I didn't want to touch this visible translated string in a
patch release so I just switched the data.
For Qt versions before 5, we continue to show the build-time version
of OpenSSL but it's clearly marked as such.
In 8ef11a38c9, we started blacklisting
SoftError for 0 seconds. But if the two sync happen with less than
1s interval, we would still prevent them to happen.
So make sure we expire if 0 seconds have expired
The backtrace looks like:
File "atomic_base.h", line 396, in QString::~QString
File "qlist.h", line 442, in OCC::FolderStatusModel::slotUpdateDirectories
This is the only QList operation, and it may crash if the list is empty.
It can be empty if the propfind returned empty results.
I'm not sure how this can be possible to have an empty list there since
the server is always supposed to return at least one entry, for the directory
itself. But it can happen if a directory was transformed in a file, or
if there is a bug on the server.
This is a possible fix for #3850, applying custom window flags (http://doc.qt.io/qt-5.8/qt.html#WindowType-enum) on the connection method dialog. These avoid resizing the dialog and displaying a maximize button on OS X and context help button on Windows.
It is possible to create files with filenames that differ
only by case in NTFS, but most operations such as stat and
open only target one of these by default.
When that happens, we want to avoid uploading incorrect data
and give up on the file.
Typically this situation should never occurr during normal use
of Windows. It can happen, however, when a NTFS partition is
mounted in another OS.
So the title of the column is fully visible.
Adjust the default size and the size policy of the label so it behave
correctly as the window get resized
Issue #5600
The crash reporter shows many crashes in OwncloudPropagator::scheduleNextJob.
We don't really know what could be the cause, but it's probably because
the _activeJobList contains dangling pointer.
So this patch makes sure to remove all the jobs from this list as they get
destroyed.
* For requests:
- reuse the original QNetworkRequest, so headers and attributes
are the same as in the original request
- determine the original http method from the reply and the request
attributes
- keep the original request body around such that it can be sent
again in case the request is redirected
* Simplify the interface that is used for creating new requests in
AbstractNetworkJob.
Backtrace from the crash reporter:
Crash: EXCEPTION_ACCESS_VIOLATION_READ at 0x401
File "moc_activitylistmo_M5OEXJ7XGJYTWT.cpp", line 92, in OCC::ActivityListModel::qt_static_metacall
File "qobject.cpp", line 3730, in QMetaObject::activate
File "moc_networkjobs_7AMNCW4BBANVRK.cpp", line 1342, in OCC::JsonApiJob::jsonReceived
File "networkjobs.cpp", line 714, in OCC::JsonApiJob::finished
File "abstractnetworkjob.cpp", line 207, in OCC::AbstractNetworkJob::slotFinished
File "moc_abstractnetwor_PFI2TXGQHRE33H.cpp", line 98, in OCC::AbstractNetworkJob::qt_static_metacall
File "qobject.cpp", line 3730, in QMetaObject::activate
File "moc_qnetworkreply.cpp", line 367, in QNetworkReply::finished
File "qnetworkreplyhttpimpl.cpp", line 2100, in QNetworkReplyHttpImplPrivate::finished
File "qnetworkreplyhttpimpl.cpp", line 279, in QNetworkReplyHttpImpl::abort
My theory is that the AccountState stored in a property of the job was destroyed.
before the job timed out.
Therefore, the qobject_cast within the qvariant_cast would call the metaObject()
virtual function on a dangling pointer.
Fix it by storing a QPointer instead which will track the deletion.
This leads to crashes since we changed the connection to the parent
jobs not to be queued anymore.
We don't really need to bubble up the finished state through
parents in that case, and it would also mean that we'd recurse
all the way through leaves as we go up to each parent. So just call
abort directly on the OwncloudPropagator and make sure the abortion
call is posted to the event loop.
Avoid using connections to report up the job tree for signals
that we can directly communicate to the OwncloudPropagator.
This slightly reduces the memory usage and avoid passing those calls
through the whole parent chain.
In preparation for the PropagateDirectory refactoring, simplify things
by removing WaitForFinishedInParentDirectory, which is currently
implemented as a one-level check.
This value is important for directory items, but is however never
used since a directory CSYNC_INSTRUCTION_RENAME item will always be in
PropagateDirectory::_firstJob, which will have to pass through its own
PropagateDirectory job's parallelism() before reaching the parent's
_subJobs optimization.
Since PropagateDirectory::parallelism can only return WaitForFinished
or FullParallelism, that value is lost. So this commit doesn't
change the behavior for directories, and allow file renames to be
scheduled in parallel across directories (which isn't a problem).
It could be possible that _firstJob is marked as finished if
aborted before its parent PropagateDirectory was marked as finished,
allowing a posted scheduleNextJob call to schedule the child job
in-between.
- I checked every occurence of a '%2' and make correct use of the
QString::arg overload that takes several argument instead of chaining
them, because the first argument can contains a '%1'
- I tried to look for every label that they either use plain text or richtext
and escape the user provided strings in there.
By default, followRedirects is true for all requests, to transparently
handle redirections. In the wizard, we have special redirect-handling
code though and that was being skipped.
Setting the flag to false allows the wizard to be aware of redirects
and to handle them in the correct way. Tested with the server described
in
https://github.com/owncloud/administration/tree/master/redirectServer
There's a second bug here, where followRedirects always converts
redirected requests to the GET verb. That means redirected PROPFINDs
will never have worked. This change un-breaks them for the wizard only.
There should be no case that previously worked that stops working now.
The test sets OWNCLOUD_MAX_PARALLEL to 1 to disable parallelism.
But since the max amount of parallelism is twice as much, that does not
work.
So change the way we compute the hardMaximumActiveJob: Use the value of
OWNCLOUD_MAX_PARALLEL to maximize this amount and base the maximum amount
of transfer jobs on it instead of the other way.
A result of this change is that, in case of bandwidth limit, we keep the
default of 6 non-transfer jobs in parallel. I believe that's fine since
the short jobs do not really use bandwidth, so we can still keep the same
amount of small jobs.
This reverts commit e1f5a49c21.
Retrying uploads with insufficent storage errors frequently leads to
high server traffic. See #5537 for links and a sketch of a correct
solution.
If we call
setConfiguration(QNetworkConfiguration());
This sets an invalid configuration on the QNAM.
But later, when we really go online because interfaces are discovered,
QNetworkAccessManagerPrivate::_q_onlineStateChanged is called (with isOnline=true).
And this will set the state to disconnected because customNetworkConfiguration is
true, and the networkConfiguration state is disabled.
The workaround we to fix another bug on Windows in which the default network
configuration was not behaving properly.
The issue on linux is hard to reproduce and only happen in some condition,
but it was reproduced on smashbox when they run two owncloudcmd at the same time.
Issues: #4720 , #3600
We were removing the wholme journal db when the user wanted to keep all files,
But that would also remove the selective sync lists.
We should only remove the metadata table.
Issue #5484
- Put all tests in the bin directory so that DLLs can be loaded
- Add missing exports
- Skip tests that use code depending on zlib
- The "GMT" timezone is named differently, use the int constructor instead
5 tests are still failing, it's not really worth fixing at the moment
since no developper is currently using Windows as its main platform.
It could be possible that _firstJob is marked as finished if
aborted before its parent PropagateDirectory was marked as finished,
allowing a posted scheduleNextJob call to schedule the child job
in-between.
Previously, we'd try migrating from legacy settings if reading
the settings failed with an error. Now, we try again after a
couple of seconds and eventually give up.
We can do that because the only changes that were in master but not in 2.3 were the
translations change and documentation change, and the support for the 'M' permission
which we want in 2.3.
The sync engine rely on the 'M' in premission to ask for confirmation
(As requested in issue #5340)
But we only want to ask the premission for the 'root' of the mounting point and not
for every subfolders within it.
So we change the discovery phase in a way that it does not keep the 'M' for
children within the external storage.
Added two checkboxes in the Account Wizard in the advanced page to change the first options.
Also added a checkbox in the general settings to ask for confirmation for external storages.
Theme options allow to hide the checkboxes in the wizard.
As described in issue #5340
We now delete subjobs as their propagation is complete. This allows us
to also release the item by making sure that nothing else is holding a
reference to it.
Remove the stored SyncFileItemVector from SyncEngine and SyncResult
and instead gather the needed info progressively as each itemCompleted
signal is emitted.
This frees some holes on the heap as propagation goes, allowing many
memory allocations without the need of requesting more virtual memory
from the OS, preventing the memory usage from increasingly growing.
This was to catch duplicate emissions for PropagateDirectory but we
don't emit this signal anymore from there.
This fixes a warning about PropagatorJob not being a registered metatype.
This reverts commit fe42c1a818.
Previously if you paused/unpaused a folder for a disconnected account
they would prepare to sync and thus display the 'Waiting...' text. With
this change, folders that can't possibly sync don't show text like
this.
When account connectivity changes, all unpaused folders will be
scheduled anyway.
The avatar image is fetched from the server async, thus connect a signal
from the account if the avatar changes.
Server feature https://github.com/owncloud/core/pull/26872 is needed.
The avatar pixmap is fetched from the server by the Connectionvalidator,
once it has validated the user name, it queries the avatar pixmap.
If the server does not have the avatar route, an empty pixmap is stored.
Stale chunks might be there because a file was removed or would just not
be uploaded, for any reason.
We just start the DeleteJob but we don't care if it success or not.
Relates to https://github.com/owncloud/core/issues/26981
One of the test is testing the case where the file is modified on the server
during the upload. So this test the precondition failed error.
The FakeGetReply logic was modified because resizing a 150MB big QByteArray
by increment of 16k just did not scale when downloading a big file.
Relates to https://github.com/owncloud/core/issues/26981
We do not track the success or error of the DeleteJob because it does not
matter. If it fails, it might be because the chunks were already removed.
If not, the chunks will be stale, but the server must anyway do a few
cleanup from time to time because we do not always remove the chunks
The current logic tried to avoid a DB lookup just to fetch whether
the file is shared or not since that info is already in the
SyncFileItem. The implementation would however need to decrease the
sync count for itself (and parents) before emitting the new status,
thus emitting the OK status for parents before that last child that
ended the propagation for that folder.
Change the implementation to achieve what we want: give the
possibility to decSyncCount to use a pre-fetched sharing state while
still doing the emission for all involved files. This ensures that
the leaf file also gets its status emitted before its parents.
Issue #4797
We currently push the SYNC status for all files that will be propagated,
and then the OK status when those files are propagated.
On top of this, we send those statuses to all clients connected, even
if the socket is kept open by an application that only needed to show
a file open dialog. On macOS we're also using an NSConnection which
means that we have to wait for the RPC call to return from the
extension, which makes bulk status changes possibly heavy.
Reduce the time spent needlessly sending status pushes by limiting
them to files requested through that socket since it connected.
To limit the data to store, only remember the parent directory of
files requested, and store those in a bloom filter.
Note that this adds a requirement to shell extensions: they should
make sure that the status cache only contains entries that have been
requested through the socket API. In other words, the status cache
must be empty when each socket client connects to the socket API.
Otherwise the cached icon type will be shown to the user, and the
SocketAPI won't push new status for that file if it didn't receive
a RETRIEVE_FILE_STATUS.
- Use the looked-up method index also for the invocation
- Do the method name concatenation already on QByteArray since we'll
convert anyway
- Use staticMetaObject instead of metaObject()
Shrinks owncloud binary by 24 KB and libowncloudsync by 14 KB.
I don't know if it has influence on memory usage or runtime speed though.
Was worth a try.
Previously this wasn't happening for errors that were not
NormalErrors because they don't end up in the blacklist.
This revises the resetting logic to be independent of the
error blacklist and make use of UploadInfo::errorCount
instead.
412 errors should reset chunked uploads because they might be
indicative of a checksum error.
Additionally, server bugs might require that additional
errors cause an upload reset. To allow that, a new capability
is added that can be used to advise the client about this.
This displayName() seemed to be based on Account::user() which used
to call _credentials->user(). But then we repurposed user() to be
davUser() and this usage wasn't updated to point back to the username
used for the credentials.
Some custom server use persistent cookies with the auth token. So we should
clear all the cookies when disconnecting.
Account::clearCookieJar is only called from the HTTPCredentials. This funciton
is not used for shibboleth.
There is probably no reasons to keep the HTTP cookie anyway.
Issue #5370
The re-enables the UI, uses Qt API for importing and
stores the certificate/key in the system keychain.
People who had set up client certs need to re-setup the account. This is ok
since it was an undocumented feature anyway.
Fixes#5408#5407.
The problem was that cleanup of the credentials page set the
credentials of the account back to dummy, thereby overriding
things like shib usernames.
This should be broken since a932eac832.
Otherwise it might happen that the model is inconsistant and this can
lead to crash in the worst case.
(For example, if there was a "fetching" label, and we hide it because it
was a 404. In this case, we would not call begin/endRemoveRows, so the
view could still call the model with an index of row 0, that used to be
for the label, but now correspond to the first element of _subs. And
because _subs is empty, this could lead to crashes)
This could make sure that the network job gets deleted if the parent job gets
deleted, and would avoid crashes like:
Crash: EXCEPTION_ACCESS_VIOLATION_READ at 0xffffffff8b008a04
File "qiodevice.cpp", line 1617, in QIODevice::errorString
File "propagatedownload.cpp", line 264, in OCC::GETFileJob::slotReadyRead
File "moc_propagatedownload.cpp", line 85, in OCC::GETFileJob::qt_static_metacall
File "qobject.cpp", line 3716, in QMetaObject::activate
File "moc_qiodevice.cpp", line 154, in QIODevice::readyRead
File "qnetworkreplyhttpimpl.cpp", line 1045, in QNetworkReplyHttpImplPrivate::replyDownloadData
(#5329)
The csync log level was only set up on startup, and for log files.
Fix the issue by making Logger::isNoop rely on being explicitly activated
for the log window instead of relying on the presence of a connected
signal, and move the csync log level logic in Logger.
The compiler seems to use signed enums and we need to reserve an extra
bit for the sign to avoid the 2 value to overflow and being interpreted
as -2 when read, and thus not being correctly compared to the full enum
value.
The rules to select the webdav url are now:
- If the server reports that the new chunking algorithm is working,
always use remote.php/dav/files/<username>
This capability can be overriden with an environment variable
- Otherwise, use the dav path provided by the theme, which defaults to
remote.php/webdav
This means that with the newer server, the branding can no longer override
the webdav URL. If there is still an usecase for the branding to do so, we
need to find another way to override it. But it is now more complicated to
configure as might need include the username and need different endpoint
depending on the operations (chunks or not)
Issue #4007
As the URL might be print on the logs.
Also don't change the scheme from http to owncloud.
This was required before when we were using neon through csync, but now that
we use QNAM for everything we don't need it. The credentials from the account
are used.
We are going to change the webdav path depending on the capabilities.
But the SyncEngine and csync might have been created before the capabilities
are retrieved.
The main raison why we gave the path to the sync engine was to pass it to csync.
But the thing is that csync don't need anymore this url as everything is done by the
discovery classes in libsync that use the network jobs that use the account for the urls.
So csync do not need the remote URI.
shortenFilename in folderstatusmodel.cpp was useless because the string is the
_file of a SyncFileItem which is the relative file name, that name never
starts with owncloud://.
All the csync test creates the folder because csync use to check if the folder
exists. But we don't need to do that anymore
Some listeners detect whether a sync is finished by checking
for isUpdatingEstimates and completedFiles >= totalFiles. But
if a sync didn't transfer any files we never sent signal
with these values. Now we do.
The "S" in the permission is only for the "Shared with me" files.
It is only used to show the shared status in the overlay icons.
But we also wish to show the shared status for files that are shared
"by" the users. We can find that out using the 'share-types' webdav
property. If set, then we are sharing the object.
We fake a 'S' in the permission as for our purpose, they mean the same.
Issue #4788
The ownCloud 9.1 server has a data-fingerprint property that the admin must
change in case of backup restoration. When this change, the client understands
that a backup was restored, and will generate conflict files and re-upload
new files.
The heuristics based system checks that there is at least two files wose mtime
is put back in the past and no files that goes forward. In that case we ask the
user before creating the conflicts.
This commit disable the heuristics for newer server that have the data-fingerpint.
And change the heuristics to two hours because we want to avoid false positive due
to some clock error, and that 2 hours of lost due to backup restoration is probably
not so bad.
We only ask the user in the heuristics based aproach so in practice this mean that
the "backup-detected" dialog will no longer appear with newer server.
Relates issues #5260, #5109
Otherwise progress listeners think it's still the last-completed
item when the next sync starts. This lead to spurious entries in
the "Recent Changes" list.
Update from commit 05ce8a23cdc12e825532dc6de06c267fb8d48b4f from
https://github.com/dragotin/QProgressIndicator
Which itself is forked from commit e5ba0fd09bfd43b067ee3646d70b294c7efcb558 from
upstream, with additional license header.
It was relicensed to MIT according to
14bb9d10e2
Relates to issues #5180 and #5184
Issue #5224
Two problems:
- In the discovery phase, we need to check the selective sync entries of
the source path in case of renames.
- When the rename is done, we need to actually update the black list in the
database.
Some tests (such as FolderManTest) can polute the config file with invalid
accounts.
(That's because most of the code, (even in libsync) always instentiate a ConfigFile)
The crash reporter shows a lot of crashes in sqlite3_clear_bindings
which seems to indicate that _stmt is null. We should guard against
a null value in order to avoid crashing.
This should only happen if the prepare call fails. We don't usually
check the return value of the prepare call, but if _stmt is null, the
exec call should return false, not true. We check the result of the
exec call, so this should then abort the sync with an error, rather
than crashing.
- Use a white icon if the context menu is visible.
- Enable `QIcon::setIsMask` if compiled on Qt >= 5.6 to allow automatic
macOS color handling.
- No changes if the colored icons are used.
One local folder can now be configured as sync target for multiple
accounts as long as their url and user differ.
Also this patch accepts that the sync folder is behind a symlink.
Also this patch fixes a bug that before the user input was taken
canonically which was not working for the symlink handling.
Instead of using the regular selective-sync UI (where it's unclear what
the "Cancel" button would even mean in this context), provide a
different set of buttons that allow the user to quickly synchronize
all pending big folders, none of them, or perform manual changes
as usual.
Tray: Workaround collection
* QDBus workaround for Qt 5.5.0 only, there were reports of the tray
working fine with 5.5.1. #5164
* OWNCLOUD_FORCE_QDBUS_TRAY_WORKAROUND to force the workaround on an off
* OWNCLOUD_TRAY_UPDATE_WHILE_VISIBLE to enable or disable updating of
the menu while it's visible - disable by default due to problems on OSX and Xubuntu.
* Track the visibility of the tray menu with aboutToShow/aboutToHide
only on OSX - the aboutToHide signal doesn't trigger reliably on linux
* Refactor such that setupContextMenu is different from updateContextMenu
* Don't use on-demand updating of the tray menu when the qdbus workaround
is active, instead to occasional (30s) updates of the tray menu.
We ignored csync log, but we also need to silent Qt debug output.
We need to ignore it at the very begining because there might be
qDebug also in account creation.
Issue #5196
Two bugs:
- The change filed are not considered as move, they are re-downloaded
but the old file was not removed from the database. The change in
owncloudpropagator.cpp takes care of removing the old entries.
- Next sync would then remove the file in the server in the old folder
This was not a problem until we start reusing the sync engine, and
that the _renamedFolders map is not cleared. We were before deleting
a non-existing file. But now we delete the actual file.
Also improve the tests to be able to do move on the server.
This include support for file id.
Issue #5192
In case of the root directory, it may happen that the _item
is empty and the _item->_status is NoStatus. But we still need to report
the proper success or error of the whole propagation. We should really
use _hasError for that. However, _hasError is also defined to NoStatus
if there was no error, so in that case we need to set Success.
This fixes the problem in which the data-fingerprint is not saved on the
database because the SyncEngine think that the sync failed. (Issue #5185)
We wwer enot connecting to the right signal from the check server
job, and therefore we were not catching the condition in which the
json was invalid. We would then never terminate the ConnectionValidator job.
Note that instanceNotFound is also emited if there is a network error.
The log looked like this:
10:25:51.247 OCC::CheckServerJob::finished: status.php from server is not valid JSON!
10:25:51.248 OCC::CheckServerJob::finished: status.php returns: QMap() QNetworkReply::NetworkError(NoError) Reply: QNetworkReplyHttpImpl(0x2b6a790)
10:25:51.248 OCC::CheckServerJob::finished: No proper answer on QUrl("http://localhost/~owncloud/status.php")
10:26:23.235 OCC::AccountState::checkConnectivity: ConnectionValidator already running, ignoring "owncloud@localhost"
10:26:55.235 OCC::AccountState::checkConnectivity: ConnectionValidator already running, ignoring "owncloud@localhost"
[...]
- Replace functions that are provided by MinGW with a Win32-based
implementation
- Explicitly export needed symbols from ocsync.dll
- Rename share.h to sharemanager.h since the name clashes with one
of the Windows headers and get included from there
- Remove the timestamp from the fallback csync stderr logging, it's
not used since we always provide a log callback
These are not understood by owncloud yet, but were requested for CernBox
OC-Total-Length in the MKCOL: The full lenght of the file
OC-Chunk-Offset in the PUT: The offset within the file in which this chunk belongs
OC-Checksum in the MOVE: The transission checksum
Two bugs:
- The change filed are not considered as move, they are re-downloaded
but the old file was not removed from the database. The change in
owncloudpropagator.cpp takes care of removing the old entries.
- Next sync would then remove the file in the server in the old folder
This was not a problem until we start reusing the sync engine, and
that the _renamedFolders map is not cleared. We were before deleting
a non-existing file. But now we delete the actual file.
Also improve the tests to be able to do move on the server.
This include support for file id.
Issue #5192
(cherry picked from commit 85b8ab178e)
We wwer enot connecting to the right signal from the check server
job, and therefore we were not catching the condition in which the
json was invalid. We would then never terminate the ConnectionValidator job.
Note that instanceNotFound is also emited if there is a network error.
The log looked like this:
10:25:51.247 OCC::CheckServerJob::finished: status.php from server is not valid JSON!
10:25:51.248 OCC::CheckServerJob::finished: status.php returns: QMap() QNetworkReply::NetworkError(NoError) Reply: QNetworkReplyHttpImpl(0x2b6a790)
10:25:51.248 OCC::CheckServerJob::finished: No proper answer on QUrl("http://localhost/~owncloud/status.php")
10:26:23.235 OCC::AccountState::checkConnectivity: ConnectionValidator already running, ignoring "owncloud@localhost"
10:26:55.235 OCC::AccountState::checkConnectivity: ConnectionValidator already running, ignoring "owncloud@localhost"
[...]
(cherry picked from commit ff701bd473)
This is not sufficient as it is not working for the Socket API.
Next commit will fix it in another layer.
Also, not ignoring paths that are not inside the folder is wrong
as it might still happen if the name has a different casing
This reverts commit d5a481f132.
(cherry picked from commit 904cd46f75)
Previously, we only checked the hiddenness of the target file and
ignored the hiddenness of the containing folders. This lead to
undesired behavior when people synced their home folders and there
was a folder watcher notification for a non-hidden file in one of
the hidden folders.
I'm not fully sure why, but sometimes notifications for .foo/bar were
already ignored, but notifications for .foo/bar/car were not. This may
be because of how we set up the FolderWatchers on Linux.
The new behavior is to check all path components for hiddenness, up
until the base path (but excluding the base path, so using a hidden
folder as a sync folder will work).
This is important because we compare the paths from the file system watcher if it
starts with this path.
Same in the SocketAPI where we need to use cannonical paths in the REGISTER_PATH command,
as the plugin themself will do this comparison.
Issue #5116
This is not sufficient as it is not working for the Socket API.
Next commit will fix it in another layer.
Also, not ignoring paths that are not inside the folder is wrong
as it might still happen if the name has a different casing
This reverts commit d5a481f132.
If the folder has different case in the settings and in the FS, we should
not ignore all the files. This is important for the files system watcher.
(cherry picked from commit 98268d102f)
Missing deleteLater when the CleanupPollsJob aborts.
This is only a problem if the SyncEngine is kept alive a long time. Which is
usually not the case in the configuration where poll jobs are used.
(cherry picked from commit 3465024898)
I got into a situation where the model would endlessly request the directory
contents from the server because we did not notice yet that the server
is actually in maintenance mode while we were expanding the tree view when
changing the tab to the account or when just expanding it by clicking.
(cherry picked from commit 524220d090)
I got into a situation where the model would endlessly request the directory
contents from the server because we did not notice yet that the server
is actually in maintenance mode while we were expanding the tree view when
changing the tab to the account or when just expanding it by clicking.
Before it was in Folder, however, the command line client does not
have the Folder class. To not duplicate code, the function to generate
the sync journal name went to SyncEngine class.
This would happen if the directory would first need to be created
through an mkdir propagation job. This job's itemCompleted signal
would trigger the directory to show as SYNC even though its children
are still propagating.
Fix the issue by tracking the sync count for each file, affecting
its parents. This allows us to get rid of the O(n) vector lookup
for each status query, and properly track the hierachical sync
status of a directory.
This also removes the itemCompleted signal emission from the
PropagateDirectory job. Since we only needed for overlay icons, and
since this job doesn't do any direct propagation, we can remove it
to ensure that we won't call itemCompleted twice for the item attached
to Propagate*Mkdir jobs (since the PropagateDirectory is backed by
the same SyncFileItem, instruction and status).
The current way of tracking the need to update the metadata without
propagation using a separate flag makes it difficult to track
priorities between the local and remote tree. The logic is also
difficult to logically cover since the possibilities matrix isn't
100% covered, leaving the flag only used in a few situations
(mostly involving folders, but not only).
The reason we need to change this is to be able to track the sync
state of files for overlay icons. The instruction alone can't be
used since CSYNC_INSTRUCTION_SYNC is used for folders even though
they won't be propagated. Removing this logic is however not possible
without using something else than CSYNC_INSTRUCTION_NONE since too
many codepath interpret (rightfully) this as meaning "nothing to do".
This patch adds a new CSYNC_INSTRUCTION_UPDATE_METADATA instruction
to let the update and reconcile steps tell the SyncEngine to update
the metadata of a file without any propagation. Other flags are left
to be interpretted by the implementation as implicitly needing
metadata update or not, as this was already the case for most file
propagation jobs. For example, CSYNC_INSTRUCTION_NEW for directories
now also implicitly update the metadata.
Since it's not impossible for folders to emit CSYNC_INSTRUCTION_SYNC
or CSYNC_INSTRUCTION_CONFLICT, the corresponding code paths in the
sync engine have been removed.
Since the reconcile step can now know if the local tree needs metadata
update while the remote side might want propagation, the
localMetadataUpdate logic in SyncEngine::treewalkFile now simply use
a CSYNC_INSTRUCTION_UPDATE_METADATA for the local side, which is now
implemented as a different database query.
Add a missing call that we currently only do in slotItemCompleted.
This would normally only affect the first sync and would have
gotten properly update at the end of the sync anyway.
To be able to test the SyncEngine efficiently, a set of server
mocking classes have been implemented on top of QNetworkAccessManager.
The local disk side hasn't been mocked since this would require adding
a large abstraction layer in csync. The SyncEngine is instead pointed
to a different temporary dir in each test and we test by interacting
with files in this directory instead.
The FakeFolder object wraps the SyncEngine with those abstractions
and allow controlling the local files, and the fake remote state
through the FileModifier interface, using a FileInfo tree structure
for the remote-side implementation as well as feeding and comparing
the states on both side in tests.
Tests run fast and require no setup to be run, but each server feature
that we want to test on the client side needs to be implemented in
this fake objects library. For example, the OC-FileId header isn't
set as of this commit, and we can't test the file move logic properly
without implementing it first.
The TestSyncFileStatusTracker tests already contain a few QEXPECT_FAIL
for what I esteem being issues that need to be fixed in order to catch
up on our test coverage without making this patch too huge.
This is a move away from the original policy where jobs
would only follow redirects in special cases.
Two restrictions are in place:
1. We do not allow protocol downgrades (https -> http)
2. We stop redirects after we find them looping (e.g. old = new url, or
indirectly when looping 10 times).
This is closer to RFC conforming behavior, although currently
we will treat 301 replies like they were 302. This is for a separate
commit.
Error handling (and display) also needs improvement.
Addresses #2791
Once upon a time, the SyncEngine was instantiated once per sync. But now that
the SyncEngine is kept between sync, we need to reset all these variable between
syncs.
Reverts commit 622017adcf
Could be the cause of #5092 and the cost is higher than the benefit if this is the case.
A network request taking more than 30 seconds isn't something unlikely in this world
and shouldn't be a good reason to abort. We should try to untangle the threads
dependencies to properly fix this if possible instead.
It opens a window and connects to a cipher test
page, showing the output from there, that helps for debugging.
The window is enabled by setting the environment variable
OWNCLOUD_SHIBBOLETH_DEBUG
Missing deleteLater when the CleanupPollsJob aborts.
This is only a problem if the SyncEngine is kept alive a long time. Which is
usually not the case in the configuration where poll jobs are used.
While loading the account, only override the server url if Theme::forceConfigAuthType
is set. This restore the behavior from the client 2.1 for theme that did not
use Theme::forceConfigAuthType.
Issue: owncloud/enterprise#1418
This was removed in 0194ebb222
because it breaks on Linux. However, it looks like it is correct
for Windows. In the meantime the surrounding ifdef has changed
from !Q_OS_MAC to Q_OS_WIN, so reverting it makes sense.
Since the SyncEngine now quits and waits for the discovery thread,
the main thread can enter a deadlock where the discovery thread waits
for its directory result.
Add a 2 seconds timer to the discovery thread wait condition
to limit the deadlock time.
This disables the workaround 487e1fdca5ee04fc98c1ed77898df70d740967c8
for servers that are new enough to support fine grained permissions
on federated shares.
The consequence is that the 'reshare' permission is now granted by
default and that users can edit permissions on the usual fine-grained
level again.
The way the client deals with servers <9.1 is unchanged.
Use a QMap to avoid using a full hashtable for only a few entries, and
clear the QMap once we're done with the measuring. This saves a few
hundred bytes per job during propagation that would otherwise only be
freed at the end of the sync.
During propagation, we create a line for each file, taking memory, but
we delete all lines passed 2000 right at the beginning of the next sync.
Since the user has little chances of being able to read past those 2000
lines in the log, we might as well keep it capped at 2000 also during
propagation to prevent it from eating memory.
The SyncRunFileLog owned by the Folder must be destroyed after the
SyncEngine since the SyncEngine will abort during destruction, resulting
in all jobs being aborted.
It's possible that this crash only happens with a debug build.
The FolderWatcher inserts files to be marked as SYNC and we
currently assume that all file statuses will be updated by the
following sync. It's however possible that the FolderWatcher
notify us of a change that csync won't consider necessary to
propagate, in which case a new status wouldn't be pushed and
the file manager would continue showing this file as syncing.
Re-push the file status when emptying the dirty files list
before propagating to avoid this issue, most likely the OK
status.
No need to allocate (and initialize to 0) a 10 MiB buffer for each files, even
when most files are much smaller than that.
So make sure the buffer that we allocate is not bigger than the file size.
And Also 10 MiB is a bit big for a buffer. 500 KiB should be more than enough.
(Too big allocations can cause problem because of memory fragmentation and such)
We first need to set the abort flag to csync and then aborting the discovery
job, otherwise, the discovery thread could start a new job in the mean time.
We also need to make sure that the thread has existed before we destroy the
exclude list.
Events from the crash reporter suggest that the QNAM and its
child replies might get deleted before returning from this method
and the only possible cause we can see is that the inner event
loop has something to do with it.
Try keeping a ref on the QNAM while in this method to make sure
that it won't get deleted by the inner event loop.
Same fix as in commit 60c101d9
From the crash reporter:
Crash
EXCEPTION_ACCESS_VIOLATION_READ at 0x4
qnetworkreply.cpp in QNetworkReply::request at line 476
propagateupload.cpp in OCC::PUTFileJob::slotTimeout at line 100
moc_abstractnetworkjob.cpp in OCC::AbstractNetworkJob::qt_static_metacall at line 98
qobject.cpp in QMetaObject::activate at line 3716
moc_qtimer.cpp in QTimer::timeout at line 192
qtimer.cpp in QTimer::timerEvent at line 247
qobject.cpp in QObject::event at line 1267
qapplication.cpp in QApplicationPrivate::notify_helper at line 3722
qapplication.cpp in QApplication::notify at line 3505
qcoreapplication.cpp in QCoreApplication::notifyInternal at line 932
FolderDefinition::save and load escapes the alias. We also need to escape
it when we remove it.
New folder can't be created with alias that needs escaping, but old folder
from old config may still exist, and we must allow user to delete them.
For issue #4927:
On Windows 10, we get a notification after the sync is finished for file that were
just downloaded. The guard we have against our "own changes" are only working when
the sync is running and the OwncloudPropagator still alive.
Setting the Environment variable only for owncloud makes in inconsistant with
other Qt application running at the same time.
The users can still set it themself for the whole desktop if they wish.
Addresses #4840
Previously rejecting any kind of certificate meant that the user
was never asked again, even if the certificate changed.
Now we keep track of which certificates were rejected and ask again
if the ones mentioned in the ssl errors change.
mitmproxy is excellent for testing this.
* Progress: Don't count dirs without propagation jobs #4856
These directory SyncFileItems are necessary for bookkeeping
but should not influence the progress display at all.
* Progress: Skip ignored files #4856
The problem in this case is if we rename the file "xxx" to "invalid\file".
The rename will fail because the new filename constains a slash, and it
will be blacklisted.
But then if the user re-rename the file to "valid_name", then we should
invalidate the blacklist entry and retry to upload. But we did not do
that because renaming don't change the mtime and we did not store the
rename target in the database
IL issue 558
This fixes an issue in which too many jobs are started un parallel
while uploading many files, which could cause too much memory usage as the
chunks are stored in memory.
Probably the fix for #4611
Issue #4855
A typo in the context string made the translation lookup fail.
But also the %Ln was not recognized as a plural form by transifex, so only
the singular was translated
The assert was there to make sure that this case wasn't happening
to eventually be properly tested. Remove the assert for now but this
codepath should eventually be unit tested using this specific situation.
Since the windows implementation first does cache lookups using the
path string, directories need to be passed identically as through
RETRIEVE_FILE_STATUS.
Change the convention to never have a trailing slash for directories
in the protocol. This allows the convention to be applied without
having to access the disk (since we'd need to know if the path is
represented by a directory) and also matches the convention of the
rest of the sync engine. Individual file manager plugins are then
responsible of handling pushed paths as not ending with a trailing
slash.
This also:
- Moves the trailing slash removal logic from the SyncFileStatusTracker
to the SocketApi class
- Remove the unneeded QString::normalized call in fileStatus, since
this should already be done by the FolderWatcher and plugins
Since the statuses are cached and that we can't invalidate the cache,
sending NOP would need to be overwritten by the default OK status
once the client successfully connected. But instead of remembering
which files we NOPed, rather wait until we are ready to sync before
sending the REGISTER_PATH message to the socket API client. It will
also prevent the client from sending unnecessary RETRIEVE_FILE_STATUS
requests.
Also remove AccountState::canSync, since it does the same as
isConnected and syncing is not an account responsibility.
Go through fileStatus like other cases to make sure that all use
cases go through the same code path. This also makes sure to use
lookupProblem which will use lower_bound which is more efficient
for larger sets of sync problems.
This also fixes the issue with lookupProblem that prevented it to
properly match an empty pathToMatch, caused by the fact that the
problem map contains relative paths not starting with a slash.
Make sure that we push the new status when the status of the SyncEngine
changed. SyncEngine::started comes a bit late, only when the propagation
starts, although it's better in this case since child folders will
only switch to Sync in aboutToPropagate.
Also fix an issue with SyncEngine::findSyncItem when using an empty
fileName; this would match and return the wrong item, even though
not currently happening with the code since fileStatus won't call
it with an empty fileName anymore.
As before, we rely on metadata-update SyncFileItem entries for parent
directories to notify us that a directory contains files to propagate,
and to know when all children were propagated through its itemCompleted
signal.
Those metadata SyncFileItems however have a None direction and we need
to add a explicit directory check to show them as Sync.
This fix also handles new files as well as existing ones, so no need
to keep a separate logic for new files.
propagatedownload.cpp:712:35: error: 'seenLockedFile' is a protected member of 'OCC::OwncloudPropagator'
Signals are protected in Qt4 but public in Qt5, mark the class accessing it
as friend when compiling with Qt4
Do the visual stuff from designer.
The previous code that was ment to change the color in red did not work
and changed it to gray instead.
Also I don't see why there should be a frame.
Issue #4773
When a conflict-rename or a temporary-rename fails, notify the
LockWatcher. It'll regularly check whether the file has become
accesible again. When it has, another sync is triggered.
owncloud/enterprise#1288
If the downloaded file is empty but the PROPFIND previously announced it
should not have been empty, this might mean the file was somehow corrupted
because of a bug on the server and that we should therefore not accept
the file.
Normaly we accept a change between the actual size of the file and what we
got during discovery because the file might have been updated to a new version
inbetween. But after this patch we won't accept the file if it was replaced
by an empty file.
Will help for issue #4583
Also requested by IL for issue 548
This uses the file watcher to keep track of files that were modified
in order to assign them the blue icon.
This is transient state that's not persisted across restarts.
Before commit 1a51b6718a, the wizard was
making sure folder had an alias but this is no longer the case.
So generate still an unique alias.
Alias is not used in the UI any longer, it's just use for internal purposes.
For issue #4737
As discussed on issue ##4460
Having the quote to be queried on subfolder is wrong in the generic case,
so add a branding option to configure it.
This partially reverts commit ff4cdc3161
In the before-propagate slot, new files that wait to be
pushed to the server are remembered in the _syncProblems
map. That way, the parents show a sync icon properly as
asked for in #4682.
After the item has been transfered properly, the item is
removed from the map again because success is the default.
Added in previous commit from pull request #4663
As discussed, we do not need this option so no need to introduce
a new dependency on the config file in the sync engine
If an app modifies the expiration date (for example the password policy
app) then on more recent versions of the server we will get the share
object back REST style. We should use that info!
Fixes#4409
* Add checksums/supportedTypes and checksums/preferredUploadType
capabilities. The default is that no checksum types are supported.
* Remove the transmissionChecksum config option. Servers must now
use the capabilities to indicate that they are fine with the
client sending checksums.
Note: This intentionally breaks brandings that overrode
Theme::transmissionChecksum. The override must be removed and the
server's capabilities must be adjusted to include the new values.
The reason is that updateFolderView is invoked by the
emitted signal folderSyncStateChange() anyway.
This will reduce the traffic over the SocketAPI nicely,
maybe this was the reason why it was slower than before.
There were two issues:
* With the refactoring of how Folder and SyncEngine relate, the
ignore_hidden_files flag on the CSync context was reset after
each sync run and not updated from the configuration again.
* The folder watcher failed to enumerate hidden folders and thus
didn't watch for changes inside them. (linux only)
SQLITE_DONE is the indicator for not more query results, which is a legal
thing and not an error condition.
Also, check _getFileRecordQuery for null pointer, as close() wipes it.
The problem with QSet is that the QDateTime was part of
the hash, but that does not make sens as it should be unique
per widget and not per <date, widget>
Instead make it a QHash so there is only one entry per widget.
The idea is that the next call to any database operation will try to
reopen the database through the checkConnect() method. So even if there
was a disconnect trom the db file, this will reestablish the connection.
Imagine tgus scenario on a read only share that you move file from
one location to a new directory in the read only share.
Creating the read only directory fails for permission error.
But we should also restore the files that have been moved.
IL issue 542
Bring back the hardcoded status logic for excluded files.
Since the activity log doesn't even mention those files on purpose,
we can't rely on the SyncEngine to notify us about the status of those files.
Looking up a/aa while an error is present in a/aab/aaba would return
a warning status since a/aa is a substring of a/aab.
Fix the issue by checking if the following character is a slash.
This prevents having to define a Problem structure with dubious
operator overloads to accomplish the same.
Also use std::map::lower_bound to quickly iterate over the
list of problems.
* Remove duplicate remote path
* Use thin progress bar
* Move bandwidth and file info to tooltip
* Shorten overall progress message
This also fixes#4562 by making the layout not dependent on the
width of the displayed text.
This also remove all smartness from the SocketApi about the status
of a file and solely use info from the current and last sync.
This simplifies the logic a lot and prevents any discrepancy between
the status shown in the activity log and the one displayed on the
overlay icon of a file.
The main benefit of the additional simplicity is that we are able
to push all new status of a file reliably (including warnings for
parent folders) to properly update the icon on overlay implementations
that don't allow us invalidating the status cache, like on OS X.
Both errors and warning from the last sync are now kept in a set,
which is used to also affect parent folders of an error.
To make sure that errors don't become warning icons on a second
sync, SyncFileItem::_hasBlacklistEntry is also interpreted as an error.
This also renames StatusIgnore to StatusWarning to match this semantic.
SyncEngine::aboutToPropagate is used in favor of SyncEngine::syncItemDiscovered
since the latter is emitted before file permission warnings are set on the
SyncFileItem. SyncEngine::finished is not used since we have all the
needed information in SyncEngine::itemCompleted.
SyncFileStatus' purpose is to track overlay icon status.
Instead of putting comments and default: clauses in switch
on both sides about unused enums, use different enums.
This also remove STATUS_NEW which is the equivalent of
STATUS_SYNC in all shell extension implementations, and
remove STATUS_UPDATED and STATUS_STAT_ERROR which have
the same semantic as STATUS_UPTODATE and STATUS__ERROR.
This currently is no-op code since the socket API isn't notified
that the tainted folder list changed, and the result is the same
since a sync will be triggered within the next 5 seconds and the
modified folder will be shown as SYNC at that point anyway.
Removing the dependency to the file watcher allows moving the
status estimation logic to libsync.
This will help moving the SyncEngine construction in the constructor
and allow moving functionalities from Folder to SyncEngine or its
delegated objects.
The policy that was said is that if a notification has no action, the
client can and should display a close-button. This patch does it.
In additon to that, the client needs a blacklist of closed notifcations
otherwise they would re-appear next time the server notifications are
fetched again.
Also, changed the cleanup of not-longer-used widgets to be more robust.
If a notification is not longer in the list that comes from the
server, the notification is removed.
That is mainly for the notifications that are created by the
announcement application
Since the quota is a per-folder value, this will make the displayed data
more useful when a single sync folder is configured.
Of course each subfolder could have a different quota again.
The date we receive from the server is an ISO8601 datetime that
includes the offset from UTC. Qt does correctly parse this
information and creates the appropriate QDateTime object.
Calling setTimeSpec(UTC) will force the timezone offset to 0 and
thereby change the referenced point in time to an incorrect one.
Soldiering on with a broken or incomplete response could lead to
incorrect sync behavior.
Since discovery uses LsCol jobs which already handle errors
correctly, this should not have a significant impact.
the size on the server might be different from the size on the client
with certain backend so it should be ignored.
(cherry picked from commit 9222db6df9b19a21e1bea5a238d745d96a6385e3)
In SQLite bindings are not cleared by sqlite3_reset() calls, so
skipping a sqlite3_bind call to create a NULL value doesn't work,
instead the previous value will be written.
To fix this, I clear all bindings in SqlQuery::reset and make sure
to explicitly bind NULL when desired in SqlQuery::bind.
To make sure there's no confusion about SqlQuery::reset and
sqlite3_reset, I rename our method to reset_and_clear_bindings().
(cherry picked from commit 7bd4f95b8c)
In SQLite bindings are not cleared by sqlite3_reset() calls, so
skipping a sqlite3_bind call to create a NULL value doesn't work,
instead the previous value will be written.
To fix this, I clear all bindings in SqlQuery::reset and make sure
to explicitly bind NULL when desired in SqlQuery::bind.
To make sure there's no confusion about SqlQuery::reset and
sqlite3_reset, I rename our method to reset_and_clear_bindings().
The isValid check should be used everywhere the capabilities
are used as the loading of the capabilities is happening
in parallel of the startup, so it is not guaranteed to be
available always.
As interaction is required, the notifications are displayed in a
separate widget above the server activity list.
Note that design and also where we display the notifications can
still be discussed and changed.
If the ownCloud server does not have the activity app enabled,
it returns 999 as status code. If all the configured accounts
do that, this code hides the entire tab with the server
activities.
This is supposed to fix#4533
Helps with small file sync #331
When I benchmarked this, it went up to 6 parallelism and
was about 1/3 faster than the previous fixed 3 parallelism.
Doing more than 6 is dangerous because QNAM limits to 6 TCP
connections and also the server might become a bottleneck.
Should also help for #4081
Previously one could accidentally call Folder::setSyncPaused() and miss
some expected side effects. Before, the correct call was to FolderMan::
slotSetFolderPaused(). Now the setter on Folder has the expected effect.
This will be useful if we ever want to store account-level gui state.
I built this originally because I thought a paused account would be
this kind of state.
The creation doesn't need to be separated from the SyncEngine anymore.
This allows the SyncEngine to be created in fewer steps if we want to
use it in tests.
This moves most of the direct csync code from Folder into the SyncEngine.
The exclude file logic for the context has been wrapped using the
existing ExcludedFiles class as well.
Given that we control all call sites, the only way that this can fail is during
OOM. Also remove the code in csync itself to make sure that it's obvious that
any new error case wouldn't be handled by call sites.
As discussed with Klaas, this seems to be a better compromise.
10MB * 3 prarralel jobs = 30MB in memory, and to retry in case of
disconnection. Which is still reasonable. And might make the upload
almost twice as fast on fast network where the amount of chunk is the
bottleneck (because of more server processing)
Relates to issue #4354
