Re: [REVIEW 3-6 ] No more prereg, no more unopkg sync

On Fri, 2012-08-10 at 16:32 +0200, Stephan Bergmann wrote:
> Quoting 
> <http://cgit.freedesktop.org/libreoffice/core/commit/?id=6dcb3d4ef46312729bb6f16c473b433474863f68>
>  
> "Related fdo#51252: No more prereg, no more unopkg sync":  "Now that 
> 5c47e5f63a79a9e72ec4a100786b1bbf65137ed4 'fdo#51252 Disable copying 
> share/prereg/bundled to avoid startup crashes' removed the use of 
> share/prereg, there is no longer need to generate it in the first place 
> (by calling 'unopkg sync' at build or installation time), and so no need 
> for the 'unopkg sync' sub-command, either.

        This looks like a really nice change. Yes it -looks- invasive, but the
weight of the LOC change is (as Tor points out) substantially redundant
code removal (I love it).

        IMHO - the less complicated and un-necessary stuff we run as we install
the MSI file the more reliable our install will be, and the fewer
un-needed dark-alley code-paths we have lying around the safer we all
are.

> Hence, I would suggest to not jeopardize LO 3.6.1 with this clean up, 
> but only backport it to libreoffice-3-6 after branch-off of 
> libreoffice-3-6-1 early next week.

        My concern is that we don't necessarily get better testing of this
(except perhaps on master ?) by merging it later - and indeed, most of
the heavy-duty upgrade testing we want done should (I hope) happen
around 3.6.1 rc1.

        IMHO I'd prefer to have an RC3 for 3.6.1 than the potential for
regression by including this in 3.6.2 - and having the state-of-the-art
code that is shared between master and -3-6 for testing seems sensible
(to me).


>   (This clean up would nicely remove 
> the root for <https://bugs.freedesktop.org/show_bug.cgi?id=44628> 
> "unopkg.exe does not find VC runtime during install (extension 
> registration)," but it might be better to address that with yet another 
> targeted band-aid fix for LO 3.6.1, like by just not calling "unopkg 
> sync" from the msi install scripts.)

        Yep - so I read it through; and I'm +1 for -3-6 (and hence -3-6-1) I'd
love someone else to read it too.

        The only gotcha I spotted was the:

@@ -222,18 +211,6 @@ NodeJava::NodeJava(Layer layer):

        constructor; but it seems we initialize m_layer earlier already so just
remove a duplicate initialization.

        Thanks !

                Michael.

-- 
michael.meeks@suse.com  <><, Pseudo Engineer, itinerant idiot