[Libreoffice-commits] .: distro-configs/LibreOfficeMinGW.conf distro-configs/LibreOfficeWin32.conf distro-configs/LibreOfficeWin64.conf

[Michael Meeks]

On Fri, 2012-03-30 at 08:58 +0200, Stephan Bergmann wrote:
> On 03/30/2012 08:51 AM, Stephan Bergmann wrote:
> > Looks like this kills MinGW builds
> 
> Ah, looks like Luboš already fixed it, with 
> e20fa170160e1bb1953ad171e092edfb3de531af.

	At root there is some *serious* badness that crept in here. I'd really
like someone to spend some time to unwind how exactly we managed to
screw up our configure.in -so- badly and it remain un-detected for so
long.

	It really needs an expert proctologist to get to the bottom of that;
there is a sizeable configure.in change:

commit f8d64ffd4a2ee3f6b3340a2bbac21eb7d2b4551c
Author: Michael Stahl <mstahl at redhat.com>
Date:   Mon Nov 7 16:58:52 2011 +0100

    configure: refactor system-libs checks:
    
    * initialze with_system variables in AC_ARG_WITH
    * do not interpret --without-system-libs as --with-system-libs

	I assume you re-configured and diff'd the before/after Host.Env.sh to
check nothing changed ? [ but then it was only the Win / Mac builds that
changed ].

	Or - was this down to the tangle with:

commit 202557da3c4f1cd57f46a4ba1c9d74e7b4d1c2db
Author: Korrawit Pruegsanusak <detective.conan.1412 at gmail.com>
Date:   Sun Aug 28 15:11:26 2011 +0700

    Also build cairo on Windows if directx disabled

	I -hope- that this was related to the VCL / rasterizer_rsvg work and
there is no knock-on effect on other configure options / defaults
elsewhere :-)

	Hmm,

		Michael.

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