To run on ancient CPUs, we compile for Windows with -arch:SSE since
8bd6bf93b7 (August 2014). Thus
_M_IX86_FP gets defined as 1. This meant that LO_SSE2_AVAILABLE did
not get defined, and that we hardcoded tools::cpuid::hasSSE2() as
always returning false. That was hardly the intent.
Change-Id: I7ee34510a774dab865c8990b74b91a5284218a96
so releaseCairoContext is a mirror of getCairoContext, so callers that don't
need to flush can drop that explicit call
Not sure, now that I see it, how there's been no sign of any problem with no
damage of the affected text area under gtk3
Change-Id: I8e508160be4102c64a0ecdf5591b6999f9171c36
but a delete does. The diff from delete to cut is the
EnterBlockNotifications/LeaveBlockNotifications calls so try that
Change-Id: I10147043d6fe358bf2f4f4bd6877210b86cbbd1f
...when SfxTabPage::Reset (sfx2/source/dialog/tabdlg.cxx) calls into
DoApplyFlags/DoRest with a null rSet (but which apparently isn't actually used
then), e.g. upon "File - Export as PDF..." in Impress.
Change-Id: Ie2f255694212356fa8fc994287c45f2b4730a5bd
... attributes that happen if both CharHighlight and CharBackColor
properties are used, because the CharBackTransparent property wasn't
taken into account, and combining the CharBackColor and
CharBackTransparent properties happens *after*
XMLTextExportPropertySetMapper::ContextFilter() runs.
Also, it looks like a transparent highlight wouldn't export properly but
apparently DomainMapper::getColorFromId() won't create such.
(regression from f880962f5b)
Change-Id: Ib628ef8bb377482f74fadb97c81afb95fbbf7184
partial revert of
commit 9149dbf173
Author: Yousuf Philips <philipz85@hotmail.com>
Date: Fri Nov 20 09:40:59 2015 +0400
tdf#91909 Add shortcuts to slide navigation
to get home/end working in text boxes again
Change-Id: I9960b94077c4c00d63483d3105d73f418cbcef0e
...location." Caused many crashed during "make check", when aLocaleNames is a
sequence of 17 ("ar-SA", "ar-DZ", ...), but aLcoations merely a sequence of 2
(".../dict-ar/ar.aff", ".../dict-ar/ar.dic").
From comments further down below, it looks like aLocations will always contain
exactly two entries ("also both files have to be in the same directory and the
file names must only differ in the extension (.aff/.dic)"), and that all the
aLocaleNames members share the same set of aLocations ("Thus here we work-around
this by adding the same dictionary several times. Once for each of its supported
locales"), so that it would appear to be OK to just check once for the existenceof aLocations[0].
(There is a check for
if (aDictIt->aLocaleNames.getLength() > 0 &&
aDictIt->aLocations.getLength() > 0)
below, so it might be that these can be empty, or it might just be "defensive
programming." Play it safe, and check here that aLocations is not empty.)
Change-Id: I82bea6571983e397a9e164b294a5ba656b511a67
instead of setting a single tooltip for the whole
panel area, wait for a tooltip request and specify
the area to which it applies.
Which is what other widgets that have different
tooltip areas do, and maps much better to the gtk3
tooltip system which functions like that too.
Change-Id: I56f452562c3c4d9dcec4109a8ee6168f34619472
... and implement it in Writer.
Otherwise there is no way we can perform e.g. page down in an expected
way. Without this, the core visible area depends on the zoom in the
document, and the client visible area can be something entirely
different.
Change-Id: Iadfb5a225da09a2551ffa41ddf503bb3d22b3eae
Instead of fighting with this obsolete thing, let's
just kill it. This fixes disabling in read-only doc,
as well as pressed/unpressed states.
Change-Id: I4d06ba3395024ffd09d77c6d22653a2c5e227115
Reviewed-on: https://gerrit.libreoffice.org/22010
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Maxim Monastirsky <momonasmon@gmail.com>
On Windows, the text is rendered with 0 of alpha, which is an issue in
this transition as it enables GL_BLEND.
Change-Id: I45fe32ccf77854e758eddc33f300791d4cdd6904
But seriously, who builds with OSL_DEBUG_LEVEL > 2?
Either debugging printouts inside such ifdefs are completely pointless
and should be removed, or then the information might actually be
useful, and the ifdefs should be removed, so that the information can
be selected to be printed using the normal SAL_LOG environment
variable mechanism (in a build where SAL_INFO is functional).
Change-Id: Ia5d2c7c7b525ed133fe3f077a005f583b8a856e7
This is similar to what is implemented in online.git's leaflet. If the
zoom factor changes, then inform LOK about it on the first keyboard hit.
Change-Id: I2db4ffed3897db5c852f4b8aea8f4e2d3500a25a
This looks obsolete, it was used in LibreOfficeLinux.conf since the
beginning but in fact since LO 3.5 it was broken and the vclplug_gen
was not actully dlopen()ing Xrandr.so but somehow magically "found"
the symbols anyway.
Change-Id: I81251a8fef0abac63183b0ff3533bb2331ca6f5a
Since commit ae0e830f9a, mirroring
is handled correctly in LO, so no need to handle it specially in
PDF import code. This obsolete handling causes double-mirroring.
Change-Id: Ia873f0426e444cafd80b24b7d209e90eb481d8f4
Reviewed-on: https://gerrit.libreoffice.org/21870
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Armin Le Grand <Armin.Le.Grand@cib.de>
This was borked since gbuildification of vcl in 2011 and the dlopen
code didn't even compile since commit
b5f1139427
Change-Id: Ieff22ed144d9d89e53dd956ce0dfea5e4d07ea13
