[Libreoffice] [tdf-discuss] Linux distros and LibO packaging

[Rene Engelhard]

[ fullquoting for discuss at dfs sake. forgot the CC. Not that it matters
much, but anyways. ]

Hi,

On Wed, Jan 19, 2011 at 11:13:48AM +0100, Rene Engelhard wrote:
> On Tue, Jan 18, 2011 at 09:13:43PM +0100, Andrea Pescetti wrote:
> > It is a good idea to track changes, but it is probably a questionable
> > practice to make changes. I expected LibreOffice to be consistent across
> 
> Nonsense. This is OSS.
> 
> > it). Are there compelling reasons why distributions should ship versions
> > of LibreOffice that have significant changes with respect to the
> > "official" version?
> 
> Define significant changes? Does ripping off the Mozilla address book
> support (implicitely, because using system-mozilla) count as that? Would
> you prefer Linux distros having a obsolete, patched and insecure Mozilla
> copy there? No, not acceptable.
> 
> > The OpenOffice.org experience, and the first distribution-specific
> > LibreOffice bugs like
> > http://www.mail-archive.com/discuss@documentfoundation.org/msg04508.html
> 
> Wow. I don't think Petr added a patch here, so it might just be system
> differences? Petr, correct me if I am wrong.
> 
> Besides that, distros will have to continue libreoffice-build, which does
> still contain patches. (Removing those would be a big regression about
> what we ship right now)
> 
> > make me think that fragmentation, while of course allowed by the
> > license, should be discouraged when it comes to functionality; I'm not
> > questioning desktop integration or branding, but I'd like to know why
> > distributions feel they have to make changes to functionality...
> 
> Because bugs should be fixed ASAP, not when you think one wants to release.
> What if Debian didn't backport important fixes to it's 3.2.1 from 3.3 or so?
> Should we release wiith known important bugs in a stable release. Living 2
> years with it? No. You have to care about quality.
> 
> Besides that, some distro-specific bugs are not by feature patches, but just
> because of other bugs, Like bugs in system-libs, new version of systen lib
> breaking XYZ (e.g. the ) wrapping issue, need to find out the bugnr caused
> by changes in the Unicode Standard and ICU 4.4), build issues etc. Those

https://bugs.freedesktop.org/show_bug.cgi?id=31271 is what I meant here.

> you can't foresee and neither does documenting every change help here.
> 
> Get some clue. And don't speak about this if you don't, kthxbye. Noone
> does this intentionally.

Sorry, I apologize for the first two sentences of this. But I am getting
annoyed by those senseless discussions. Should we repeat the errors
Oracle did again?

Grüße/Regards,

René