Age | Commit message (Collapse) | Author |
|
This skips invisible tabs in many more places and at least fixes
the osx bug of the missing tabs of the NBB. It also fixes the
multiple highlighted tabs assertion for multiple rows from
TabControl::ImplGetItem.
I'm quite sure the tab rect calculation for multiple rows still
has the wrong height, so there might be minor overlapping.
Change-Id: I3a0a06ee73204b98e367563700fce4a49c066c1a
Reviewed-on: https://gerrit.libreoffice.org/75230
Tested-by: Xisco Faulí <xiscofauli@libreoffice.org>
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Change-Id: I7737c8b19dca787eab70b733d07a34a5e5f01c9c
Reviewed-on: https://gerrit.libreoffice.org/75208
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I80e197ffcda4ebc5d50e611b7a4a5c3fd84fc839
Reviewed-on: https://gerrit.libreoffice.org/75201
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
See tdf#94879 for motivation.
Change-Id: I1b84376b550a2b8dd57cc1046077578efbee7bed
Reviewed-on: https://gerrit.libreoffice.org/75191
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
Currently the cursor stays at the old position when you right click
in a different position in the text.
This causes some issues, e.g. when you right click on a hyperlink in
draw, you don't get the context menu for the link if the cursor is not
on the link (see tdf#98575 for related bug).
So generally, when right clicking on a piece of text, move the cursor
to where the click happened.
The new behavior is now matching the behavior of Word, PowerPoint, etc.
Change-Id: I5e0ac37b7ac6c859d3056c5b8ed453f97c747360
Reviewed-on: https://gerrit.libreoffice.org/75127
Tested-by: Jenkins
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
|
|
Since we no longer install on such systems anyway
Change-Id: Ia3b5c202e1276857c97017bb5a7eb9f087d5626e
Reviewed-on: https://gerrit.libreoffice.org/74961
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
Change-Id: Id8f72781d0338344bc96efb5cf62b67174aa80eb
Reviewed-on: https://gerrit.libreoffice.org/75152
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
Tested-by: Julien Nabet <serval2412@yahoo.fr>
|
|
Change-Id: Ic2ee5c2e122244e11770ab5fb73f65800828439a
Reviewed-on: https://gerrit.libreoffice.org/75128
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Fixes regressed unit test failures from commit 3f7e8ddea89f ("Enable
many more unit tests on all archs").
Change-Id: I76a8f7560675b9be3765237f8368d05cff22591d
Reviewed-on: https://gerrit.libreoffice.org/75155
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
This hides all the data, which shouldn't change after init. Real
const makes a lot of problems for copying, so this is the 2nd
option to just add getters for private data. While at it use
typed_flags for the GlyphItemFlags.
Change-Id: Ic1eeabe2398f6c30080fdd516285b72c620b11be
Reviewed-on: https://gerrit.libreoffice.org/75147
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
This depends on commit "Make font-based unit test depend on instdir
fonts, not that it's sure that this really fixes the problem, as its
origin is really unknown.
It especially enables all the font-based tests I could find on all
archs. Same for many more test where I couldn't see any reason they
don't work generally. To get rid of even more ifdefs, it moves these
from the class to the functions, so there is actually just one needed
for any test. As a result some few tests run but do nothing.
There is still some problem with embedded fonts on MacOS and with
delayed graphics loading on Windows, so these ifdefs are kept.
Change-Id: I63f8424e9debda6cbf3e5777c93245e09f8eb0f2
Reviewed-on: https://gerrit.libreoffice.org/74719
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Instead of some manual decoding of the .fot file, this extracts
all the relevant information of the registered font using the
default routines already used when enumerating the system fonts.
We still have to extract the font family from the .fot file to
know the registered family name of the embedded font.
And it uses the OSL functions to create the temporary font resource
file name and generally adds more SAL_WARNs when some part of the
private font registration fails.
Change-Id: I5e118bf1f3bc1fb7bbf4c68a383a45892dffaf98
Reviewed-on: https://gerrit.libreoffice.org/75102
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
This way we can get rid of the embedded fonts without the reqirement
to re-process all of LO's shared fonts.
Change-Id: I25661a611d43ae05052e5cb9cc21e74ccd06b172
Reviewed-on: https://gerrit.libreoffice.org/75101
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
This separates the system registration (AddFontResourceEx) of
LO's shared fonts from the later PhysicalFontCollection
population using EnumFontFamiliesEx.
Calling AddTempDevFont from GetDevFontList creates a new
WinFontFaces, but the later EnumFontFamiliesEx also creates
WinFontFaces for all registered fonts, resulting in duplicate
WinFontFace entries in the PhysicalFontCollection.
Also currently the provided WinFontFace from AddTempDevFont is
much less accurate, compated to the EnumFontFamiliesEx one.
Change-Id: I2cd8af77197429508d0a7112ff98706d7f449659
Reviewed-on: https://gerrit.libreoffice.org/75067
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Change-Id: I9cce47c132345e40d3500ba69178e871d68bf764
Reviewed-on: https://gerrit.libreoffice.org/75130
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Reduces load time of this file by 10%.
This logic is here since
commit 8ab086b6cc054501bfbf7ef6fa509c393691e860
Date: Mon Sep 18 16:07:07 2000 +0000
initial import
and there does not seem to be any good reason to immediately trigger a swapout.
Our balancing logic will perform a swapout as necessary.
Change-Id: I02592dfd3bde1a00bcca2fcbcd4d38aacc8f3de4
Reviewed-on: https://gerrit.libreoffice.org/75086
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: Iac52b72831868a917fc445e970edabc1c2585b55
Reviewed-on: https://gerrit.libreoffice.org/75080
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
This is needed to be able to load e.g. TIFF files in Online, which does
not have code in the chroot since online.git commit
ad21717498cc214d5d5d5c0ff378238552137e2f (Avoid copying most code into
the chroot jail., 2018-03-10).
Change-Id: Ib095191ee690d9da6269878a82b1c6995df50dfa
Reviewed-on: https://gerrit.libreoffice.org/75077
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
i.e. behind BitmapEx
Change-Id: Ibe5a20ffe127acf7fb5bcf6341dcd046371aa761
Reviewed-on: https://gerrit.libreoffice.org/75044
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I7e6994ec6bf7f0c7380df36c49d0b05a9d27c673
Reviewed-on: https://gerrit.libreoffice.org/75041
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Followup to commit 3a99d7f621036, which did a good job of removing use
of GetOutDevType() in OutputDevice initialization. Commit changes
ClipRegionIntersectionRectangle() to ClipToDeviceBounds().
ClipRegionIntersectionRectangle() takes a region and clips it, however
it breaks the contract in Printer and PDFWriterImpl where it does a noop
- any caller on these classes will expect the region to be clipped to an
intersecting rectangle.
Instead, I have renamed it to ClipToDeviceBounds(), which is what the
bClipBounds check shows it is doing in the original code. I have made it
return the new clipped region as this seems more logical.
I have also moved the comment to a doxygen comment on the function
signature in OutputDevice.
Change-Id: I64276920bb452839df04c16f0efc3d5581162673
Reviewed-on: https://gerrit.libreoffice.org/74967
Tested-by: Jenkins
Reviewed-by: Adrien Ollier <adr.ollier@hotmail.fr>
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
See tdf#42949 for motivation.
Change-Id: I89c082d62409b40c8472b865cc60b0d10923c9c1
Reviewed-on: https://gerrit.libreoffice.org/75015
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
Related tdf#93853 make sure we use the opensymbol nice
checkmark (0x2713) to compel viewers to use that.
Change-Id: Id5028b0d062e0491b1cc9c36e2d9d4e4a7ab14a1
Reviewed-on: https://gerrit.libreoffice.org/74995
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
|
|
Change-Id: Ie84b19a3dfaec32184bb825b7593ec33a5c4145c
Reviewed-on: https://gerrit.libreoffice.org/74994
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
|
|
Change-Id: Ie90e0f75018edee178f1f02056e59b35956c4142
Signed-off-by: Adrien Ollier <adr.ollier@hotmail.fr>
Reviewed-on: https://gerrit.libreoffice.org/74464
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: I386e913f9002eed164c26137a0e184993d010b86
Reviewed-on: https://gerrit.libreoffice.org/74090
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
This adds drawing the wave line (typically used to underline the
wrongly spelled words) with bezier curves. Previously the wave
lines were drawn with drawing pixels, which didn't look that good,
especially on HiDPI display, so the looks of wave lines now is
therefor much better.
The creation of the wave line as a polygon has been added to the
basegfx module, so it can be reused if needed.
In addition, everytime we draw the waveline, we have to enable
antialiasing, to have a much better quality of the curves. By
default the antialiasing is disabled for some reason.
This also adds ScopedStates.hxx file which currently includes
ScopedAntialiasing, which sets the antialiasing to a certain state
for the time the object is in scope, and then sets it back to the
original state.
Change-Id: I4b866fc5d69725eb7f6f78a1acf4176b1205aa73
Reviewed-on: https://gerrit.libreoffice.org/74810
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
Change-Id: I2d6e35c68d1ab99a19b5b2256b1f8fae4d614720
Signed-off-by: Adrien Ollier <adr.ollier@hotmail.fr>
Reviewed-on: https://gerrit.libreoffice.org/72327
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
Tested-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
makes OutputDevice::ImplClearFontData clean
Change-Id: Iee0683bd4567f85e20d5017b8eaa8a46490678db
Signed-off-by: Adrien Ollier <adr.ollier@hotmail.fr>
Reviewed-on: https://gerrit.libreoffice.org/72084
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
This reverts commit 624530549f30572cb7fb137cbadecfab9a75fc8c.
Change-Id: I4bc0216cc12af637ad291f3ba6af090d1d34f107
Reviewed-on: https://gerrit.libreoffice.org/74916
Tested-by: Jenkins
Reviewed-by: Xisco Faulí <xiscofauli@libreoffice.org>
|
|
Finished parts from old logger:
1) UNOCommands
2) StarterCommands
3) UIObjectCommand
4) DialogCommand
5) SpecialCommand
Change-Id: Ia521efef0abe1a351b9a4fcabaab6dbf20e8fc89
|
|
Regression from
commit 319c57d2af5d26d3910db4b02dca145d8881af
tdf#120837 File saving at least 5 times slower
which made BitmapEx::operator== rely more heavily on
Bitmap::operator==, which in turn relies on the platform-specific
SalBitmap subclass to calculate a checksum, which was not happening
with WinSalBitmap.
Consequently, make Bitmap::operator== more robust by returning false
if either of the underlying bitmaps cannot generate a checksum
Change-Id: Iad8a4ce19544839a6303f0e8d006b138c2d75a31
Reviewed-on: https://gerrit.libreoffice.org/74834
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I079fb2bb786b66b843e8e3960643cf2db6387c79
Reviewed-on: https://gerrit.libreoffice.org/74821
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: Ie7f40abaa5893bee46d64b85404ee733c0c11f85
Reviewed-on: https://gerrit.libreoffice.org/74797
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I52a0e3aaed903f48c39c900a085cee3cb5131fca
Reviewed-on: https://gerrit.libreoffice.org/74768
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
|
|
a) use EditEngine instead of TextEngine as the former can be hosted in a
foreign widget
b) use a SfxGrabBagItem to hold the custom pseudo-field info inside the
EditEngine
c) use a proxy XDropTargetListener/XDropTarget to massage the treelist entry
drag source into the textengine drop target destination
Change-Id: I6bf1613ed91341c24b77564405f38b5dab0aacd7
Reviewed-on: https://gerrit.libreoffice.org/74699
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: Ic2e36b8921d329d6f9cd2eb68524ed111f6fda75
Reviewed-on: https://gerrit.libreoffice.org/74754
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
This patch does not solve the main LO 6.2 problem in bug 125609,
but fixes the related problem from comments 10/11.
This patch fixes two different scenarios:
1.) Sometimes buttons could fall through the other safety nets and
have multiple radio buttons selected in a single group at display time.
Since ImplInitStyle can be called multiple times for the same
button, ensure that once everything is imported (IsRadioCheckEnabled)
that any calls to this double-check that only one button is enabled.
First come, first served if by programmer error multiple are marked as
IsChecked in the same group. This problem existed pre-LO 6.2,
so it had nothing to do with tabstops specifically.
2.) This patch specifically fixes the 6.2 regression with the
Grammalecte extension Graphic Options dialog box. In this case
Dialog::GrabFocusToFirstControl ended up being treated as a "click"
on a tabstopped, unchecked radio, because setting the state
never called the SetState/Checked function, but only the Init,
and so the uncheckAllOther function never removed any tabstops.
Change-Id: I98272d6c81cf582a15c2e6bf04e6ed60da1d9c71
Reviewed-on: https://gerrit.libreoffice.org/74108
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
this document appears to have just enough fallback fonts on each page to
make the cache useless, so just bump up the size of the cache
considerably, since the cache key/values are really small
Change-Id: I4983de92f66410fee3f04e1e126b2ddb4ec2cbf9
Reviewed-on: https://gerrit.libreoffice.org/74735
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
if we configure something to be dropped on, then make it track drags
Change-Id: I7beacae7012b8929f9bcfa4c72d713ed04ba576c
Reviewed-on: https://gerrit.libreoffice.org/74733
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
And the vcl_fontmetric test must depend on more_fonts. This way
there should never be a missing font, so just assert on it.
Change-Id: If3b71d7c3e52f1beeda73197c56f5efc6a83dbb4
Reviewed-on: https://gerrit.libreoffice.org/74717
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Change-Id: Ib79da121941863421bc12afa27e27cdd7e2e81e3
Reviewed-on: https://gerrit.libreoffice.org/74695
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
SalMenu's event handling is not really compatible with QAction.
LO expects to get something like an activate event, which happens
without any menu state update. The item handler will then update
the item state, as expected by LO.
I'm not sure how this could be implemented in a good way in Qt, so
this patch is just a hack. If the item is checkable, we invert its
state before handling it. This just works correctly for single items,
not radio groups, but it does the trick for this bug. And since
setChecked() just emits toggled(), triggered() won't run twice.
Maybe someone can come up with a better way using event filters, if
ever needed.
Change-Id: Ic32d07525ce8f394b83d04f53689d1669ddc9d83
Reviewed-on: https://gerrit.libreoffice.org/74590
Tested-by: Jenkins
Reviewed-by: Michael Weghorn <m.weghorn@posteo.de>
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Change-Id: I57b667384813b0f6d8412da17c83877a9deed5b1
|
|
The time here is all in the nice rescaling we do these days. Speed it up
by using int ops instead of float ops.
This takes the time from 5m to 1m30 for me.
Change-Id: Ic1dcd9e49eef1894f4a4fdb416015b69c6ef96da
Reviewed-on: https://gerrit.libreoffice.org/74689
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: If863d28c6db470faa0d22273020888d4219e069e
Reviewed-on: https://gerrit.libreoffice.org/74559
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: Ic861b0a593505828a900fe2163125d6f5584a956
Reviewed-on: https://gerrit.libreoffice.org/74634
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I5ed8af625cb46998649c4f11255d7676a11bdc94
Reviewed-on: https://gerrit.libreoffice.org/74605
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I1e173ecd241526b50a8426d75f3aaa13f6ab5c0d
Reviewed-on: https://gerrit.libreoffice.org/74622
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
Tested-by: Julien Nabet <serval2412@yahoo.fr>
|
|
At least maths element docking window has a floating height of zero.
So the original gtk comment about broken values is still true.
And the initial SalFrameGeometry is ok, so no need to initialize it.
And set default size and position on show, if not explicitly set.
Change-Id: Ibe4969a164b50b666745ce010f5100af4b3ff2f6
Reviewed-on: https://gerrit.libreoffice.org/74614
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|