Revert "Resolves: tdf#119688 under X full-screen doesn't work before gtk_widget_show"

This reverts commit 88fc17e4.

Change-Id: I5f2996c0d940eec459725cc91c5dee7fd7dee267
Reviewed-on: https://gerrit.libreoffice.org/66502
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Follow-up to d8106b6d "Non-standard sed -i
option is different on macOS than with GNU sed", thanks to mst for spotting it.

Change-Id: Ida51cfa7c4dd216a5164df4f8143ac89cf2c6b7f
Reviewed-on: https://gerrit.libreoffice.org/66498
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

At least <https://ci.libreoffice.org//job/lo_tb_master_win/20394/> once failed
with

> Value in Thread #1 is 0
> Value in Thread #2 is 5
> C:/cygwin/home/tdf/lode/jenkins/workspace/lo_tb_master_win/sal/qa/rtl/doublelock/rtl_doublelocking.cxx:202:rtl_DoubleLocking::getValue::getValue_002
> assertion failed
> - Expression: nValueOK != 0
> - getValue() failed, wrong value expected.

and there is indeed no guarantee that either of the OGetThreads has already been
through its while loop by the time the main thread calls terminate().  So make
execution of the threads deterministic (and take out the sleeps that are no
longer necessary, and use the now-deterministic expected values in the result
checks).

Change-Id: I61726906f646ffae7a21b1e08b1d973a4e51265b
Reviewed-on: https://gerrit.libreoffice.org/66494
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

Change-Id: I6cdfc50b2385c426e20ce0e9b216b18c763249b8
Reviewed-on: https://gerrit.libreoffice.org/66506
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: I439bf27b3f57838d9d0ea19605fd1b684ad4f777
Reviewed-on: https://gerrit.libreoffice.org/66474
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

The default case is that all managers have assistants/employees, so
nodes under a manager can only use the horizontal space under the
manager to avoid overlapping.

But in case the previous / next sibling of the manager have no child
nodes (assistant/employee) then we can use that space to make the child
nodes larger. This improves readability of the chart's text a lot and
brings the layout closer to what PowerPoint does for the same input.

Handle all this in the hierChild algorithm, i.e. the container for a
list of assistants or a list of employees, which means "parent" in this
context always refers to a manager node.

Change-Id: Ib4125ea2a113339ab7bbcd78e43c5d204f442996
Reviewed-on: https://gerrit.libreoffice.org/66504Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins

Change-Id: I36b2bda0fcb0bfd14e62c2529ecb79e11eab34f2
Reviewed-on: https://gerrit.libreoffice.org/66482Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins

Change-Id: I56058a2069600a4c8acba861a38e7e5d84f44d81
Reviewed-on: https://gerrit.libreoffice.org/66499
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: I9400c8fa5b7763debc13274f7e001b6425d8392e
Reviewed-on: https://gerrit.libreoffice.org/66496
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

The scenario is like this (see the unit test):

* double-click a word to select it (the handles appear)
* click somewhere else (the selection and handles disappear)
* double-click the same word again (the handles did not appear in this
  case)

The reason was that the old state was remembered and the now thought
duplicate state was discarded.

Change-Id: Ia49200f12907c520067258b7570d4e21b365a8dd
Reviewed-on: https://gerrit.libreoffice.org/66171
Tested-by: Jenkins
Reviewed-by: Jan Holesovsky <kendy@collabora.com>

While with GNU sed the optional backup suffix is given directly as part of the
-i[SUFFIX] (or --in-place[=SUFFIX]) argument, on macOS it must unconditionally
be given as a (potentially empty) argument following the -i argument.  That
means that on macOS the -e that happens to follow the -i in these sed
invocations (and that happens to not be necessary to introduce the following
script in these invocations) was mistaken as the backup suffix, causing creation
of pointless workdir/CustomTarget/*/generated/tokens.cxx-e backup files.

Change-Id: Icd865c76139afa348d03813eb83653afd7221a0e
Reviewed-on: https://gerrit.libreoffice.org/66491
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

Found while checking include/vcl/fontcapabilities.hxx

Change-Id: Ib598978672d08d3cae7bf6f96d1b874ca13fcb5f
Reviewed-on: https://gerrit.libreoffice.org/66254
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>

- also use std::make_unique
- also avoid multiple declarations in a single statement

Change-Id: Ib5429730fcf14131b7b68628a8724a812dcc018f
Reviewed-on: https://gerrit.libreoffice.org/66479
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>

Change-Id: I2c32a47fac06b2026d61a0f369e6d569c9f57502
Reviewed-on: https://gerrit.libreoffice.org/66392
Tested-by: Jenkins
Reviewed-by: Bartosz Kosiorek <gang65@poczta.onet.pl>

Change-Id: Ie0baed90e3d0f51ec15c110dbd7ff41dcf7baa40
Reviewed-on: https://gerrit.libreoffice.org/66324
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

Use range-based loop or replace with STL functions

Change-Id: I261dcba1778a9d895491452fb3fee3dd4e8dfc33
Reviewed-on: https://gerrit.libreoffice.org/66422
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

Change-Id: Ib920de1dfbe5ddf4c8576a00fcc7a557b5a6b2da
Reviewed-on: https://gerrit.libreoffice.org/66492
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>

Change-Id: Icc57e3d06c13b444bbbd2936112daafbe410c34b
Reviewed-on: https://gerrit.libreoffice.org/66489
Tested-by: Jenkins
Reviewed-by: Jens Carl <j.carl43@gmx.de>

Move XElementAccess Java tests to C++ for ScTableConditionalFormat.

Change-Id: Ib2575920e05d56feed68c6f4c15b485ab7993600
Reviewed-on: https://gerrit.libreoffice.org/66488
Tested-by: Jenkins
Reviewed-by: Jens Carl <j.carl43@gmx.de>

Move XIndexAccess Java tests to C++ for ScTableConditionalFormat.

Change-Id: I4eb40fb4c44967f6d6dba7ea093818adfc4aacfb
Reviewed-on: https://gerrit.libreoffice.org/66486
Tested-by: Jenkins
Reviewed-by: Jens Carl <j.carl43@gmx.de>

Move XNameAccess Java tests to C++ for ScTableConditionalFormat.

Change-Id: I675d5d5aea24f76c19507d2a4181715ab70dd48d
Reviewed-on: https://gerrit.libreoffice.org/66485
Tested-by: Jenkins
Reviewed-by: Jens Carl <j.carl43@gmx.de>

Rename ScTableConditionalFormatObj to ScTableConditionalFormat to match
the object name to test against.

Change-Id: Ic14bc1181a6c0724e0973955119d9f2e731d0940
Reviewed-on: https://gerrit.libreoffice.org/66484
Tested-by: Jenkins
Reviewed-by: Jens Carl <j.carl43@gmx.de>

This change fixes displaying of elements of 'Styles' menu of Writer

Change-Id: I455a4f1c37aa7bc6641ebdd15eec23f18a8430d9
Reviewed-on: https://gerrit.libreoffice.org/66387
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>

Remove unnecessary destructor

Move MenuBar clearing out of DoFullMenuUpdate function

Change-Id: I0fb916c6935b2f8a675e75adf696d99203a2c016
Reviewed-on: https://gerrit.libreoffice.org/66386
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>

When QEvent::KeyPress event is processed, QEvent::ShortcutOverride event is created
and processed before finishing processing original event.

If it's an active shortcut, and QEvent::ShortcutOverride event is not accepted,
the shortcut is activated, but QEvent::KeyPress event is omitted.
If it's an active shortcut, and QEvent::ShortcutOverride event is accepted,
the shortcut is not activated, and QEvent::KeyPress event is sent as usual.

But it looks like even if shortcut activation is suppressed,
shortcut is still called somewhere from insides SalEvent::KeyInput
event processing callback from Qt5Widget::handleKeyEvent function.

Shortcuts are still assigned to menu actions
since it shows the key combinations used to activate them.

Change-Id: I1ffc63c2530dd61ae67ab66a96ea3efe767218c0
Reviewed-on: https://gerrit.libreoffice.org/66444
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>

Change-Id: I57aea80f0489c4176990d7233a4c218e5199a993
Reviewed-on: https://gerrit.libreoffice.org/66467
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

* Update helpcontent2 from branch 'master'
  - Mute and simplify l10n for GetPathSeparator()
    
    Change-Id: I5b5736f56a21bb6abef75519557d60b8494a09be
    Reviewed-on: https://gerrit.libreoffice.org/66454
    Tested-by: Jenkins
Reviewed-by: Olivier Hallot <olivier.hallot@libreoffice.org>
    

* Update helpcontent2 from branch 'master'
  - Format string and save l10n work
    
    make Syntax: and Return value: headings same as other functions
    
    Change-Id: Ie715b6c9172ebb8154051a8083b1e2c246d93ba8
    Reviewed-on: https://gerrit.libreoffice.org/66481
    Tested-by: Jenkins
Reviewed-by: Olivier Hallot <olivier.hallot@libreoffice.org>
    

...and make sure that the generated name is actually short enough to be stored
in a sockaddr_un::sun_path.  (oslPipeImpl::m_Name is only used to store such a
name, to be used in osl_closePipe, so just make it of the same size to guarantee
that copying between the two with strcpy will always work.)

Change-Id: I6d1d0c5518e6d09eff129d682a1a334d12201e90
Reviewed-on: https://gerrit.libreoffice.org/66469
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

Change-Id: I00a4b6aef307621b1607abbe75b3a64e6f8fb40a
Reviewed-on: https://gerrit.libreoffice.org/66464
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

Change-Id: Ie27432549e835622ab8480acfb3ed613aef623ab
Reviewed-on: https://gerrit.libreoffice.org/66411
Tested-by: Jenkins
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@libreoffice.org>

Call to pFormat->HasPositiveBracketPlaceholder() with nullptr as
obtained for the not existing nIndex was the immediate cause of
the crash.

Also obtain the type from the format if found instead of the
separate GetType() call that again would need to find the format
first. This clarifies that the other places accessing pFormat
don't need additional checks because eType isn't UNDEFINED there.

Change-Id: I5d78ad73eeb4c3820c45d6a91b0e729ed22951a4
Reviewed-on: https://gerrit.libreoffice.org/66476Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins

... which kicks in if SvxNumberFormatShell::FindEntry() determined
a format index of NUMBERFORMAT_ENTRY_NEW_CURRENCY instead of a
real existing index.

Regression from

    commit e607f1da
    CommitDate: Thu Oct 4 15:02:13 2018 +0200

        Select the current format if multiple instead of first matching format code

The old (now removed as unused)
SvxNumberFormatShell::GetListPos4Entry(const OUString&)
used before took care of that, add the functionality back but
combine into the existing GetListPos4Entry() function, the
additional FindEntry() call wasn't necessary anyway because it was
just called before in the caller.

The actual crash happened elsewhere in
SvNumberFormatter::GenerateFormat() because no format exists for
the then determined entry, which should be changed independently
to prevent such abuses.

Change-Id: Id42efa152693119fbde82b234f367679f818cfab
Reviewed-on: https://gerrit.libreoffice.org/66472Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins

Change-Id: Ibba0464d17a9517eb48f3f33b46db2455093ac52
Reviewed-on: https://gerrit.libreoffice.org/66413
Tested-by: Jenkins
Reviewed-by: Jens Carl <j.carl43@gmx.de>

... which decides whether number cells are stored as numbers
or quoted like strings, as documented in
<https://wiki.openoffice.org/wiki/Documentation/DevGuide/Spreadsheets/Filter_Options#Token_8.2C_csv_export>
but long unimplemented.

Note that in this implementation the false value of the above token
is superseded and ignored if either "Fixed column width" (1st token)
or "Save cell content as shown" (9th token) option is on.

Change-Id: Ib4ff02a2be81a8590e1fc249725f02cd83e91118
Reviewed-on: https://gerrit.libreoffice.org/65604
Tested-by: Jenkins
Reviewed-by: Eike Rathke <erack@redhat.com>

its nigh impossible to have a GtkNotebook without an active tab, so
try and theme it into invisibility with no width

Change-Id: I312b81ae51abe9fdd222fb5d5c45cb41603b80cb
Reviewed-on: https://gerrit.libreoffice.org/66443
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

no need to store such small struct on the heap

Change-Id: I8d9c12082a40e1d24cf7d25cca7369e1bb034c7a
Reviewed-on: https://gerrit.libreoffice.org/66323
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

(cherry picked from commit 81b26cc1)

Change-Id: I5f443d60959490f54063288463ed497ed3cb07e1
Reviewed-on: https://gerrit.libreoffice.org/66438
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: Ia01e95ea80c99881756dd5bd5d6e0b82c305329a
Reviewed-on: https://gerrit.libreoffice.org/66456Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
Tested-by: Julien Nabet <serval2412@yahoo.fr>

Change-Id: If1187cbb428d329fa10070662282d7fc3aeaf9de
Reviewed-on: https://gerrit.libreoffice.org/66441
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>