And move the validation of rename sessionId
where it belongs.
Change-Id: I3fb16ff7b648accaa20b041bf3fd6b82f516cb9b
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This fixes a regression where the
per_document.idlesave_duration_secs setting
wasn't effective.
Idle save also takes into account the time
uploading finished, or if we had a recent
save issued. This is particularly helpful
for large documents and/or slow networks
where uploading takes dozens of seconds.
In that case, we don't idle-save too
frequently.
Change-Id: I76ee1869003073b5cf9bd420f610bd8a12ff71b3
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This avoids an unnecessary lookup in
_sessions, as we already have the
ClientSession instance.
Change-Id: I3dfdfd322fb907f916235a5ccc5f3b76e0eb4855
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Each test loads one or more document. At the end
of the test we may load a new test (a new
UnitWSD/Kit instance). However, because of the
static interface that the tests have, there
could be only one instance at a time. This
means that when we switch the current test
instance, an existing DocBroker might fire
events into the wrong (new) test, making it
fail.
By capturing the current test, we make sure
that DocBroker can never run against different
tests.
Change-Id: I161ec3c9ce1617f41d03b53d2d5fb19682930b78
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This helps generate the public URL
easily and unifies the processing
of embedded video.
Change-Id: I07ef7f7722f368957a85e41493fe175de38464f7
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This is needed, for example, for date type cells
for which the format language is set to default.
In this case the load language is used and it
determines the display/output format.
Signed-off-by: Gabriel Masei <gabriel.masei@1and1.ro>
Change-Id: I49ec4940377b261971224d37ea5fbe8ed7006157
There are a number of races with having Storage
track the attributes. To fix them, we move all
attributes to DocBroker and correct a number
of issues.
The idea of the design is based on the fact that
we want to capture the attributes between
uploads, but based on the saved document.
That is, when we upload a document version, we
want to pass to the storage whether from the
perspective of the *Storage* there has been
any user-modifications or not. Since saving
to disk may happen multiple times between
uploads (not least because of failures), and
since saving resets the modified flag, we need
to capture the modified flag at each save and
propagate it until we upload successfully.
Upon uploading successfully, we reset the
attributes.
For this reason we have two attribute instances;
one is the 'current' attributes as being uploaded
and the other the 'next' one. We capture the
current state at saving into 'next' and we merge
with 'current' when saving succeeds and we
aren't already uploading (otherwise, we update
it and then discard it when uploading succeeds,
losing the last attributes).
Furthermore, because the modified flag is
reset after each save, and because we might
save and upload immediately after a
modification, we may not have the modified flag.
This means that we need some heuristics to
decide if there has been user-issued
modifications. (It is better to be conservative
here.) We try to detect this by introspecting
the commands we receive from users.
In effect, we capture the attributes when issuing
an internal save, we transfer the captured
attributes only when saving succeeds and we aren't
uploading, and from then on, uploading has to
succeed to reset the 'current' attributes. In the
meantime, if we fail to upload and issue another
save, the new attributes will be captured and
merged with the 'current' and the next upload
will not have any lost attributes.
Change-Id: I8c5e75d25ac235c6232318343678bf5c0933c31e
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
There is a race between the time of modifying
a document, receiving the modified flag, and
saving. This can happen when, for example,
the user modifies the document and closes
immediately. In that case, when uploading
we will not have the modified flag and will
not set the User-Modified attribute.
While this isn't 100% accurate, and it can
never be, it's still better to be conservative
and flag a version in storage as user-modified
than otherwise.
Change-Id: Ia504a7cddd4839bcbfeaaf9bf6c90ed8b68efa91
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This introduces a new config entry to
control the minimum time between uploads
after a failed attempt. In the case of
a successful upload, the minimum time
between uploads is bounded by the
min_time_between_save_ms config.
Change-Id: Ic7e37f97a52af2c5ec82897b6e81f57211d16553
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This adds a new config entry to help tune
how often to save after the last save
request. This is used regardless of
the auto-save interval.
Change-Id: I58c291e69a63a3b98e72584817b408b54d9f80f1
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
And dump the saving timeout in the state dump.
Change-Id: If203d5434ee17f646009ee241b8484ca597840ad
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Cleanup some debugging, distinguish wsd/viewport delta fetches from
client-side tilecombines.
Update tests for deltas where they are expected.
Change-Id: I3dca09e4ceb094abf03db2aed2618f19351f3e28
Signed-off-by: Michael Meeks <michael.meeks@collabora.com>
The new helper corrects inconsistencies in
detecting writable sessions and improves
readability.
A writable session is one that is either
not read-only or can change comments
(in PDFs, which are otherwise read-only
documents).
Also fixes a bug in reusing the last editor
session to save. The condition should've
been !isReadOnly(), since sessions used
for saving cannot be read-only. Indeed,
the last editor couldn't possibly have
been read-only by definition.
Change-Id: I9860be287057050eca692568d43f77371964f80b
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
When deciding whether the file on disk needs uploading
or not, we compare it with the last modified timestamp.
For this to work they need to be different.
When there is no file to upload (i.e. when there is
nothing new to save), the timestamp we get is
unix-time of 0. This is always different from 'now'.
So we incorrectly think there is data to be uploaded
when there isn't.
Better to initialize the last modified timestamp to
unix-time of 0, so in this case we correctly
conclude that there is nothing to upload.
Change-Id: I545e9021e9c76d009e08758342effaf3e9b964c7
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
When disconnecting from Kit we shouldn't bluntly stop
DocBroker, because even though we cannot save, we
can still upload the latest document version from
disk.
What makes this particularly problematic is that
we request to kill Kit when the last client is
disconnected.
This was caught through unit-tests, albeit it was
not very common or easy to reproduce.
This fixes cases where always_save_on_exit
doesn't save/upload.
Change-Id: Ib66281fc68d4c0264602852b0acdc453b6f79963
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
The timestamps now helpfully include how
long ago they were.
This also improves the conversion between
chrono clocks by adding generic functions.
Change-Id: I18081330564c8d969b9364cf1b9e33546e241280
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>