Change-Id: Ief06d3ea0288e8ebf34f2f67f3d1dfc9405b6d49
Reviewed-on: https://gerrit.libreoffice.org/29180Reviewed-by: Jochen Nitschke <j.nitschke+logerrit@ok.de>
Tested-by: Jochen Nitschke <j.nitschke+logerrit@ok.de>

This reverts commit d141d5b3.

seems like I pushed it too soon, it depends on other commits not in yet.

Change-Id: Id4d3f80b3906395838afaffb2b0d00437d17f6dd
Reviewed-on: https://gerrit.libreoffice.org/29182Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>

reduced size because it is stored in widely used SfxPoolItem.
converted 'if' chains to 'switch case'

Change-Id: Id9b5e375b681c88e8c91c561abd1a50610aa4815
Reviewed-on: https://gerrit.libreoffice.org/29186Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>

In zforlist.cxx the function SvNumberFormatter::GetPreviewStringGuess
switches between the system locale and en_US in order to guess the
formatting of a number.  This results in poor performance, because in
the created SvNumberFormat only the last used locale is cached.

Caching more than 1 entries improves the load time for the document
attached to tdf#53698 from 90s to 60s in case of non en_US locales.

Change-Id: Id0eb4447dea6213015e2d958d8303a1e7892487a
Reviewed-on: https://gerrit.libreoffice.org/29136Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>

Change-Id: Ifb6045885049733415895f58cdd911256f48323c
Reviewed-on: https://gerrit.libreoffice.org/29187Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>

Change-Id: I70fcf95dfd3db05b4fd6e5cee37866f673d3afa8
Reviewed-on: https://gerrit.libreoffice.org/29183Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>

The problem, as seen by the user is that during shape text edit,
sometimes just traveling with the cursor causes invalidations, which is
unexpected.

What happens is sd::Window::KeyInput() invokes
sd::DrawViewShell::MakeVisible(), which does a pixel-to-logic
conversion, but because the map mode is in general disabled in the LOK
case, nothing happens. Then a bit later aVisArea.IsInside(rRect) fails,
as it compares pixel and logic units, which results in
sd::ViewOverlayManager::UpdateTags() scheduling an async call to
sd::ViewOverlayManager::UpdateTagsHdl(), which at the end causes a full
invalidation in sd::SmartTagSet::remove().

Fix the situation by temporarily enabling map mode, so we don't detect
a visible cursor area as a non-visible one.

Change-Id: I6d16f24d13143dc263a89317fbc38111e652c0ac

Change-Id: Ice2657c8e8ecccb67d1b14292514a42ff37caa39

KDE theme engines can explicitly support rollover menubars via
styleHint( QStyle::SH_MenuBar_MouseTracking ).

So this adds support for this feature to the vcl::Menu.

Change-Id: Ica923ff2d06cfd54e548ba858b3d90f70d9c255a

Selecting radio and check boxes is handled via
QStyleOptionMenuItem::checked. Actually forwarding the
QStyle::State_Sunken creates strange artifacts in all
menus.

Change-Id: Ia0b0864b9bec5961023b4f25315a8acf1fae4683

This will always redraw the progress bar over the last one.
It stops the flickering at least in KDE4 from re-painting the
background, but obvously will break any transparent or alaiasing
effect, if it was used.

There is actually a mbProgressNeedsErase in the global settings,
so actually use it when setting a new progress value.

Change-Id: Ic62784da4f6d08bf0e03a9e04848a5cc687ee236

Bug introduced by commit 4f5fe008.
The code was moved from ImplInitSettings to ApplySettings and somehow
a not (!) got lost.

Change-Id: Ie3fa4fb99d030d106aab712cd6bdafada8c6d423

Libre office just has a general disabled color. Instead of the
general QPalette::Mid, use the color QPalette::WindowText of
the palette QPalette::Disabled.

Change-Id: Ie6943b455d3e6545289b241ee4ab85128f4fd7ab

... and just fill the painted image once at the beginning.

Also adapt background based on the control information from Qt4.

Change-Id: I39c03b2ca779381920e2597b4bea7973fea6ae2a

Change-Id: If8778c62ec7c88e036c86ea7cb1fb018c251235e
Reviewed-on: https://gerrit.libreoffice.org/29147Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

Change-Id: Ia185cbf7b72ffe014b0f2a3f99bf21c0a33cee06
Reviewed-on: https://gerrit.libreoffice.org/29153Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

if you request a size you might not get it.

This attempt dating through...

commit 6dc1d270
Author: Caolán McNamara <caolanm@redhat.com>
Date:   Mon Jun 15 10:56:33 2015 +0100

    another stab at tdf#91393

    block paints only if the new requested size is larger than the original and
    unblock on explicit expose events as well as configure ones

...

commit 8f324aeb
Author: Caolán McNamara <caolanm@redhat.com>
Date:   Sun Jun 14 15:49:56 2015 +0100

    Resolves: tdf#91393 autotext (etc) not fully drawn

...

commit e6a19560
Author: Jan Holesovsky <kendy@collabora.com>
Date:   Fri Jun 12 15:36:03 2015 +0200

    tdf#91301: Don't cache incomplete tabs.

    After introduction of the Idle processing, something has changed so
    that the underlying GetGdkWindow() does not update its size fast enough;
    even though the gtk_window_resize() is called before the Window::Erase()
    (that actually paints the background) etc.

is all junk. I don't see the original problem of tdf#91393 or tdf#91301 anymore
under gtk2 after reverting this yet anyway to see if they need another fix.

Change-Id: Ide071eba279de726a9c6a80884bc9021c8914d9d

This did not take effect anyway because unxgcc.mk unconditionally
overwrites it; see
https://lists.freedesktop.org/archives/libreoffice/2016-September/075218.html

Nobody knows why it was added so just remove it.

Change-Id: Ib1a6d06ed8e380e63c588a72b6a65bfbeaacb924

This did not take effect anyway because unxgcc.mk unconditionally
overwrites it; see
https://lists.freedesktop.org/archives/libreoffice/2016-September/075218.html

Change-Id: I33a2ae9aae44d974d6b5466a2d7dae61ae5cc92c

... since commit 8541943b

Change-Id: I56b12ebd59b8387878f20d625381d3f743558a43

Change-Id: Ib0cd605cf8e71e94f4442c07300726aff1d9ce03

Change-Id: I4da3374e31d3c7407b401d66275da7f56ae83d30
Reviewed-on: https://gerrit.libreoffice.org/29178Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>

Change-Id: I02eb887101f2ebe9eb5abe8e2291f8cf5d79d2ce
Reviewed-on: https://gerrit.libreoffice.org/29141Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

Change-Id: I7b1aa65a557f3855939ad8b183bba8ff656946f8
Reviewed-on: https://gerrit.libreoffice.org/29173Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

Change-Id: I8b4430f9f8282ef7a6b23964ca51d3444a450305
Reviewed-on: https://gerrit.libreoffice.org/29174Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>

That's where the non-view variant is emitted, and SwVisibleCursor
said visibility=false even if the cursor was expected to be
visible.

Also fix the test callback that checked if the payload is equal to the
"true" literal, but actually the payload was a JSON.

Change-Id: Ifa0e23eb274925e8ab8ef5e7c21370ad2cd22282

Change-Id: I72eea07e775fa1987081b952523dc4e32483fc8a

In preparation of moving SfxSimpleHint that has GetId() to SfxHint.

Change-Id: If781ba9578931d9543875ee288ce37c727bfba10

Removes 220 warnings from build log (at least under Windows).
Requires Python 2.5+

Change-Id: I06b18b14a08f2c27528fcb009faddff081ffd54e
Reviewed-on: https://gerrit.libreoffice.org/29172Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>

... in ResourceMenuController based top-level menus.
MenuBarManager doesn't set frame action listener for
top-level. This wasn't an issue when top-level always
was the menubar, but now top-level can be a popup menu.

Change-Id: Ibc9116389239eb06f84a2e53b31eacc0a09d438e

This unit test class relies on calling waitForEventIdle to wait for
UI transitions to complete.  On Windows this is not reliable, due to
VCL's use of PostMessage (Windows message queue) to asynchronously
handle some events such as focus changes.

This issue causes the test to intermittently fail when the dispatcher
for the action the test wants to perform (like closing the print
preview) is not available yet.

 - Change the test code to explicitly wait for the desired dispatcher
   to be available

 - Remove the sleep which was added earlier this year to address this
   symptom in one location

Change-Id: Id701209a28cd0c806ab0a548508d31f9f11211f5
Reviewed-on: https://gerrit.libreoffice.org/29167Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>

Change-Id: Id952f3dc236d2dc2c6839ce75943bda162fabf82

Change-Id: I9d4b4e55673f73dd6c518bb169416958fde72a32

Change-Id: I96956d6631ed574c06df62e9b633d7e1e4d5a5f2

* each module have own list of implementations
* deleted placeholder .ui files from Calc and Impress

Change-Id: Ib7d4cafbb332447e769ca753003d31e9f3025f0a
Reviewed-on: https://gerrit.libreoffice.org/29143Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>

Change-Id: Iec52b93746b2fd9bbd6ad67667092717f39222d7

..on the new sheet

Bug introduced with
    commit 86e9ef21
    Author: Noel Grandin <noelgrandin@gmail.com>
    Date:   Tue Mar 8 20:41:21 2016 +0200
    sequence->vector in sc

Change-Id: I7a6539005f42c7a7036c60b12a9ff2f77f71ba6b

Bulk scope prevents repeated broadcast/notify/track cycles of identical areas
during mass broadcasts.

Brings insertion time of tdf#87101 bugdoc from 3min30 down to 2min30

Change-Id: I360fd52b622a8a327cdede47ed398260b7e8b4f6

This reverts commit 0792aef9.

It created a bottleneck with massive use of OFFSET() functions mixed
with other references, see tdf#87101 bugdoc. Will have to come up with a
different approach to solve the "broadcast cell move" problem.

Additionally, the unit test for tdf#83067 and the original problem still
work with the reverted change, so we seem to generally have fixed
something else in the meantime..

 Conflicts:
	sc/source/core/data/formulacell.cxx

Change-Id: Ida2e26398f99ff5878d43ffb46035bc4441dfff6