages, ...

Stefan