Pass --enable-tiled-annotations to gtktiledviewer on command line
to turn comments on again, by default they are turned off now.
Change-Id: I40e93ab683ec9538e088ae9aa83cda2eaaf9939e
Reviewed-on: https://gerrit.libreoffice.org/33622
Reviewed-by: pranavk <pranavk@collabora.co.uk>
Tested-by: pranavk <pranavk@collabora.co.uk>
string::compare() will be faster when string::find's result is
compared with 0, because it will not scan the whole string.
Change-Id: I78596a6d796fe9779f88b7c7b91da09aa27b7035
Reviewed-on: https://gerrit.libreoffice.org/32430
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
this reverts
commit 610f59651563205564a0f4818b5c1d347ca02132
fix "cppcheck:stlIfStrFind"
and
commit 5c1bb338ef0d781648d5b33ffc9d64c2be6d9926
cppcheck:stlIfStrFind
which are deficient in one way or another, the original logic, while
inefficient, is still most accurate
Change-Id: I4907d2569c9c2976533f8ac9f4b8b45eb1807b72
What's new:
1) RectangleAndPart handles "EMPTY" payloads
2) LOK_CALLBACK_INVALIDATE_TILES msg type with "EMPTY" payload are
handled in CallbackFlushHandler::queue
3) gtktiledviewer handles "EMPTY" LOK_CALLBACK_INVALIDATE_TILES msg
even if the part number is included in the payload
Change-Id: I21f4a71ec875d24f4bbd100e4aacf8437d745ae4
And since this is now too much code for inline move implementation to math.cxx
Which again made it necessary to give libreofficekit lokdocview.cxx its own
implementation that doesn't even claim to build against sal ...
Change-Id: I0f80be9d9172ee20693b9babde715206f2c3d8c1
Reviewed-on: https://gerrit.libreoffice.org/29428
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>
Since desktop/ code queues, compresses and only emits callbacks on idle,
it's possible that two invalidations are in the queue, and there was a
setPart() call between them. In this case it's impossible to tell what
part the invalidation was sent for.
Fix this by conditionally including the part number in the invalidation
payload. It's off by default, a new feature flag is added to request
this behavior.
gtktiledviewer enables this feature flag by default, though just to show
the part number in the debug output. Android doesn't enable it.
Change-Id: I73e6def848c0eb61d64e71026002c7a0e750aab4
These colors are used in the tiles, so it's a good idea if the client
can use matching colors for cursors and selections. But to be able to do
that, we need an API to expose these colors.
Change-Id: Ia688c07e6c300fecdf8dc428d5a3f000d1857387
If a too large rectangle is parsed, the width or the height may be
larger than std::numeric_limits<int>::max(), in that case just set
width/height to that max value, instead of allowing an overflow.
Change-Id: Ic01319b01a3f9286501c346ea765868be57466a1
As this only returns the number of views, not the actual views. Since
it's a rename, it's just an API (but not an ABI) change.
Change-Id: Ib4f0ea56a90e5ae9c80ee1781aa2f29aff4259e7
Calc omits the document size in the payload in several cases, and
online.git handles that, so handle it in lokdocview as well for now,
instead of fixing up all the sc code to always emit the doc size in the
payload.
Change-Id: Ib2cca1770d2a160e32540e3a3158eb00bf13207b
An alternative would be to follow the Manage Changes dialog approach and
subscribe to the SFX_HINT_DOCCHANGED notification in SwDocShell, cache
the old redline table and find out the differences to the current one,
but that way sound much more complex without benefits.
Change-Id: I20a45285b88255ccea9d6646c0b5288ac1c91879
Open two views, and type into both of them when a Writer doc with
redlining enabled is open: the manage changes dialog now shows how the
correct author is used when creating the redline items.
Change-Id: I48fb90301bfcc04b06d5be5544324ca76fe7b3d7
This shows the full undo and redo stack, with all the metadata available
via the LOK API.
Also fix SfxUndoManager::GetRedoActionsInfo(), so it's easy to show the
undo/redo stack in linear time; and fix a use-after-free in lokdocview.
Change-Id: I66625ed453efa61b5738d99d7d1ad8f468908240
Reviewed-on: https://gerrit.libreoffice.org/27913
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
Hopefully less confusing, the crossed out rectangle is also used inside
the tiles for deleted comments, and the two are independent.
Change-Id: Id06fbf6ec1b21dfbab1c126c3c432f91cf51430c
Reviewed-on: https://gerrit.libreoffice.org/27503
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
When we're after SdrBeginTextEdit(), but before SdrEndTextEdit(), and
have multiple views, then only the active view paints the edited text,
the other views look like the shape has no text at all.
Add a new callback that exposes the position and size of the rectangle
where the shape text will be painted after text edit ended, so clients
can draw some kind of locking indicator there. This way the rendered
result can differ in the "shape has no text" and the "shape text is
edited in an other view" cases.
Change-Id: I6096479a8a05c2547d15222e6d997b848af02945
Reviewed-on: https://gerrit.libreoffice.org/27441
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
Writer pages are exposed as parts, but it still makes sense showing
selections/cursors from other parts in that case.
Change-Id: Ic76d93291bde2d959c149cf2ef5eba7ed33a45e8
Reviewed-on: https://gerrit.libreoffice.org/27321
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
This way we show view selections/cursors from other views only in case
the part number matches. Anything else looks buggy in Calc/Impress.
Change-Id: If3ecbfc4b05350606730319b1a570afc15fd9d0a
Reviewed-on: https://gerrit.libreoffice.org/27315
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
Callbacks are processed on idle on the main thread, so by the time we
parse them, possibly the widget is already gone, avoid that problem.
Change-Id: Ie8e16423d1ffe087e0dd21425026f7a5d644c27b
So that e.g. it's possible to see which invalidation affects which view.
Change-Id: I6b6db2fa07eaecd1315ce8160c3b3b86e9e5a348
Reviewed-on: https://gerrit.libreoffice.org/27138
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
With this, in case a text cursor is turned into a graphic selection in
view#0, then view#1 can also hide the text cursor of view#0.
Change-Id: I7de89b8537ef8b0985336793b719d93733604bff
Reviewed-on: https://gerrit.libreoffice.org/27044
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
So a view can be aware where the graphic selections of other views are.
Change-Id: I0cc420cfe4bf3824fbfa1a58da889cac5e9a7b60
Reviewed-on: https://gerrit.libreoffice.org/26863
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
If we draw a black graphic handle manually, then it's possible to color
it later, this isn't easy if a bitmap is painted.
Change-Id: Ib4456fd5155862d52e3ffa79ee49c7bfd16fb742
Reviewed-on: https://gerrit.libreoffice.org/26860
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
So a view can be aware where the cell cursors of other views are.
Change-Id: Ifcf06c0019c6af8b859e2e92222e4f3fd18da74f
Reviewed-on: https://gerrit.libreoffice.org/26844
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
This fixes the following use-case:
1) Start gtktiledviewer, click New View
2) Click Edit in the first view
3) Click somewhere in the document in the first view -> nothing happens
Change-Id: I79d63538607f03b78851a639adf158d918745276
Reviewed-on: https://gerrit.libreoffice.org/26789
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
Commit 45c2410041c48c22bd860efb42d4daadad7869b0 (LOK: change type of
view ids to uintptr_t, 2016-06-17) fixed the problem of view IDs being
reused for the price of random IDs, which makes debugging harder.
Implement a simple shellToView() function that makes sure view IDs are
not reused, and stop exposing view shell pointer addresses, which allows
reverting the LOK API change.
Change-Id: I63089e6de08ee7e1c7706757d43a11f6cf4d6e06
Reviewed-on: https://gerrit.libreoffice.org/26773
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
Otherwise it's possible that a keystroke is sent in for a different view,
when that other view reacts to an invalidation (invoking paintTile())
caused by a previous keystroke.
I.e. open two views, place the cursor at different positions, type fast,
and some of the characters appeared at the incorrect view.
Change-Id: Ie5e471f1b9c2d69adaa87111fba74d4abe184ef8
Reviewed-on: https://gerrit.libreoffice.org/26562
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
It's similar to the normal selection, but it's colored and has no
handles.
Change-Id: Ibd9594b4834ff4f9b1cfd85912ed5cee3c8b8c71
Reviewed-on: https://gerrit.libreoffice.org/26543
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
So a view can be aware where selections of other views are.
Change-Id: I5026b1ff2b99a4eedfd0bde32a05ceb8e2f424bc
Reviewed-on: https://gerrit.libreoffice.org/26542
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
It's similar to the normal cursor, but it's colored and does not blink.
Change-Id: I6a869a98f46979946f320905426e016fe011cbc6
Reviewed-on: https://gerrit.libreoffice.org/26522
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
So a view can be aware where cursors of other views are.
Change-Id: I6133fb55aa2869843c0284b7d76264bab3b3d5da
Reviewed-on: https://gerrit.libreoffice.org/26513
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
This fixes the following problem:
- createView() = 1
- createView() = 2
- destroyView(1)
and then view #2 was renumbered to 1.
Instead expose the pointer address of the SfxViewShell as the ID, which
is not changing in such a situation.
Note that the SfxViewShell <-> ID mapping is an implementation detail of
SfxLokHelper, and only pointers are converted to IDs, user-supplied IDs
are never converted back to pointers.
Change-Id: If79ef8b99ba391011b5d82b219ad13447d44cd5a
Reviewed-on: https://gerrit.libreoffice.org/26423
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
