...it just appeared to work on machines that happen to have /usr/bin/clang. So
hoist CLANGDIR/CLANGBUILD from compilerplugins/Makefile-clang.mk to configure.ac
instead.
Change-Id: Ie5d9c6bb8e9d0caa1583d78c8693f06b69873095
Remove the oooblogger extension. It is unmaintained and the original
author (caolan) wants it dead.
Change-Id: Iff89f5badbb74f465c57e068a9e1acbbc13bc235
It can be used to show extra build indentification in the about dialog
just under the version string.
The plan is to show information about tinderboxes and make the life easier
for QA guys.
Change-Id: I4714a34cf13492c7bb92798b4b6d11385093f223
..and add few fixes, like:
* use 'set -o pipefail' to fail meaningfully
* include download.lst at the end of config_host.mk
* MOZ_ZIP_* do not have md5sum, do not check for it
Change-Id: I23759aa2f5fdd55b662588fdb0a31568a686aef7
- this is needed for packaging, so that all possible tarballs can be
available in the complete source package
Do not define tarball names in configure. Move the variables into
download.lst, which is going to be included in config_host.mk.
We already have conditions for getting tarballs in Makefile.fetch.
Change-Id: I160aff6426a253814eea07d2467e4942478f653a
Currently there are 4 different debug options:
--enable-dbgutil (the recommended one)
--enable-debug
--enable-selective-debuginfo
--enable-symbols (for advanced users only)
In this table the properties of each option is shown:
----------------------------------------
options\properties | O | S | D | T | U |
----------------------------------------
production-code | x | - | - | - | - |
----------------------------------------
--enable-symbols | x | x | - | - | - |
----------------------------------------
--enable-debug | - | x | x | x | - |
----------------------------------------
--enable-dbgutil | - | x | x | x | x |
----------------------------------------
where
O: optimization
S: debug symbols
D: debug STL
T: trace facility
U: additional debug utility (object counting)
Note:
--enable-selective-debuginfo has the same properties as --enable-debug
Change-Id: Ib8a28c6162f47526d6bb33f81f53835cd11894b2
Even for an Android (or iOS, presumably) app with just viewer
functionality, some export code gets included in the single .so or
executable.
For now, when we know that we are producing only viewer apps for
Android and iOS anyway, it might be a good idea to be able to
carefully bypass some particularly pathological places in the code
that generate large amounts of code and/or data that is useless in a
viewer app.
Change-Id: I59b97024ff2855e7bbe53d6ec8782797705f61d8
This hopefully should help in situations like Jonathan Aquilina's,
where /usr/bin/install_name_tool does not correspond to the tool-chain
used (but is from an earlier version of Xcode).
Change-Id: I5b4ca9e5e163fb4a84967577d2146cdbe8344f03
The immediate trigger was 5e5c11c664 "fdo#42070
Fix RTL support in presenter console" causing build failures on Mac OS X when
linking the extension against vcl, but there should be more benefits of going
from a bundled-anyway extension to plain code. (Not the least to get rid of the
com.sun.star.drawing.XPresenterHelper hack.)
To avoid unnecessary confusion between the newly plain code and any instance of
the old extension still installed (per-user or shared), I renamed all relevant
identifiers as follows:
* UNO implementation com.sun.star.comp.Draw.framework.PresenterScreenJob ->
org.libreoffice.comp.PresenterScreenJob
* UNO implementation com.sun.star.sdext.presenter.PresenterProtocolHandler ->
org.libreoffice.comp.PresenterScreenProtocolHandler
* protocol handler schema vnd.com.sun.star.comp.PresenterScreen ->
vnd.org.libreoffice.presenterscreen
* configuration schema /org.openoffice.Office.extension.PresenterScreen ->
/org.openoffice.Office.PresenterScreen (it appears this contains little to no
user-changeable data anyway, so not migrating it to a new user profile due to
the schema name change should not be problematic)
* job ID onDocumentOpenedJob -> org.libreoffice.PresenterScreen
Even with these precautions, having the presenter screen installed both as plain
code and as a (per-user or shared) extension still leads to a crash when
activating presentation mode (likely due to how both codes want to take control
of the screen). To mitigate this, existing installations of the extension are
explicitly not migrated to new user profiles.
The sdext/source/presenter/bitmaps/*.png files were moved to
icon-themes/galaxy/sd/res/presenterscreen-*.png and are now accessed via SdResId
(adding the relevant data to sd/source/ui/inc/res_bmp.hrc and
sd/source/ui/app/res_bmp.src; not sure whether these locations are already
ideal).
The code itself has been left mostly unchanged in sdext/source/presenter/, and
it still clumsily communicates with sd core code via XPresenterHelper. There is
a lot of room for improvement here.
The help data is left untouched at sdext/source/presenter/help/ and needs to
be incorporated properly into helpcontent2 in a follow-up commit.
The --disable-ext-presenter-console configure switch is gone.
Change-Id: I71adb7ae6dcdbd1802151fce6e3871d8a2026332
This is a rework of f9059d4eee "Key all browser
plugin features to --enable-nsplugin." The problem with that was that Mac OS X
supports pluging browser plugins into LO documents (which was originally
controlled by --enable-mozilla) but not plugging LO into browser windows (which
was originally controlled by --enable-nsplugin), so controlling the former with
the same switch as the latter did not actually work.
Thus I replaced the single ENABLE_NSPLUGIN feature flag with two dedicated ones,
ENABLE_NPAPI_FROM_BROWSER (for plugging browser plugins into LO documents) and
ENABLE_NPAPI_INTO_BROWSER (for plugging LO into browser windows). The
--enable-nsplugin configure switch is gone completely, setting the feature flags
is always done fully automatically now.
Change-Id: Iecf706637465e865c987563b5de489fa90b4c904
The current configure option --enable-bluetooth is a bit misleading,
it doesn't really tell what it's actually for. Therefore renamed it,
so it's more clear that it's an sdremote backend using bluetooth.
Change-Id: Ia8b46ee001ea112b80521baa502dcab2bb7e83aa
Reviewed-on: https://gerrit.libreoffice.org/1086
Reviewed-by: Rene Engelhard <rene.engelhard.re@googlemail.com>
Tested-by: Rene Engelhard <rene.engelhard.re@googlemail.com>
...to avoid confusion what that is actually about. (Especially, it is
completely unrelated to --enable-mozilla.)
ATTENTION: This renames the configure option --with-system-mozilla-headers to
--with-system-npapi-headers.
Change-Id: I48f3c79d69c4d4f445a94e7ddbeab06ea667becc
This allows us to simply pass the already exported variables to invoked
configures, where the non-empty values will override the detection logic
there.
Change-Id: I7353a313814c58922e34843aeae380e5afa442ca
(PT Serif, Source Code Pro and Souce Sans Pro under OFL; Open Sans
under ALv2)
Change-Id: Ic783f0d5841cafce6dba60213b0a658978ff2cfd
Reviewed-on: https://gerrit.libreoffice.org/1007
Reviewed-by: Andras Timar <atimar@suse.com>
Tested-by: Andras Timar <atimar@suse.com>
Now with librsvg no longer used, and liblangtag no longer using glib,
a bunch of bundled (in some configurations, always on some platforms)
3rd-party libraries are no longer needed.
Initial work by rene, continued by tml.
Change-Id: I76edd7aea5452e3487499f0b9ed9f767cf760194
peter's gbuildifications caused and ocean of red, that
1/ was not followed up by any attempt by the author to fix
2/ I wasted a good part of the weekend to de-entangle with
only partial sucess
reverting the whole mess for now...
to be resubmited when a bit less borked...
This reverts commit c4c160a451.
This reverts commit faef2e51d0.
This reverts commit 057ce1fe29.
This reverts commit a7d34af344.
This reverts commit befae0ceb8.
This reverts commit 433b43bcd2.
This reverts commit 93e2c4a9d2.
use plain copy instead of git-new-workdir when using --with-linked-git
configure switch (point it to your master copy of the core-repo)
also remove obsolete bin/create_bootstrap_links, bin/mkworkdir and
bin/repo-list[.in]
Change-Id: I37028badf2c46a24a3d628e80476efe4dd365d06
Reviewed-on: https://gerrit.libreoffice.org/912
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
Rationale:
- it is advised to use max-jobs and num-cpus with the same value in wiki
- max-jobs was used only for lcms2 and few gbuild
modules outside of tail_build anyway.
Also fixes:
- really use CHECK_PARALLELISM when meant to
- EXTMAXPROCESS is not defined in gbuild;
use parent's jobservers in sub-make where possible
Change-Id: I501de732d223ce0c935081bd1d73da611d16ee88
Reviewed-on: https://gerrit.libreoffice.org/930
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>