Change-Id: I7c22869028a78f48836741a4eb4cd5bbff16e884

Change-Id: I98a3428578b52c6d9d7210a8fb0211876d6f7cae

Two different xmlparse libraries are created: ascii_expat_xmlparse and
expat_xmlparse. One without -DXML_UNICODE and one with. Source file are
duplicated and renamed with gb_UnpackedTarball_set_post_action function
to be able to add artifacts twice to gbuild machinery.

On windows 64 bit additional two librares are created: expat_xmlparse_x64 and
expat_xmltok_x64. That is due the problem with shell/shlxthandler (comment):
------------------------------------------------------
use UNICODE only because shell/shlxthandler
doesn't link against ascii_expat_xmlparse
------------------------------------------------------

Include files are delivered to $(OUTDIR)/inc/external/expat
now and not to $(OUTDIR)/inc/external any more.
set_include call is added in RepositoryExternal.mk.

To define dependency between StaticLibrary and ExternalProject
new function was introduced: gb_StaticLibrary_use_external_project.

Change-Id: I3b3aa40f39ef82c70f6f28790b582c83e48bdf76

Change-Id: I770cdbb23e9a42308e95505235378d715c1fcd79

Mainly because it seems we hit this assert several times and right now
it seems harmless.

Change-Id: I9176741dd25f405c9eb49daef9c703cfde4386c4

All the other messages are already > 1, makes no sense to have this as
> 0.

Change-Id: I816239903cef650f34d9c57446018843e2acb64a

… in this function"
declaring those functions static breaks the build on MacPPC / gcc 4.0.1

Change-Id: I3a389530d51ae0ddd38418159ca4234016bf93ba
Reviewed-on: https://gerrit.libreoffice.org/911Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
Tested-by: Stephan Bergmann <sbergman@redhat.com>

so we can keep translations markup free, and it matches
better our existing font use in controls

Change-Id: Ib762030f8bf831f2b0afabdd643508e660ad9266
See: https://live.gnome.org/GnomeGoals/RemoveMarkupInMessages

Change-Id: I33ca2c27d557d24e3900d7ebf8f3ea2d99b3cfb0

Change-Id: I2cd4e0965452c376df8c5e43353fbdd6f62fb27d

Change-Id: I765762f11813c6b612416a8db2707bf94114c876

Change-Id: Ided6c480969764073056830722c8996e0df52285

Change-Id: I62c5f9206b6bc0321b7ee2b9c679ab7b5400a8ef

Change-Id: I2bbc2763eca7c878ead2a325589d8c4bb9210bda
Reviewed-on: https://gerrit.libreoffice.org/868Reviewed-by: Marc-André Laverdière <marc-andre@atc.tcs.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: Icde740ea12e8839b0a4c8e0317143dbe2157a964
Reviewed-on: https://gerrit.libreoffice.org/894Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

...by folding the contents of java_accessibility.jar back into
java_uno_accessbridge.jar.

In the old build system there were two jars, java_uno_accessbridge.jar
containing the handful of org.openoffice.accessibility classes and all
org.openoffice.java.accessibility classes (though how the latter got included
was fairly obscure in the makefile.mk) and unused java_accessibility.jar that
contained all org.openoffice.java.accessibility classes.  When adapting this to
gbuild, the unused java_accessibility.jar was carried over, but all its
org.openoffice.accessibility classes were inadvertently droped from
java_uno_accessbridge.jar.

Change-Id: I9b582ba22667b1dae635828e85c4cc5b530353ac

Change-Id: I58cc6aa84f6c4df00b538f290f1856de54bfa4c0

Change-Id: I915b02f9ac0140afcf392210001e7948c5c2be4a

Change-Id: I55b6bf11468ce958aea7c98894d3a76343b53a9a

Change-Id: I5835696a3246b4d8f401a1617d2b53ec8c7c5626

Packagers get warning from autotools about using wrong configure opts.
If they decide to ignore it then its their problem.
With this approach we would have to die on all the configure script
changes which is bit annoying.

Change-Id: Ie20d28ab0d71f995761ccdb2c83fe89638f7a313

Change-Id: Ie2c7b219f57a2bdf14fc07311a03a2c66f6c75aa

Change-Id: I1c0c015f371346efc5712ead7abe9e7f2ec0f9d0

Change-Id: I8b76ba6d6a27cebefca3307a648bd5baafd27262

Change-Id: Id87d02b6144168b2731a814c4ec5bc3c026dcda4

... but the tests still crash and are disabled.

Change-Id: I5df66f21865cac98c38d4ef3684c562784e093b9

Change-Id: I723f70224a00972993d5b8592dd5e3834be7fd84

... the ExtensionTarget requires the description.xml to have a
translation in the --with-lang case.

Change-Id: I155e47d8a44590995ffcd63cb9ac8f02822a82e9

Change-Id: I58a547db30dda833cf1dd6a7563cd1589876431e

Change-Id: I9a6668a785d1dd6d6f02716519dedb01fb90af00

Change-Id: Ia53ddbfcf68524b7714fe6df447a9dbb1a68bcb7

Change-Id: I9812079bd9987c841dd1856573de768d7192873b

Change-Id: Ie853d55f78f27a7249b8e960587c8d2bd833d1a7

Change-Id: I0ccc91ea686e599171845ff2cff5e536bbdb7b77

Change-Id: I7607ddc9e207c87d6126715c43970ac22d3aed62

Change-Id: Ic6477adbd8e8eaa24f851befe5d18d34061e88ce
Reviewed-on: https://gerrit.libreoffice.org/905Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

...as 49a92ab3 "commit changes in dictionries,
helpcontent2, and translations" had erroneously moved them back to old commits.

Change-Id: Iea6165ab2011de0b25129759442948cb3d0e3083

Add document::XEventBroadcaster interface to XGlobalEventBroadcaster,
because a client uses it, and the service always implements it.

This is safe to do because we haven't released an LO version with
XGlobalEventBroadcaster in it yet.