Change-Id: If0759525326f963b497a6e011027719fc376eaf9
Reviewed-on: https://gerrit.libreoffice.org/26226Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>

Need to populate it & register handlers for the non-first windows as
well.

Change-Id: I27231027c3f33fb081768c51ecc291b9f2a7e99b
Reviewed-on: https://gerrit.libreoffice.org/26225Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>

With this, gtktiledviewer no longer crashes on load when opening a
sample ODP file.

Change-Id: I19857d6dcfab74c9fa282754e450c951f7ca564c
Reviewed-on: https://gerrit.libreoffice.org/26223Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>

cf. d150eab8 "loplugin:passstuffbyref: For now
disable 'pass parm by value' warnings".  At least all the other changes in
4d49c960 "Let loplugin:passstuffbyref also look
at fn defn not preceded by any decl" were OK but the one reverted with
b3e93997 "coverity#1362680 Pointer to local
outside scope".

Change-Id: I022125fbcb592e7da3c288c0fd09079dd2e87928

ScRefCellValue in interpreter context has to be used with GetCellValue()
and GetCellString(), using raw ScRefCellValue getString() and getValue()
is not enough.

This also changes behavior with referenced empty cells that
sc/qa/unit/data/functions/fods/Functions_Excel_2016.fods
tests for, but I assume that tested our implementation, not what Excel
does in these cases. Needs checking.

Change-Id: I16194ad59cce3d15271e7610e8ed90ac1786bcaa

the problem is the hidden window used to be the temp parent for
the ref edit and button when this dialog is in shrink-ed to reference mode
from the cell-range option on the criteria page

Change-Id: Ic11299aae159b4f58f2158442b0a787fb0607c03

Change-Id: I05c0b2ba42cf395a2b30ce658e1f1d034e95f21f
Reviewed-on: https://gerrit.libreoffice.org/26217Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>

Change-Id: I9f2be193b995d43d4e3440e55a025629fe5bcaed
Reviewed-on: https://gerrit.libreoffice.org/26222Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>

didn't mean to push that just yet

This reverts commit 9c13ba3b.

to the pre vectorization code layout before...

commit e5d378b1
Author: Kohei Yoshida <kohei.yoshida@gmail.com>
Date:   Tue Apr 3 18:26:29 2012 -0400

    Check for empty-ness of vector before accessing the first element.

    I hope this will keep Stephan's tinderbox happy.  If not...

changed it, now that c++11 data is available to us

Change-Id: Ie5e584c02f51d9cc8dc3b3f65f7cfc90822f65c2

which was added in...

commit 82abd23f
Author: Caolán McNamara <caolanm@redhat.com>
Date:   Thu Nov 12 16:05:44 2015 +0000

    Resolves: tdf#93317 Modified Document Dialog misses focus on Gtk3

because it reportedly causes a windows-only regression

and instead fix tdf#93317 by presenting windows and taking focus
using the last received user input time to ensure that the ToTop
on the main window doesn't take precedence over the following
gtk_widget_show of the quit dialog

Change-Id: I08d1889232af75ca214ccafe9e4e9364df74fbe2

These callbacks were invoked from the model, so need to iterate over the
view shells and invoke them on all the views.

Change-Id: Ie8107f4782b513b874d0b94464ea372994dabbe9
Reviewed-on: https://gerrit.libreoffice.org/26221Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>

Change-Id: I43df4ba611d7b377829ac3b5f17ff26077e9f13d

After registration and invalidation is handled per-view, the scrollbar
and the text control of the comments is rendered correctly.

Change-Id: I5cfa8b43fdc3a2ce7a008269b0283ee25d141440
Reviewed-on: https://gerrit.libreoffice.org/26218Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>

Change-Id: I3cf9e13dafb59031ab8dc4e678de9d502a376d52

Change-Id: I2a7d5d507a875e555441a9d30d54e638e5dea026
Reviewed-on: https://gerrit.libreoffice.org/26099Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
Tested-by: Stephan Bergmann <sbergman@redhat.com>

There may be more elegant ways to accomplish this, go and find one..

Change-Id: Iceaa0783db9cf3d3e1aa20f075fe7e0618a1feb6

Change-Id: I0f4143cb2a34cf18e6c926432d707d84ef0ea579

Change-Id: Iaf93ff58229e9362b03b3e810611e969ef8baf38
Reviewed-on: https://gerrit.libreoffice.org/26206Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

These were the last callback invocations in editeng that did not support
the per-view callback.

Change-Id: Ic703f6708a37f059dcdbbc321d6b18a7fd04edae
Reviewed-on: https://gerrit.libreoffice.org/26216Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>

Change-Id: I1bd6dc18277aa009c23c75b85932d1c93fc13b5e
Reviewed-on: https://gerrit.libreoffice.org/26190Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

- insertByName
- replaceByName
- removeByName

Change-Id: I964aa0dc2e7f5a5be9eaec719b8944e847eb9d6a
Reviewed-on: https://gerrit.libreoffice.org/26037Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

We have two major groups of text fitting options in draw/impress

ones that apply only to custom shapes
 * Word wrap text in shape
 * Resize shape to fit text

those that apply to other elements, e.g. text boxes, legacy rectangles and lines
 * Fit width to text
 * Fit height to text
 * Fit to frame
 * Adjust to contour

Of the second group, only "fit to frame" is always available. Text boxes
have "Fit width to text" and "Fit height to text", while stuff like legacy
rectangles and lines etc have "Adjust to contour" instead.

A fun issue is that (currently anyway) the bit for "resize shape to
fit text" in custom shapes and the bit for "fit height to text" in text boxes
is the same bit, SDRATTR_TEXT_AUTOGROWHEIGHT.

So before this change in all circumstances the same collection of checkboxes
for all settings was visible. With context used to enable or disable which
ones could be set according to the type of shape it was. Simultaneously there
is logic to enable/disable checkboxes depending on if other checkboxes that
control contradictionary options were checked/unchecked. e.g. "Fit to frame"
disabled if "fit height to text" is enabled. So its not apparently why some
can be enabled and some disabled by clicking about the place in the shared
collection.

In this commit we split the sizing options into their two families, one frame and
column for each of "custom shapes" and "everything else". When adjusting
a single selected object we use context to determine which column to show
or hide. When editing multiple objects or the underlying graphic styles we show
both columns. When editing a presentation style we show just the text box
column. (The use of HasText in the original code is a concern, cause it doesn't
make sense to me, using it like that means that in the original dialog
format->text on an empty text box from F2 shows the contour option, but after
entering text and then format->text it shows a set of different text box sizing
options, so that's dropped here)

Because (currently) the same SDRATTR_TEXT_AUTOGROWHEIGHT bit it used for two
apparently different things then we visually toggle on all things that use that
bit in multi-column mode when its togged on, i.e. on editing a style visually
both "fit height to text" and "resize shape to fix text" are kept in sync when
you toggle one or the other. We don't disable the "resize shape to fit text"
checkbox (unlike the fit height to text checkbox which controls the same bit)
if "fit to frame" and "adjust to contour" are set, and give it additional
powers to unset those if clicked.

This hopefully makes the ui describe the way things actually are.

Because this SDRATTR_TEXT_AUTOGROWHEIGHT is currently sort of doing two purposes
selecting a freshly drawn custom rectangle and using format->default formatting
will cause it to change properties because the underlying
SDRATTR_TEXT_AUTOGROWHEIGHT bit is cleared.

The way things should probably be I guess is that there should be a
SDRATTR_TEXT_AUTOGROWSIZE property (which used to exist but wasn't hooked up to
anything) just for custom shapes which overrides the old family of options if
set.

Change-Id: I49241c90d919eeb5caa8775beab57746d5c6df04

Change-Id: Iab1e9888d35ecad7b1ca9713a571f32d59389cd1

Change-Id: I2bce1aba3e01e5c8d67d47b88cddca1f7ad37a93

Also assert that we don't try to invoke the model callback in editeng in
the the view-callback case.

Change-Id: I9f7b7d74da154ac8d0fe3404b7d8114008f74773
Reviewed-on: https://gerrit.libreoffice.org/26211Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>

As per l10n request
https://listarchives.libreoffice.org/global/l10n/msg09997.html

Change-Id: I4f83c85abc15c12f77ac478589736cd799760832

Change-Id: Ia18ad7e283898418118237c1997a6ee72c468341
Reviewed-on: https://gerrit.libreoffice.org/26209Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>

put more useful info on the same line, and indent following lines, to
visually separate them from new log messages.

Change-Id: Ieda71aebf68420a0521ce98c621b0236d533bc53

Change-Id: I6ce045db236b81aa8bc7bce2a0e20c4132c5931f

Change-Id: I5da1a04d2fce6583fad0934d06dcba452181f256

Change-Id: I19ba749ae4cca43602330834f155eca89638c711

Change-Id: I95e42012d8503493a89aee94fd05b2a878535f06
Reviewed-on: https://gerrit.libreoffice.org/26198Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>

Converted libmar into static library and removed duplicate code.

Change-Id: I51c58ae9a6a1f1128422b4a79491d0cd62e6ced6
Reviewed-on: https://gerrit.libreoffice.org/23312Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Reviewed-by: Jan Holesovsky <kendy@collabora.com>

this doesn't seem to make sense anymore since...

commit 4d49c960
Author: Stephan Bergmann <sbergman@redhat.com>
Date:   Wed Jun 8 17:14:34 2016 +0200

    Let loplugin:passstuffbyref also look at fn defn not preceded by any decl

so lets revert that to be safe

Change-Id: I78fe8ae48dc4c4f3d5786c7232d7e791ff50d9b4
Reviewed-on: https://gerrit.libreoffice.org/26189Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

That needs fixing, to check that the parm is not bound to a reference, cf.
<https://gerrit.libreoffice.org/#/c/26189/> "coverity#1362680 Pointer to local
outside scope".

Change-Id: I3656354ccd10affafa006c9e46cf1db608b5b2a7

Change-Id: If054c98f887b62adaf6c73ee35e664507a69b6df

Change-Id: I3567c3f9c5fe8374abd81d3f6e9a4a41e6a4c23d

Change-Id: Ideec43617ff214b7e0dcbb0a1969955b34ee19e4
Reviewed-on: https://gerrit.libreoffice.org/26196Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>

Change-Id: Ib28e76ed6947431c6466d11aaadca8d95abfc275