[FFE] LibreOffice 3.4 for Oneiric

[Björn Michaelsen]

LibreOffice 3.4 should be included in Oneirc, it is the most recent stable release and it is being stablized in sync with out progress to beta.

[Björn Michaelsen (bjoern-michaelsen)]

Release notes are here:
http://wiki.documentfoundation.org/ReleaseNotes/3.4

[Björn Michaelsen (bjoern-michaelsen)]

The current 3.3.2 package in oneiric ftbfs because of mono dendency troubles.

[Björn Michaelsen (bjoern-michaelsen)]

Builds libreoffice and libreoffice-l10n 1:3.4.1-4ubuntu1 succeeded in:
https://launchpad.net/~bjoern-michaelsen/+archive/libreoffice-oneirictest-20110718

[Martin Pitt (pitti)]

Approved. We planned for 3.4 all along.

[Matthias Klose (doko)]

I don't like how this FFe was granted; apparently LO 3.4 was never built before on armel and powerpc, and promptly failed to build. There is no reason, why such an upload cannot happen before feature freeze, and no reason that these architectures are not tested before an upload.

Please re-consider this before the next approval of such a high profile package.

[Martin Pitt (pitti)]

Matthias Klose [2011-08-23 18:41 -0000]:
> There is no reason, why such an upload cannot happen before feature
> freeze,

There were some blockers which prevented the upload so far.

> and no reason that these architectures are not tested before an
> upload.

Not having ports support for PPAs does make this quite nontrivial,
though. Also, 3.4.2-2 did build in Debian experimental on arm and
powerpc, so it should be quite possible to fix it for Ubuntu, too?

[Björn Michaelsen (bjoern-michaelsen)]

> I don't like how this FFe was granted; apparently LO 3.4 was never built
> before on armel and powerpc, and promptly failed to build.

That does not match up with your requests on IRC to _not_ test in PPAs but
directly in the distro -- unless you are searching for reasons to complain.

> There is no reason, why such an upload cannot happen before feature freeze,
> and no reason that these architectures are not tested before an upload.

There are very good reasons for other non-x86 archs to be tested last: Build
time, compile cylces, availability on PPAs are just a few.

And yes, next time when I bite my nails out with some problem like the mono one
here, I will first see if somebody already caused a FTBFS with the current LO
version by changes in one of LOs deps before trying to find what errors was
introduced in LO.

But rather than wasting my time argueing here, I'd prefer to get something
constructive done, so lets please put this to rest here.

[Dmitry Shachnev (mitya57)]

And what about 3.4.3 bugfix? http://blog.documentfoundation.org/2011/08/31/the-document-foundation-announces-libreoffice-3-4-3/

[Matthias Klose (doko)]

On 08/23/2011 11:54 PM, Björn Michaelsen wrote:
>> I don't like how this FFe was granted; apparently LO 3.4 was never built
>> before on armel and powerpc, and promptly failed to build.,
>
> That does not match up with your requests on IRC to _not_ test in PPAs but
> directly in the distro -- unless you are searching for reasons to complain.

wrong. I didn't see any uploads to the libreoffice PPA either. please get some
exposure to the community early. I did hope that the development behind the
curtain did have an end with OOo, and didn't propagate to LO.

>> There is no reason, why such an upload cannot happen before feature freeze,
>> and no reason that these architectures are not tested before an upload.
>
> There are very good reasons for other non-x86 archs to be tested last: Build
> time, compile cylces, availability on PPAs are just a few.

there's no reason to build at least once on a developer machine. These are
available for a reason.

> And yes, next time when I bite my nails out with some problem like the mono one
> here, I will first see if somebody already caused a FTBFS with the current LO
> version by changes in one of LOs deps before trying to find what errors was
> introduced in LO.

this is development. detect it early and file reports. if mono is such a pain,
please consider dropping support for mono.

> But rather than wasting my time argueing here, I'd prefer to get something
> constructive done, so lets please put this to rest here.

if you want to be constructive, please address these issues.