With --enable-pch=full there's not much difference between a "public"
header in <module>/inc and a private one in <module>/src/somewhere/inc .
And since the script searches recursively, this apparently helps to
find even more headers for lower pch levels.
Change-Id: I8483d0aa5b4fea5a59107c20a8aa5f1ef694af0a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87799
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
This normally wouldn't make a difference, but it does when building
with Clang's -fmodules-codegen to collect duplicated templates,
inlines, etc. into extra .o file. The inline calling the template
is emitted in that .o, but the template is defined only in the .cxx
and not referenced from it, so it's never emitted.
Change-Id: I8b49030f2be6146ca1e2cd5e35734c0db6cca394
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87798
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
this is the first honggfuzz (honggfuzz_asan_libreoffice) report I've seen to date
Change-Id: Iac733aa63c7c94d9454b9c8596340dc2286393c0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87803
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
* Update helpcontent2 from branch 'master'
to a7c86802627ba84c6990354e020aa34b2738ec57
- tdf#130064 adjustments in help page about legacy list style positioning
- add "access to"
- remove "heading" / change position of note
- add help on "Apply" control
- add "Load/Save" control
- corrections in description of controls (see bug report for details)
Change-Id: I8a822c44cfd0e80e5b850378a76f70f1727abe14
Reviewed-on: https://gerrit.libreoffice.org/c/help/+/87541
Tested-by: Jenkins
Reviewed-by: Seth Chaiklin <sdc.blanco@youmail.dk>
tracked it down to...
Removed default initialization, use GLM_FORCE_CTOR_INIT to restore the old behavior
so adding in GLM_FORCE_CTOR_INIT to get them working again
Change-Id: I1c6e7d8eb748fce40f0c518ff708708e5fb1e3d2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87789
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
which is much better for CPU cache, since the representation is more
compact, and since we almost always do insert() in pointer order,
there is surprisingly little sorting
Also add a count() method for compatibility with std::set and the
proposed std::flat_set
Change-Id: I2a3211dc59919cfec5cac1497530a4c3600d50ca
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87793
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
... resulting in error message
C:/lo/src/build/workdir/UnpackedTarball/skia/src/gpu/gl/GrGLGpu.cpp(3771):
error C2039: 'f__faststorefence': is not a member of 'GrGLInterface::Functions'
Change-Id: If69c0067abe287dd276c685cde6be3b31a1ce950
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87780
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
which is always a better choice when dealing with pointers, for
2 reasons: way more cache dense, and we very often iterate pointers
roughly in-order, which means very little sorting needs to be done
Change-Id: Ib38f56018b2fc16414c9af910421cc45f069e1f3
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87782
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
This mainly replaces the whole AWK code with the git helper
"interpret-trailers", which was added in git v2.2 end of 2014.
It also moves the argument checks from the original Gerrit hook
to the front of our tests to verify the script arguments.
Change-Id: I38c831bf7c9d399419a598d6966e48166d31ea6f
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87369
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
Create specific option to toggle display of the bookmark positions
instead of relying on generic IsViewMetaChars().
Add a Bookmarks checkbox to Writer->Formatting Aids page in Options.
Add property ShowBookmarks to com::sun:⭐:text::ViewSettings.
Add config setting:
org/openoffice/Office/Writer/Content/NonprintingCharacter/Bookmarks
Change-Id: I998ecdc16d8e4a5baeeb1bbcff0bf5f1f5636692
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87774
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@cib.de>
in case of all chart types except pie chart.
Clean up commit 4223ff2be6
(tdf#48436 Chart: add CustomLabelPosition UNO API property).
Note: use the correct default OOXML label placement in case of
radar charts.
Change-Id: I9a8f509304b3c70d879c8c6a95bc91d15ac28521
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87759
Reviewed-by: László Németh <nemeth@numbertext.org>
Tested-by: László Németh <nemeth@numbertext.org>
check indentation of braces in namespace decls,
and the comments that often appear with them.
This is my penance for messing up the indentation with
clang-tidy-modernize-namespaces.
As such I have limited it to new-style namespaces for now,
and the check is off by default.
Change-Id: I4db7f10a81c79bc0eece8f8e3ee564da8bc7f168
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87723
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
paragraph style to avoid missing text lines later
in vertically merged table cells at page break.
From commit 49f453755b
("tdf#89714 - enable Widow/Orphan in default style"),
Table Contents got unnecessary orphan/window
control. Unfortunately, recent table layout code
cannot ignore these settings completely, causing known
problems, see for example tdf#128959 (FILEOPEN DOCX
Table row content disappears when broken between pages).
Change-Id: Idd570f17b0a11af85072a65f3422535b993db306
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87730
Tested-by: Jenkins
Reviewed-by: László Németh <nemeth@numbertext.org>
Tested-by: László Németh <nemeth@numbertext.org>
This means that in case a user field is exported to DOCX and the user
updates the field, the result will be still correct, not empty.
Change-Id: I2b52292c70aa6f597f92af95e16c773839247efa
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87748
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
Since tdf#121441 we parse custom footnotes to get at least the
DOCX footnote text, even if we can't represent the formating. This
might push additional contexts to the parser stack. Therefore it's
now not sufficient to check the current context for a footnote,
but one has to check the global parser for a footnote context.
The actual bug is the unsupported footnote page break, which was
not correctly ignored and added a paragraph context to the stack,
resulting in the async substream input and output stack size.
Change-Id: I143254e7df37a619cb4efb542b58d3eff3afffa7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87114
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
I was correct to see in 202146901b that the black bitmap was
in premultiplied alpha, but what I missed what that some controls
keep the alpha set at zero (and only some work properly). So go back
to the algorithm of synthetizing alpha from the red channel, compute
it properly (before it was using alpha channel by mistake), and
treat the data properly as premultiplied.
This hopefully finally makes all Windows control widgets work.
Change-Id: If2716eb8ecf623fcc57ee1db5904edfaee679aa9
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87734
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
Add a new SwBookmarkPortion, derived from SwControlCharPortion.
There is no character for the bookmark in the text so the portion has
length of 0, which makes things quite a bit more tricky.
Formatting:
* SwBookmarkPortion is created last in WhichFirstPortion().
* In an empty paragraph the SwTextFrame::FormatEmpty() must be disabled.
* If there's a bookmark at the end of a paragraph,
SwTextFormatter::CalcAscent() must use the font of the previous
character, not the paragraph font, because that could grow the line
if it's higher.
* The SwMultiPortion complicates matters, because it uses a nested
SwTextFormatInfo and thus we need some extra steps to prevent
duplicate SwBookmarkPortions; this is particluarly a problem for
rotated text portions.
- SwTextFormatter::BuildPortions() must advance the outer
SwTextFormatInfo's bookmark position because BuildMultiPortion() has
already created the SwBookmarkPortion.
- If a SwBookmarkPortion is at the start of a SwMultiPortion, it will
be created before the SwMultiPortion but must be painted inside the
SwMultiPortion because its font is going to be initialised as inside
the SwMultiPortion (e.g. it will be rotated) so its position must
also be adapted to be inside, and only
SwTextPainter::PaintMultiPortion() does the setup for that; add a
hack to move it in SwTextFormatter::BuildMultiPortion().
Painting:
* Using the original font seems rather difficult, hard to predict what
some character is going to look like, and how it scales if the size is
increased; use OpenSymbol instead. Unfortunately OpenSymbol doesn't
have a good glyph that could be used for both the end of a bookmark
and start of another bookmark at the same position.
* SwLinePortion::PrePaint() wants to avoid moving the portion
half-outside the frame but often it looks better that way (previously
it was misaligned, now it's half-outside).
* Not sure if it makes sense to draw a field shading in
SwTextPaintInfo::DrawViewOpt() too; let's try it out, but dependent on
the same IsViewMetaChars() setting as the bookmark character itself,
not on the field shading setting.
Change-Id: I1ab94afb417221e278dbb3afd6c978a05fc78497
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87364
Reviewed-by: Michael Stahl <michael.stahl@cib.de>
Tested-by: Michael Stahl <michael.stahl@cib.de>
The sign was flipped for the 90° case, resulting in painting too far
up the page, ever since this was added in
293ed704cf.
In SwControlCharPortion::Paint() rotation wasn't handled at all.
Change-Id: I83a7eef8b2f6126e070d8c48f75378b6cf6d50e1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87707
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@cib.de>
