This regressed with 7c1e1120c6
Calling .initialize() on socket object is not enough anymore after above
mentioned commit; it doesn't re-establish the socket connection as
was the intent earlier in following cases.
Change-Id: Icd0a87a875eb275ed3c914474cdc502b460e2d16
Reviewed-on: https://gerrit.libreoffice.org/42431
Reviewed-by: Henry Castro <hcastro@collabora.com>
Tested-by: Henry Castro <hcastro@collabora.com>
total_mem can be confused with total system memory or total memory
availabe to loolwsd process. Change the API to say what it actually
returns.
Change-Id: I55c246b8a1d0c5c52767520150e55616266a6d4f
Though no harm as of now because we are calling resetAppLoaded whenever
we are emitting 'docloaded' with false status at the caller site, better
safe than sorry for future.
Change-Id: Id6d86ca2ae01a6eb32c8120df435d0535d906938
Now the Close_Session Post message command can be used to close the
documents while it is loading.
Change-Id: If60391834fedc90b3dcb63848aa72503eb43888d
The timer is cancelled right after 'statusindicator' is received, so a
document never really go to inactive mode after set inactivity time
period (unless user refocuses the document). Call _activate() to ensure
that timer is started after document is completely loaded.
Also timer shouldn't really start when the document is not yet loaded.
Change-Id: I58f5d7718c65cc37da9c3feb99ee6b16741a22a2
First, use the docloaded event to see whether we should be dimming the
document (make it inactive) or not. There is no point in inactivating a
document before document has completely loaded; so don't start the timer
until after the document is completely loaded.
It turns out that we were only emitting the WOPI's App_LoadingStatus
with DocumentLoaded for the first document load, but not for subsequent
reconnections. The problem here was that doclayerinit event is only
emitted during first document load (when TileLayer is -actually-
initialized), which was responsible for emitting this event. By
bifurcating the document load from document layer initialization, we
solve this problem.
Change-Id: I0b7b97fc6244ba9ebd6318d68d78d3abef2c0c08
Reply was added when clicking Cancel. This is a regression from
0760786392
Additionally, when focus was lost on an empty, initial commit,
an empty, defunct comment box remained until reload.
Change-Id: I49a5224ff70de65623496cfcba538acb6a8086f8
Reviewed-on: https://gerrit.libreoffice.org/40915
Reviewed-by: pranavk <pranavk@collabora.co.uk>
Tested-by: pranavk <pranavk@collabora.co.uk>
This is not required in any other browser, but Edge 40 requires it
otherwise the popup window upon clicking the combo box goes behind the
main #map element.
Change-Id: Ie95263f08c8af27b16a56ba2c3c7a9faa308400a
An `npm update` on my local tries to update all the modules to latest
version; surely we don't want it everytime only one module update is
required.
Change-Id: I2f79e77ae23edb9b8f8248ac7535e014b7b3218d
... because the Makefile did not work on SLED11SP4 (?!)
Change-Id: I4e7468149f1c5461aeeb05d3a68e62cda1e329e5
Reviewed-on: https://gerrit.libreoffice.org/39778
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Tested-by: Michael Meeks <michael.meeks@collabora.com>
Start killing documents when memory usage goes above threshold.
Also make it possible to close documents from admin instance.
In DocumentBroker::closeDocument, just set the _stop flag and wake
up the polling thread which will terminate the children, instead of
manually terminating the children.
Change-Id: Ie70e05b3fb6ea816a87b6dcfaed92cdddb94aa90
This is so because we require this also when document is opened in
readonly mode and we don't have any toolbar in the readonly mode.
Change-Id: I085d1640f52843e01370d528d6eea656b743d908
