Commit graph

2687 commits

Author SHA1 Message Date
Ashod Nakashian
4cba61525c loolwsd: test notification replay on re-activation
Change-Id: I46471a2a3ecce219cf1b922aeebf0fdb02f66614
Reviewed-on: https://gerrit.libreoffice.org/24783
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-09 05:26:15 +00:00
Ashod Nakashian
b18246569d loolwsd: verify the number lokit instance before and after each test
Change-Id: Ia38475dc78ef5b4dc116a2d6f9791c7012985b5d
Reviewed-on: https://gerrit.libreoffice.org/24782
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-09 05:24:25 +00:00
Ashod Nakashian
b0843ba443 loolwsd: test helpers accept smart pointer WebSocket too
Change-Id: I83e62c7ce0daa6e44fb7f163607436640c9b9f08
Reviewed-on: https://gerrit.libreoffice.org/24781
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-09 05:22:43 +00:00
Ashod Nakashian
f20a9e94b6 loolwsd: templatize parsing helpers
Change-Id: Ia52dbf59bf5f2a816578b75111ee9accdeaaf816
Reviewed-on: https://gerrit.libreoffice.org/24780
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-09 05:21:37 +00:00
Ashod Nakashian
50cc60e6f6 loleaflet: invalidate on activation
Invalidation for spreadsheet and presentation
are part-sensitive. The handler will take
care of this distinction and construct the correct
invalidate command.

Also, set active flag before firing invalidation.

Change-Id: I9e897750e0768b54cc1521d47d69b5dbb1a36244
Reviewed-on: https://gerrit.libreoffice.org/24779
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-09 05:21:05 +00:00
Tor Lillqvist
ab92acac20 Spelling correction 2016-05-09 07:30:26 +03:00
Ashod Nakashian
f9d7de86db loleaflet: don't show the start/end selection handles for viewers
The handles imply (and indeed are designed for) changing the
selection by dragging them around. Viewers can't change the
selection, so no reason to give them the selection handles either.

Change-Id: Ifa793ac982fd48be877e7ed7709f70890d8de194
Reviewed-on: https://gerrit.libreoffice.org/24770
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 16:14:15 +00:00
Ashod Nakashian
d8f6aa7146 loleaflet: remove unnecessary duplicate message
This message is already sent from CalcTilelayer.js.

Change-Id: I9f84d5f365dbfea38061e206d5d52e2201be22ed
Reviewed-on: https://gerrit.libreoffice.org/24769
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 16:13:11 +00:00
Ashod Nakashian
3cefa3c6e4 loleaflet: announce the current part after taking editing
Make sure to sync both lokit and viewers as to which
part is being edited when a previously-viewer takes
editing.

Change-Id: Ia70f647a26029447bb4360633f79233fa9596059
Reviewed-on: https://gerrit.libreoffice.org/24768
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 15:50:27 +00:00
Ashod Nakashian
702dd48f1f loolwsd: safer document saving
Relying on the filesystem to tell us when the document
was last modified (to decide whether to upload to storage or not,)
proved unreliable.

Now we always upload to storage if there is only one client.
This both minimizes the risk and also avoids the file timestamp
check as a workaround to the problem of re-uploading documents
as many time as there were clients. Since with one client we
can only upload no more than once per save, which is reasonable.

Furthermore, when a client disconnects we auto-save automatically
as a matter of precaution. However, when there are other clients
still connected, we don't wait for the save to complete, rather
we let that job to the very last one.

Change-Id: I94a2e4bddaed30a6c9c0e69f8006667d33c5b8ee
Reviewed-on: https://gerrit.libreoffice.org/24767
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 15:49:41 +00:00
Ashod Nakashian
ce36141662 loolwsd: verify type before parsing part names/codes
Change-Id: Ib841167264d756850cb62e20f8a16a3ba147de37
Reviewed-on: https://gerrit.libreoffice.org/24763
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 13:08:53 +00:00
Ashod Nakashian
a820ec382b loolwsd: remove extra space between 'status: ' and the message
Change-Id: I122931f3aba2cf02f382f41531e01cb9f3eb7968
Reviewed-on: https://gerrit.libreoffice.org/24762
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 13:08:35 +00:00
Ashod Nakashian
dd164f905f loolwsd: no need to send clients current part on invalidation
Invalidation messages already contain the part they
are referring to (with the exception of EMPTY).

Sending clients the current part makes it hard
to let clients view different parts of a doc
while an editor types.

It also implies that the part has changed,
which isn't necessarily true. If it is, a
separate curpart notification will be sent.

Change-Id: I49afcebef8994c892ef8be633d3ce3982b031d2b
Reviewed-on: https://gerrit.libreoffice.org/24750
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:39:07 +00:00
Ashod Nakashian
e7d2ed499f loolwsd: more notifications to replay on re-activating clients
Change-Id: Ieb91eee0cb4ff66cd6bea129039f8fe328939888
Reviewed-on: https://gerrit.libreoffice.org/24749
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:37:14 +00:00
Ashod Nakashian
2829febd15 loolwsd: logging improvements
Change-Id: I5591e309ac9fbcbd79c816b26e65f51b77018bb2
Reviewed-on: https://gerrit.libreoffice.org/24748
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:36:00 +00:00
Ashod Nakashian
b3acb4001d loolwsd: add tests to validate Impress page changes
Validate the pages before/after insert/delete and
undo/redo operations.

Change-Id: I16aa14f022a8bb0a4b126ea9bc222b788dc7d2dc
Reviewed-on: https://gerrit.libreoffice.org/24747
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:33:27 +00:00
Ashod Nakashian
710134e1be loolwsd: cleanup Impress InsertDelete test
Change-Id: I9dfc5ee36797d0b6a8bff7b5c7e54007c6115bb0
Reviewed-on: https://gerrit.libreoffice.org/24746
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:32:34 +00:00
Ashod Nakashian
d2f4a516a6 loolwsd: spawn ForKit process after initializing server
WSD server must be up and listening to incoming
connections before firing up ForKit and spawning
children.

Having the order reversed caused tests to fail
randomly when WSD was slow to listen to incoming
connections from children already initialized.

Change-Id: I4eaf4a658c65da024101efc096c39222ebfa3c00
Reviewed-on: https://gerrit.libreoffice.org/24745
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:31:44 +00:00
Ashod Nakashian
1a896de0a7 loolwsd: getStatus and status formatting cleanup
Change-Id: Ifd842d7ded1be1f4c89d8371790f19a3909c9729
Reviewed-on: https://gerrit.libreoffice.org/24744
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:28:18 +00:00
Ashod Nakashian
f02deef936 loolwsd: proper logging in Prefork Unit and other cleanups
Change-Id: I23017d3d34edb3608cfe8bc0b5c5eaa340231033
Reviewed-on: https://gerrit.libreoffice.org/24743
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:27:09 +00:00
Ashod Nakashian
641038ef47 loolwsd: noisy failure when test can't collect stats
Change-Id: I6a1693954d84d175478b08f3a49e07e9447e1517
Reviewed-on: https://gerrit.libreoffice.org/24742
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:26:36 +00:00
Ashod Nakashian
5c47a42fea loolwsd: handle socket errors in Prefork Unit
Change-Id: Idc5a2ea776e0b8f0cde43526d690468e1969e611
Reviewed-on: https://gerrit.libreoffice.org/24741
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:26:00 +00:00
Ashod Nakashian
21bf2ff129 loolwsd: give children time to settle after forking before checking
Change-Id: If29d564f16a29f94791c28f5cb03be2ebf7219f3
Reviewed-on: https://gerrit.libreoffice.org/24740
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:24:51 +00:00
Ashod Nakashian
6f39c0f282 loolwsd: state the child PID which has no DocumentBroker
Change-Id: Ifa679ac67cc0a1d12bba4a44f24831e7a291d97d
Reviewed-on: https://gerrit.libreoffice.org/24739
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:21:41 +00:00
Ashod Nakashian
4369cb1ff3 loolwsd: UnitPrefork times out while waiting on child to respond
This prevents deadlocking in case of test failure.

Change-Id: I3026311e9a67543a26acb0316546e6b5aacf61dc
Reviewed-on: https://gerrit.libreoffice.org/24738
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:18:36 +00:00
Ashod Nakashian
9e7b086587 loolwsd: refactored new child addition
Change-Id: I5f554dd0c2d37be3f90059bc8a943cbbe7ac094b
Reviewed-on: https://gerrit.libreoffice.org/24737
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-08 04:17:06 +00:00
Ashod Nakashian
8e480d7ac1 loolwsd: use paintPartTile and remove workarounds
Change-Id: Ice59004a23919c7a77a00782f7210a83bc1464e7
Reviewed-on: https://gerrit.libreoffice.org/24703
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-06 15:06:37 +00:00
Pranav Kant
27497ba4dc loolwsd: Fix tile previews for viewing sessions
Also, don't let these tile previews change part of the document
permanently. This is a temporary solution till we have some
better API from LOKit to deal with such a situation.

Change-Id: I8dfefd2b7ad8cf3e7a57afb95b57994ef0bb3b6c
2016-05-06 16:20:27 +05:30
Pranav Kant
1f990d3fc2 loolwsd: Fix incorrect position of id=, editlock= params
editlock= is always the last parameter. This fixes the slide
previews in impress documents for editing session.

Change-Id: I3531c7f52e09e655524fa0afd6fe66da504b5d70
2016-05-06 14:44:03 +05:30
Pranav Kant
1864491e23 loleaflet: Enable/Disable buttons for edit/view modes, ccu#1768
Change-Id: Id14ef66adb9bda33246625d89c093cb88927f2bb
2016-05-06 13:59:21 +05:30
Pranav Kant
2d69745d2c loleaflet: Use tabs consistently
Change-Id: I1d24ba1c8956a919eb02347dfc379927f536173e
2016-05-06 13:06:26 +05:30
Henry Castro
e03d4f7b77 loolwsd: fix -Werror=shadow 2016-05-05 14:53:34 -04:00
Ashod Nakashian
1643938a8e loolwsd: support the tile id param when broadcasting tiles to subscribers
Obviously this is dangerous, since the id is not part of the
subscription key (the filename) so different clients could
have different ids on the same part, but in practice I
expect this not to happen. Though that clearly depends on
how clients use the id.

Change-Id: I52a0b043c9b5e5ad1111b692e4216cc9ffec5b2b
Reviewed-on: https://gerrit.libreoffice.org/24680
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-05 14:23:22 +00:00
Ashod Nakashian
346a063d0a loolwsd: fix regression to handling tile command
Change-Id: Idedf25f3756ea1e04daf17540b1a57e1e7af8448
Reviewed-on: https://gerrit.libreoffice.org/24679
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-05 14:22:53 +00:00
Ashod Nakashian
376b36b4c8 loleaflet: writer shouldn't signal page change when nothing changed
Change-Id: I1393f0c385f75c07d2024c7045d04eabc6937718
Reviewed-on: https://gerrit.libreoffice.org/24677
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-05 11:39:37 +00:00
Ashod Nakashian
83307291ec Document id argument of tile and tilecombine
Change-Id: Ia43ceae31936f68d2bed04b5291a25eeffc637b3
Reviewed-on: https://gerrit.libreoffice.org/24676
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-05 11:37:24 +00:00
Pranav Kant
3d02fb7497 loleaflet: Enable/disable spreadsheet tabs for edit/view mode
Change-Id: I084f676e77fc161c109a5271dd1cad60a3e5474f
2016-05-05 14:30:00 +05:30
Pranav Kant
7734f2557a loolwsd: Do not allow viewing sessions to change parts
bccu#1433, bccu#1757 related.

Piggyback editlock information to tile messages so that kit can
use that information to allow changing parts only for messages
with editlock.

... and decline tile render request for tile messages without editlock
information

Change-Id: I9cedb870cd977741375665cb258d04c818481a14
2016-05-05 14:28:03 +05:30
Pranav Kant
207f4b5fc9 loleaflet: Emit 'editlock' only if changed
Change-Id: I0d08241b82ad105902ad6ca03f015efd0eb9fb2e
2016-05-05 12:33:57 +05:30
Ashod Nakashian
909c996a6d bccu#1774: poor API never fails to frustrate
The tile and tilecombine messages apparently have optional
appendages at their rear ends. Not one, but two (at least).

While the first (timestamp) seems to be truely optional
(in the sense that leaving it out doesn't break anything,)
the same can't be said of the second (id).

For Impress slides this id is used to identify the slide
to which the tile belongs. Or rather the slide being
rendered, as it seems meaningful only for the slide
thumbnails.

Previously the complete arguments of tile were copied
verbatim from the input to the output (i.e. back to the
client) and so any extra payload was also echoed back.

But when id is missing (when expected) loleaflet not
only fails to show these tiles (understandably,) but
it also fails to show the scrollbar for said slide
thumbnails altogether!

With the new logic to move the tile communication to
the child socket instead of the clients, the arguments
are parsed and then serialized back in the response.
So all fields must be explicitly known in advance.

This change is necessary because tilecombine is broken
to tile commands and so both share common code. This
means that echoing back the request verbatim will
break loleaflet since tilecombine arguments (which
is a list) is not a valid response (which has the
format of tile). So the internal representation
has to be something neutral/common.

The fix is to parse the timestamp and id only when
provided and to echo back the id only in that case.

Change-Id: Ic97cf0de4083d412a4b3543e8b9a8713ac27a27c
Reviewed-on: https://gerrit.libreoffice.org/24669
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-05 02:41:59 +00:00
Ashod Nakashian
cd8d48e6bb loolwsd: tile logs
Change-Id: Id5c8b688c36038dc6f02ad95ad1c43412dcd2c31
Reviewed-on: https://gerrit.libreoffice.org/24668
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-05 02:34:13 +00:00
Ashod Nakashian
fedb3f520a loolwsd: factored out common tile message
Change-Id: I39600821e3c9c40f3a63343b167cddb3e42b7072
Reviewed-on: https://gerrit.libreoffice.org/24667
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-05 02:33:52 +00:00
Henry Castro
4d52c200c5 loolwsd: test: add testSlideShow 2016-05-04 16:19:47 -04:00
Henry Castro
ae80052e80 loolwsd: add mime type svg 2016-05-04 16:17:29 -04:00
Andras Timar
627cbb73de admincreds' argument is not directory 2016-05-04 20:33:13 +02:00
Ashod Nakashian
25c550ff8e loolwsd: test that slow/hung clients don't block others
Change-Id: I16697ab3afd1d209a77d48506604ec7babc1bfff
Reviewed-on: https://gerrit.libreoffice.org/24655
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Ashod Nakashian <ashnakash@gmail.com>
2016-05-04 12:38:09 +00:00
Jan Holesovsky
68ed15628d Another place that hardcodes the URI. 2016-05-04 13:56:39 +02:00
Jan Holesovsky
f40167cd6e Initialize the _url correctly. 2016-05-04 13:39:32 +02:00
Jan Holesovsky
191ee00f2c Run the tests on ports that do not conflict with a 'normal' loolwsd.
For the moment, it will allow running 'make check' that does not conflict with
an already running loolwsd (eg. from 'make run').  Later we can consider
running more tests in parallel.
2016-05-04 13:39:32 +02:00
Jan Holesovsky
f6f4681556 Use a separate client port for running the tests.
So that they can run in parallel to a 'production' loolwsd (like one from
'make run' or so) in the future; for the moment it is not possible, as the
MASTER_PORT is hardcoded.
2016-05-04 13:39:32 +02:00