configure uses $PYTHON to determine lxml availability, so the make rule
should use the same instead of letting gla11y find a python interpreter
through "env", in case e.g. $PYTHON is python3, and lxml is available in
python3 but not in python (actually python2).

Along the way, rather use PYTHON_FOR_BUILD for coherency.

Change-Id: Ied4c05aca462cc16685c61f36b56bb9e8612d90e
Reviewed-on: https://gerrit.libreoffice.org/50087Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

Change-Id: I358e9673edcba81336b3895479e787d981fad7a5

This was more or less consistent before, but the recent automatic
rewrites broke consistency, just give up on manual formatting here.

Change-Id: Ieb387e3d7230d227e33ba0b420cfa572309c5330
Reviewed-on: https://gerrit.libreoffice.org/50119Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

There were also two cases here:

- changing the selection with the keyboard or single-click only updated
  the selection in the navigator
- pressing enter or double-click also selected the shape in the main
  window and gave the focus away

Introduce a 3rd case for single-click: update the shape selection but
doesn't give the focus away. This way double-click is not needed to sync
navigator -> main doc selection but keyboard navigation should still
work.

An additional trick is to make sure that the current shell is the draw
shell (and not the slide sorter) after navigation, so follow-up
operations work with the selected object and not with the whole slide.

Finally, a third related problem was that the selection jumped back to
the item of the slide after clicking on a shape in the navigator. The
reason for this was the navigator list was constantly cleared and
re-filled in SdNavigatorWin::InitTreeLB(), as
SdPageObjsTLB::IsEqualToDoc() returned false (even if the list was up to
date) in case of shapes which had children but no name. Fix this by
using the same SdrIterMode::Flat iteration mode that
SdPageObjsTLB::AddShapeList() does, so the fill and the equality check
of the navigator iterates the same way.

Change-Id: I0bfc3e8b49f7ef01d5797a68284616dcd2a81c5d
Reviewed-on: https://gerrit.libreoffice.org/50118Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>

Change-Id: Ia9ec032eb40e4152e881b466311f7e34c845addc

Change-Id: I34e28a30a4f1fd264c18c901cd94094531543271

Change-Id: Ia29e495130f5b83f3422d68597dedcb3b0b27a07
Reviewed-on: https://gerrit.libreoffice.org/50120Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
Tested-by: Stephan Bergmann <sbergman@redhat.com>

...as it is included from other include/vcl/*.hxx files.  (And moving it there
helps the heuristic in an upcoming improvement of loplugin:dyncastvisibility.)

Change-Id: I15e061fbaba8e82235afb5b312f7b38f26eec900
Reviewed-on: https://gerrit.libreoffice.org/50123Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

Change-Id: I224cc955d49ee100d328e0171da710f38068d2d4
Reviewed-on: https://gerrit.libreoffice.org/50114Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

Change-Id: I3d87ee59258567b5a82cb459248539faea1b3ca9
Reviewed-on: https://gerrit.libreoffice.org/50101Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

and fix the sal_Int32() conversion operator in Color, which was
always returning 0

Change-Id: I0aefaedc34ee0a8bccd43238ccfe8196d2dd5c5c
Reviewed-on: https://gerrit.libreoffice.org/50051Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

...so that the dynamic_cast also works on macOS (where RTTI equivalence is
determined by address, not by strcmp).  The two relevant dynamic_casts are in
BezierObjectBar::GetAttrState and BezierObjectBar::Execute (both in
sd/source/ui/view/drbezob.cxx).  (They can be triggered from the Impress UI by
adding any "Motion Paths" animation to a shape, then selecting and manipulating
the path's control points.)

The source of those dynamic_casts appears to always be sd::MotionPathTag,
defined in sd/source/ui/animations/motionpathtag.hxx in the same library, so
wouldn't technically require IPolyPolygonEditorController to be
SAL_DLLPUBLIC_RTTI.  However, while 367e8743
"INTEGRATION: CWS pathfinder01" introduced that sd::MotionPathTag deriving from
IPolyPolygonEditorController, 346e0ffe
"INTEGRATION: CWS pathfinder01: #i41800# added inteface
IPolyPolygonEditorControler to allow edit of non marked path objects" also added
IPolyPolygonEditorController to SdrPolyEditView in include/svx/svdpoev.hxx, for
unclear reasons.  So better be conservative and assume SAL_DLLPUBLIC_RTTI is
needed after all.

Change-Id: Ieefabf7b00357876046014bbff098076108f9dea
Reviewed-on: https://gerrit.libreoffice.org/50109Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

To find out why this test fails for some linux tinderbox.

Change-Id: I49d1afbbac631abe49378b12b888615e3e0e1bef
Reviewed-on: https://gerrit.libreoffice.org/50121Reviewed-by: Tamás Zolnai <tamas.zolnai@collabora.com>
Tested-by: Tamás Zolnai <tamas.zolnai@collabora.com>

Change-Id: I9deadfdd324f9b9014fef8370593e107302ed843

When saving we need to determine which mime type to use for the
graphic. Return this information back to the caller so the
metadata can be written into the file if necessary.

Also add function to determine the mime type from the XGraphic
by converting it to a stream first (only then we can determine the
format that will actually be used and written to the storage file.

Change-Id: I796da715e47f29d77dea605a2769a2217ecd6ba5
Reviewed-on: https://gerrit.libreoffice.org/49992Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>

Change-Id: Id5481a975dce99a51cc5619e200e5ea46ad3ad1b
Reviewed-on: https://gerrit.libreoffice.org/50106Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>

...$PRODUCTNAME happens to be -n (which is notoriously problematic with various
versions of echo).

Change-Id: I09dfb562689138aa0135982725bf2c254ce06e59
Reviewed-on: https://gerrit.libreoffice.org/50104Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

This reverts commit f9c2bcc8.  Drawing all sorts
of curves and polygons in Draw causes various crashes in
svx/source/svdraw/svdopath.cxx, apparently because in the past ImpXPolygon's
pPointAry had a "buffer" of more elements (nSize) than it had nPoints, and the
code heavily relied on that, assigning to such excess elements of pPointAry past
nPoints (and presumably calling XPolygon::SetPointCount later?).

Conflicts:
	svx/source/xoutdev/_xpoly.cxx
(Reverted the mostly, if not completely, automatic changes there done by
924c0e34 "loplugin:changetoolsgen in svx"; they
will need to be re-done.)

Change-Id: I6cb41fd218c3bdd0b3bbeb45f5b3d68649901a48
Reviewed-on: https://gerrit.libreoffice.org/50057Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

Thanks to Antti Levomäki and Christian Jalio from Forcepoint.

Change-Id: I5681f9123a21fee100a2eda1e8284e559ede51b4
Reviewed-on: https://gerrit.libreoffice.org/50064Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: Id3ffa2333946a9551da9d0f9ecbd2885716b9376
Reviewed-on: https://gerrit.libreoffice.org/50102Reviewed-by: pranavk <pranavk@collabora.co.uk>
Tested-by: pranavk <pranavk@collabora.co.uk>

Change-Id: I079850116648dc69b8c92fa23b30233f63d9b6e7

Specifically for each operator==() call two temporaries were
copied, for operator!=() even twice as much.. in
for(...:GetColumnsRange())

Change-Id: I4b426ae855454544e50efca35aa73303138f7ba7

This reverts commit fb4d5885:

"For one, as Michael already responded, many builds (local, Gerrit/Jenkins,
other tinderboxes) now fail UITest_writerperfect_epubexport with what looks like
random 'raise DialogNotClosedException()'.  Gut feeling is that it's way more
than used to fail UITest_writerperfect_epubexport with a deadlock."
(<https://lists.freedesktop.org/archives/libreoffice/2018-February/079494.html>
"Re: UITest_writerperfect_epubexport hang (only sometimes)")  For example,
<https://gerrit.libreoffice.org/#/c/50070/> "tdf#115853: save current custom
property when adding line" Jenkins linux_clang_dbgutil_64 now failed twice in a
row with what looks like random "raise DialogNotClosedException()" failures in
UITest_writerperfect_epubexport.

And for another:  "As long as its unclear what the underlying problem is, what
benefit does it have to add a timeout failure path in the test?  Without the
added timeout, a local failed test will be ready to be inspected (and moggi now
gave suggestions what exactly to inspect in another mail in this thread).  And
failed tests on the tinderboxes will eventually time out either way."
(<https://lists.freedesktop.org/archives/libreoffice/2018-February/079491.html>
"Re: UITest_writerperfect_epubexport hang (only sometimes)")

Change-Id: I36db7526b68a1cf6be596aa82c57095e4067c150
Reviewed-on: https://gerrit.libreoffice.org/50080Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

Change-Id: I93e69e589cd81fbc9606b1b08233da7fb8f5d043
Reviewed-on: https://gerrit.libreoffice.org/50089Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Tor Lillqvist <tml@collabora.com>

Oh dear how much do I dislike fixed size dialogs. Everytime they
are enlarged to fit new content (esp. gtk3 is rather space-greedy
there) someone starts to whine the dialog is now too large

Doing what e6d7d737 does is not really an option, since there
are some 30+ tabpages to try and test for optimal size here.

Change-Id: I3ffb2d794f76781eb62186093b928289593585cf
Reviewed-on: https://gerrit.libreoffice.org/50032Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
Tested-by: Katarina Behrens <Katarina.Behrens@cib.de>

Change-Id: If07efe4c15cfc28df38a9327856d39313ca78d50
Reviewed-on: https://gerrit.libreoffice.org/50078Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

Project: help  e256ca7d98c9e4da4e14bd405e648edf394a56c4

tdf#115905 Impress custom animation fix

Custom animations are now in menu Format

Change-Id: Iade534e2278de56ff3b7b5647d18196ead6973b8
Reviewed-on: https://gerrit.libreoffice.org/50098Reviewed-by: Olivier Hallot <olivier.hallot@libreoffice.org>
Tested-by: Olivier Hallot <olivier.hallot@libreoffice.org>

Thanks to Antti Levomäki and Christian Jalio from Forcepoint.

Change-Id: I569ca80267ad9b5a21da0029ba903d2a4c45a035
Reviewed-on: https://gerrit.libreoffice.org/50065Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: If1e6727e4b5bb225495e20d5dfb78fa5da770f75
Reviewed-on: https://gerrit.libreoffice.org/50060Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Use the same solution as for images in this commit:
9fee132c

Place the inserted text shape to the center of the slide
for LOK, because we don't get correct window size here.

Change-Id: Ie0c3717d22da6b6a988ab8d566bd9759e03f86c8
Reviewed-on: https://gerrit.libreoffice.org/50074Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Tamás Zolnai <tamas.zolnai@collabora.com>

ROLE_TEXT is for editable text, ROLE_LABEL should be used for static text,
so that screen readers know they should speak it.

Change-Id: I8f781a42a2838f641ccdfe7310046d74f9faced6
Reviewed-on: https://gerrit.libreoffice.org/50052Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>

This avoids build failures when a directory uses UIConfig but does not add
any ui file.

Change-Id: I63413f87fc5515ffb9e3dd9baf25c6a4ba70c2be
Reviewed-on: https://gerrit.libreoffice.org/50071Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>

Change-Id: I9e699eea04231eeeff4146657124ee4f0e68412d

Thanks to Antti Levomäki and Christian Jalio from Forcepoint.

Change-Id: Ic5e9a9d0f8f8d35cb6b39e68338ba029948a4ce0
Reviewed-on: https://gerrit.libreoffice.org/50053Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Thanks to Antti Levomäki and Christian Jalio from Forcepoint.

Change-Id: Ie2b644a3c4c1c165334768eea73d451f07f97def
Reviewed-on: https://gerrit.libreoffice.org/50054Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

it has exactly the same effect as just letting the constant be converted
to Color via the Color(ColorData) constructor.

This commit is part of the process of removing ColorData.

Change-Id: Ie12ad451ba3f43a627ebdbb132a5369f5760ba52
Reviewed-on: https://gerrit.libreoffice.org/50038Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

Change-Id: I7c07f3a6269cd4b92cc50f7aaf847faf04665029
Reviewed-on: https://gerrit.libreoffice.org/50068Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Andras Timar <andras.timar@collabora.com>

Previously textual value like "10,5 pt" was converted to int as simply
10 (multiplied by 10, it became 100), which compared as different from
unchanged value of 105. This made the fractional values to be treated
as always changed.

This patch uses the same code to convert saved value as is used for
current edit box value.

Change-Id: I09a84a6bf33b17e0192b79b31af21ef14d7e9c63
Reviewed-on: https://gerrit.libreoffice.org/50066Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>

Change-Id: Ifb434589ef08428ce609bc7a40b015d4df13224c
Reviewed-on: https://gerrit.libreoffice.org/50048Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

Change-Id: I6b5b6b15f86d452b73fe02df5dec5490b29daed8
Reviewed-on: https://gerrit.libreoffice.org/50050Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>