Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: Ic3667c0daeeeb5b09eeaae66899affda65e2e9a5
|
|
Looks like the Windows specific code in ImplUpdateDisplayBmp is
not needed anymore and causes problems with OpenGL. The icons after
the change still seem to be drawn correctly with OpenGL enabled
or disabled.
Change-Id: I3ae1a0ceb947254aaadbc3d772f0d662b291b594
|
|
This adds recoloring of images/icons (suited for breeze and sifr)
so they are more visible when a dark theme is used (for example gtk3
can switch to a dark theme). LO must be started with environment
variable VCL_ICONS_FOR_DARK_THEME set to 1.
open issues - currently all images/icons are recolored but not all
images should get recolored.
Change-Id: Ibc42e30af79bb4c4f04c67b760019311b97b2cc6
|
|
Change-Id: I787be4f2da514017fd14100f56ed323d30dbe99d
|
|
Change-Id: Ib9a493af0f64c5f3eb50f392805c3278ed2164a0
Reviewed-on: https://gerrit.libreoffice.org/19304
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
|
|
"Is the mutex really necessary? Normally outside of the chart OpenGL
only the thread that has taken the SolarMutex should access the OpenGL
code. In general we have bigger problems if several threads start
rendering in parallel."
Change-Id: I5f8fa8c771d5e2e0df4cfe903b15d00e5cc162b1
|
|
Change-Id: Ia9b95232f14b2a6f5c06c0720a91f44217f5c80c
|
|
so close all open frames, not just the active one
and send events to the first discovered visible window and not the focus
window
Change-Id: I7339bc6743cd3106ceddea049136796121819eb3
|
|
uninitialized memory warnings on --convert-to pdf
of ooo115771-3.doc
Change-Id: I2dac6fe29d23d6e25c2f1d1511886d8b585e498c
|
|
Change-Id: Ic7c496259258bf36cd885ed68912b43117f348f0
|
|
Change-Id: I11b99fca91b266bdf6c29dc142511bf01dbd4bed
|
|
Change-Id: Ib9eae95b8ace37d31e905bc098c67de97b7816e3
|
|
...to silence a clang-analyzer-deadcode.DeadStores on the last update of nLeft
and nRight, each. Also, adding a gap of 3 to nRight in case no further buttons
will be layed out next to maDockRect resp. maHideRect is irrelevant, as nRight
will not be used for any further layout computations in those cases.
Change-Id: I251d4470099b1f98782ebdfd0c776d881bb7fae5
|
|
i.e. to reset to the initial post-startup state
performance still isn't great unfortunately, 0.7
cycles per second before this, 2.6 after this
Change-Id: Ib9d83b70f0cf052af36a49ed5bacef295fe9ed11
Reviewed-on: https://gerrit.libreoffice.org/19269
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I8a6ea395192d31391868f44152ab3b543a07aeaa
|
|
Change-Id: I9e2cf9b06fcf0b0a9dfd642fa788f203ade19fe3
|
|
...assuming the success of both steps 2 and 3 should be checked (though the of
this code is lost to history past 40d960cbdba0dbd445d624f62a523b7e78d5f5c9
"divide pdfwriter_impl.cxx")
Change-Id: Icea5de98e5a0559dbc0f837650e842fb49bc41e0
|
|
Change-Id: Ie1d3afb0b9ca59232214d9ab6da9fa761a3fd449
|
|
...to silence a clang-analyzer-deadcode.DeadStores on the last update of nUsed
Change-Id: Ib470f09155bd181b05bc0342543a36e59b8dd937
|
|
nSxe, nSye were unused ever since WMFReader::GetPlaceableBound was introduced
(seemingly, involving some copy/paste) in
fa149f402b9222d361e79769933bc6c58101ddef "#98505# if there is no WinExt we try
to calculate the extension instead of using (1/1) default which causes a crash
on unix and linux"
Change-Id: Ica6fb9ec1d25bd5314e73610eccef2a4df0c310f
|
|
Change-Id: Icef974fccdc487b8abef3b1200b0b34b7be703ec
|
|
Change-Id: I4924a3cecb1e12874659eeccde51126fab4d4e45
|
|
Change-Id: Ic76cf540896ba217f7df6045be31197b64298bf3
|
|
Change-Id: Iabb2d50cced788ab49d21e1aadcf5ab685107511
|
|
Change-Id: Iad11e6bcb64976c7eb4181ee327602e6c3a7e26f
|
|
...follow-up to 63899beb6c033f4dc305a046c558c348a907ea40 "WaE: unused variable"
removing the nVerticalSamplingFactor part
Change-Id: I2830dc0647cf651014e1199ef62a7ea50e4a1719
|
|
Change-Id: Ic2d49519475db8292fecf18eb5c90a74ace8dfe7
|
|
Change-Id: I68ab6da7c90b869499f1b7a4d2433dac77d36889
|
|
Change-Id: I499f00e13543d2191008a0d7a4aa0391d6889315
|
|
Revert partly http://cgit.freedesktop.org/libreoffice/core/commit/?id=76f33f10309b0ee384a75a7a854858b068d60495
author Julien Nabet <serval2412@yahoo.fr> 2015-03-16 21:31:23 (GMT)
committer Caolán McNamara <caolanm@redhat.com> 2015-03-26 14:14:44 (GMT)
commit 76f33f10309b0ee384a75a7a854858b068d60495 (patch)
tree e12e3b6f2ca46cb9998b93b5c1bb3406408ad3cb
parent 6c4100cf81e1678ff386549c0cdaa2f73a154ddc (diff)
tdf#67990: Management of case in combobox
See comments of tdf#94557 for more information
Change-Id: I0432db0c3f76c20e55d8fecb69734952dc0cf53b
Reviewed-on: https://gerrit.libreoffice.org/19260
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
|
|
otherwise we get stuck if the menubar gets activated
Change-Id: I23a613de41202fa0a542d1da6e10d190225a5f44
|
|
Change-Id: I10b4badaba01630e1a054b3d4d9cc4720822511a
|
|
which uses PostUserEvent like Application::Quit so can ensure
EndAllDialogs gets called before Quit
Change-Id: I9eb7dbd4f012be2e9b9cea7a73c90c5aaba141aa
|
|
regression from 1e67e94f1a308ca60d4934e9fe9d5c048225ebe8
"convert Link<> to typed"
Change-Id: Ibb20366f0edcb9425d61ce5e9d5fb1a34533785b
|
|
Change-Id: Idd4a1e6d50652a879493d8411c59605ca1a53dfb
|
|
Change-Id: Ie313009ee0d10872f5e705d017e6a73ff83c2761
|
|
Change-Id: I59f5a38f6c62b97eb73d85177be408996568e2cd
|
|
the mpStr and mnLength arguments are always created from an OUString
and the OUString aRun(rArgs.mpStr); creates a duplicate of the original
string. Its way faster to pass the original OUString around instead.
Change-Id: I3de0cc31654eb215362548b6aee435f9c0c3c9db
Reviewed-on: https://gerrit.libreoffice.org/19221
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
close all open dialogs on quitting to avoid triggering
the hang timeout and get things moving faster
Change-Id: I7469a32549f20d6b0971932e3334cb322f57b1ff
|
|
time stamp doesn't matter apparently for keyboard events
so ditching it gives less surface area to fuzz
repeat showed some interesting effects, but mostly just allows
vast paragraphs to be generated causing timeout hangs which isn't
what I'm super interested in right now.
Change-Id: I955b91745a26d5044f0d53acd3fda0a160e53855
|
|
because fontsubst.cxx is not built on mac.
move PreMatchFontSubstitution to PhysicalFontCollection.cxx ?
This reverts commit 349700228e4449ace956b30bc65208b9196f0a78.
Change-Id: I007f24313ddfd0eb3db3fd599abcf1eae3da4e9c
|
|
Change-Id: I9f56beebae1ea02a8914b7b25fe7565f7b0df053
|
|
Change-Id: I0cb515a8b394c5e4b43fe84282383872c0680f15
Reviewed-on: https://gerrit.libreoffice.org/19216
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
Tested-by: Julien Nabet <serval2412@yahoo.fr>
|
|
Change-Id: I1933951c52adc75ed36db2c083c232f29b6140d6
|
|
Change-Id: I150baadc442e57ee604563bc52965daa9d2e41af
|
|
replaced use of PTR_CAST, IS_TYPE, ISA in
idl, editeng, sc, sd, sw, sfx2, sot, starmath
Change-Id: I4a5bba4fdc4829099618c09b690c83f876a3d653
Reviewed-on: https://gerrit.libreoffice.org/19132
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Oliver Specht <oliver.specht@cib.de>
|
|
and move the boost-depending code inside the module, no need to expose
that dependency and include that header in all client code.
Change-Id: Ib991572d7a54a9ad16825d688e76f8e3a824803c
|
|
Currently we create a "disabled" version of bitmaps in "Image"
class (used for toolbar icons for example) in every draw call.
This is not really optimal. The solution is to calculate the
checksum of the original and create a "disabled" version only if
the checksum changes.
Change-Id: I6ed2d08c64fee028b479fedb35e4384f112e98e2
|
|
Change-Id: Ic3f7fddcea36c18ffe43c4c633d415f596a58cbc
Reviewed-on: https://gerrit.libreoffice.org/19094
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>
|
|
Non ASCII characters like åäö can be used as accelerator keys (Alt+<x>).
This already works however at the moment it is not exposed properly
to accessibility tools such as screen readers. After this patch at least
NVDA gets information about accelerator keys that are non ascii. On
Linux we need to adapt code in vcl/unx/atk/a11y/atkaction.cxx.
There is also vcl/source/window/dlgctrl.cxx#Window::GetActivationKey
should probably have similar treatment and I guess that it is related
to accelerator keys in dialogs but I need to confirm that.
I'm a bit unsure how to best do that so I'll save it for later.
Change-Id: I1f1509a26d781011d217670266bd10b2107811d5
Reviewed-on: https://gerrit.libreoffice.org/19078
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|