On Apple devices shortcuts like Copy, Paste, etc use the Cmd key
(also known as the ⌘ key), not the Ctrl key. We shouldn't check
KeyboardEvent.ctrlKey but e.metaKey. Introduce a function _isCtrlKey
that checks the proper property depending on platform.
That Cmd-V still doesn't work for plan text put on the pasteboard by
another app needs to be fixed in core.
Change-Id: I4e8a05afaed0bde193d00a7ded52fcebf3536b5f
Signed-off-by: Tor Lillqvist <tml@collabora.com>
Scroll bar animation was flickering after mouse leaves its area.
Signed-off-by: Gökay Şatır <gokay.satir@collabora.com>
Change-Id: Ie4d96d98dee806e6b02c8c8a23dd4eaccabbf7b0
Fix description:
The view jump is because after the zoom animation has ended, the map
center is set to the pinch center and not the last frame's center. The
patch computes the last frame's center and uses that to set the map
view's center. This only fixes the view jump and not the flicker after
the animation which is an independent issue.
Caveat:
Since we show only one discontinuous frame at the "final allowable zoom"
at the end of the zoom animation, there will be perceived view jump at
the end of the animation itself. This apparent if the user pays
attention to the contents of the document at the corners of the view
before and after the is pinch stopped. This jump is more pronounced if
the pinch zoom factor is much different from the "final-zoom".
Note: The final zoom is determined from the final pinch zoom as
var finalZoom = this._map._limitZoom(zoomDelta > 0 ? Math.ceil(this._zoom) : Math.floor(this._zoom));
--
A possible solution for this problem is to render a few frames at the
end of the animation that are closer to the "final zoom" from the where
the pinch left off.
--
Signed-off-by: Dennis Francis <dennis.francis@collabora.com>
Change-Id: I0a6989248970aeb60acc451d9b0bc7a08a0dbf06
(cherry picked from commit 5b2daac61edb78150ad9700ce11fab8bdfb7dd27)
Move the ODF entries for the ODF file types from the
UTExportedTypeDeclarations section to the UTImportedTypeDeclarations
section in the Info.plist. This, for some unclear reason, is necessary
to make ODF documents not appear greyed out for folders provided by
the Google Drive file provider extension. Sadly the documentation of
exact semantics of the Info.plist file leaves much to be desired.
This fixes https://github.com/CollaboraOnline/online/issues/1509
Change-Id: I7ff35cf9f6c717d1579adc7a4d7350f78d7bee8e
Signed-off-by: Tor Lillqvist <tml@collabora.com>
When user moves cursor out of the scroll bar area or at the end of a
scrolling event, scroll bar will slowly disappear.
Signed-off-by: Gökay Şatır <gokay.satir@collabora.com>
Change-Id: Ic1c993869bdee3be342330c847b6d89e66784f69
User can set stopping events, duration or they can directly stop, reset
the animation.
In order to use stoppingFunctionList, canvas element should get events.
Currently, CanvasTileLayer is calling callbacks of
CanvasSectionContainer.
So if one needs to use stoppingFunctionList, they should update _mirrorEventsFromSourceToCanvasSectionContainer
function and use "element.dispatchEvent" feature. We need this because
of leaflet.
I didn't change "_mirrorEventsFromSourceToCanvasSectionContainer"
function because it would have to be tested with different browsers. But
i tested the dispatchEvent feature in order to test
stoppingFunctionList.
Signed-off-by: Gökay Şatır <gokay.satir@collabora.com>
Change-Id: Ifb335d433af4656294223fa6c16378819abbb950
And explain the use of linear search in header().
Change-Id: If6c39bffa738861d0c86684f9560ed6e4a6970e1
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
User-Agent is designed for client-side use only,
in http requests. For servers, the Server header
is designed to announce the server name and version.
This tries to normalize the use and documents
the proper intent and usage.
Change-Id: I42d68d65611cab64c45adf03fe74f9466798b093
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
In many cases we write directly to the socket
outgoing buffer directly. The only way to flush
this buffer to the socket is to use send(),
which is artificial if we don't have any more
data to pass to send().
This makes it possible to safely call flush()
and, if there is buffered data, we flush it
to the socket.
Change-Id: I025e7ba5d1898e72dd0001547e104642608ab20a
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This adds writeData() member to write into the output
buffer of the socket directly, instead of serializing
into a string and then copying it over to the buffer.
Change-Id: I42aaa23e96a6120c738bc0bb36bc01f4ed5e70c5
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
An anti-feature, Poco's HexBinaryEncoder inserts
gratuitous line breaks (0x0A) every 32 bytes,
which is neither expected nor necessary.
No idea what use line breaks could have in a
hex encoder (unless it assumes it's only
possible use is to dump data for investigation).
We hadn't observed this because we generate
random hex strings of 8 to 16 bytes long.
But having used it for random URLs that
are up to 1024 bytes long, I started getting
invalid URLs.
Another reason to hasten the removal of Poco,
especially when we have our hex converter
anyway (not that it would matter if we didn't).
Change-Id: Ib674e8ed607db1effef476f1f3478da76c4f6464
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Previously, we were parsing the Status Line
and the header, both together. This of course
doesn't work when we had removed the Status
Line already. This wasn't an issue in parsing
responses from servers, but doesn't work when
parsing requests from clients (i.e. in a server).
Also, it's simpler.
Tests extended accordingly.
Change-Id: Id1c9a6385080c86b6072130fa8a169a699c42462
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Here we are more tolerant with regards to the
minimum Status Line length, since a Reason
Phrase is informative and might be omitted.
While technically it should be provided, per
the RFC, we shouldn't break when it's missing.
Change-Id: Ic702db61e5bf4272e18f09d127405033277e5943
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This was not necessary in the end; and a bad idea in general probably,
so given that has never appeared in a release yet, let's revert it.
This reverts commit 5999a5b7cc.
Signed-off-by: Jan Holesovsky <kendy@collabora.com>
Change-Id: I5ee98ea85324f584db7de702a8d98ff9a6780027
We don't need to have both a panend event (as generated by Hammer.js)
and a mouseup event cause the _endTabstopDrag function to be called
when ending a touch drag of a tabstop, especially as for the mouseup
event the result was that the tabstop jumped oddly from where you have
dragged it.
Also other simplifications of events used in the iOS app case. Likely
even further simplifications would be possible.
This does not perhaps fix what
https://github.com/CollaboraOnline/online/issues/732 specifically
mentions, but at least for me it makes moving tabstops in the iOS app
work much more reliably.
Change-Id: I14ff5baeb775f0017368c0c53300df39ab0ab6e6
Signed-off-by: Tor Lillqvist <tml@collabora.com>
it was not possible to paste or copy to/from
the insert hyperlink dialog or search field.
Fixes this problem also in other browsers
Change-Id: Id4cef6681a53e451eeb1f7d9e6d3b9c26b49d66f
Signed-off-by: Szymon Kłos <szymon.klos@collabora.com>
Signed-off-by: Dennis Francis <dennis.francis@collabora.com>
Change-Id: Ic23d772efa1beac31a920257ef08e597fdc668d4
(cherry picked from commit 9ff97f11b1682ff125870f89b6874fefaf1fc6f2)
... and avoid functionality breakage.
This is important as we now need a real 'isVisible' state encoded as
private member 'visible' and this cannot no more be reused as "is cursor
DOM attached?"
So maintain another state variable 'domAttached' and introduce and use
its getter for all previous isVisible() calls.
Signed-off-by: Dennis Francis <dennis.francis@collabora.com>
Change-Id: Ic952611c9bdc138414bd3361e3e166122b2842b8
(cherry picked from commit fd92f4f802744ccb22e3cb98f2e7409bf4970561)
Signed-off-by: Dennis Francis <dennis.francis@collabora.com>
Change-Id: I8b2676b3b576609dce4f6edcb891f5ba85eeac97
(cherry picked from commit e92c7cf925ee4bcbf07dd355899722b51c3f0779)
Fix description:
Update the 'visible' flag if cursor has gone out of view. Also Sync the
visibility of the cursor with its header.
Signed-off-by: Dennis Francis <dennis.francis@collabora.com>
Change-Id: If71736d1b042be12474dc4162a8a2db0d30613b7
(cherry picked from commit ce9f750a05bf3fdcf7771e72400dd57168eb5fb9)
Signed-off-by: Dennis Francis <dennis.francis@collabora.com>
Change-Id: I409c6e63aed864d30a802fb64a352352b172a12a
(cherry picked from commit 17bc92f8210c80660582aba1b014ba3bdb6fc110)
Fix description:
Only try to update the view selection drawing if map still has a record
for that view.
Signed-off-by: Dennis Francis <dennis.francis@collabora.com>
Change-Id: I59c389f9034c4481ea1cc4d07bf46b36e52db40d
(cherry picked from commit 1f482119062a9bfc9fb865c6faa59033126d5cea)
Problem description:
* Open a spreadsheet, then switch to another sheet,
* Open the same spreadsheet with another view (view B),
* Wtih view B, switch to the other sheet where view A is, and click on a
different cell.
* View A's cell selector should be shown on the cell view B clicked away
from in the last step, but it isn't.
Fix description:
The "view" messages (view-cursor and view-selections at least) need to
be replayed after a sheet switch to update(add or remove in this case)
the overlays. Also make sure when view cell cursors are removed from
overlay renderer, they are removed from the doc-layer cache too.
This also restores d11bea88f862eddb1e9ecf65dbaf0fd23fa89881 to let
joining user's cursor be visible on already existing views.
Signed-off-by: Dennis Francis <dennis.francis@collabora.com>
Change-Id: Ib97cf1ef7014d0cecb444823c9ff0abc6b2c8f3a
(cherry picked from commit 1dad869e7748addb82736d553434508a3fcb7f6c)
this will position correctly canvas espacially with groups
in Calc both on desktop and tablets. Simplify the code.
Also fix incorrect size of the canvas on tablets:
if we update 'left' property of the container, recalculate size
Change-Id: I9000839370ff8bc136094517b80bdf350a0db467
Signed-off-by: Szymon Kłos <szymon.klos@collabora.com>
Before it was sent for mobile and tablet only after
entering the edit mode what caused to not showing
the rows/cols in calc on tablet in readonly mode
Change-Id: Iefbdbe66c543faa56228a6ec8eed4b84fe417c4a
Signed-off-by: Szymon Kłos <szymon.klos@collabora.com>
position: fixed in IE11 doesn't work as in other browsers
so calculate absolute position instead of relative
Change-Id: I1ed07a26ba924273e527694e7d982f3814752400
Signed-off-by: Szymon Kłos <szymon.klos@collabora.com>
IE11 doesn't support custom css vars so we values
were not read. use standard properties instead.
Change-Id: I3d1ccd74cc54a048e30431d162b5b80a49b5e06f
Signed-off-by: Szymon Kłos <szymon.klos@collabora.com>
IE11 doesn't support Array.includes but we can't add
polyfill to the prototype because then we will break
all code which iterates over arrays (it will iterate
also over added method...) so add member function here
Change-Id: I122204b3621a76f5678088e3fda6f44de611f5a4
Signed-off-by: Szymon Kłos <szymon.klos@collabora.com>
