687e38c409
Merge pull request #2970 from nextcloud/notify-self-test
...
Add self-test for files_external:notify
2017-01-30 17:15:57 -06:00
6fd7d4efd0
remove obsolete hack ( #26928 )
...
* remove obsolete hack
> // file_get_contents() has a memory leak: https://bugs.php.net/bug.php?id=61961
was closed 4 years ago. we could also use the Common implementation
Signed-off-by: Morris Jobke <hey@morrisjobke.de>
2017-01-27 12:20:48 -06:00
269747985e
Add a more powerful notify mechanism
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2017-01-27 10:44:33 +01:00
29f5ab1758
Merge pull request #3238 from nextcloud/dav-test
...
Make DAV external storage test more reliable
2017-01-26 18:43:36 +01:00
95ab46e3d2
Merge pull request #3030 from nextcloud/filesapi-renamefailcase
...
Fix files node API failed rename/copy
2017-01-24 21:23:59 -06:00
38ec2a35f2
Make DAV external storage test more reliable
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2017-01-24 16:56:24 +01:00
d4d116503c
Merge pull request #3212 from mwalbeck/mimetype-hidden-files
...
Mimetype detection for hidden files
2017-01-23 21:45:16 +01:00
3d99ec6832
Reduced code for filename trim
...
Signed-off-by: Magnus Walbeck <mw@mwalbeck.org>
2017-01-23 17:14:39 +01:00
eadc1ac637
Mimetype detection for hidden files
...
Signed-off-by: Magnus Walbeck <mw@mwalbeck.org>
2017-01-23 13:28:35 +01:00
0aaf209c66
Prevent PHP request to get killed when using fclose callback ( #26775 )
...
* Prevent PHP request to get killed when using fclose callback
* Add ignore_user_abort everywhere where the time limit is set to 0
Signed-off-by: Robin Appelman <robin@icewind.nl>
2017-01-23 12:33:20 +01:00
e667b28298
Fix files node API failed rename/copy
...
Whenever a rename or copy operation failed on the view, we must throw
an exception instead of just ignoring.
Signed-off-by: Vincent Petry <pvince81@owncloud.com>
2017-01-12 13:52:59 +01:00
eb5ea0e260
Refactor files node API for more code reuse
...
Signed-off-by: Vincent Petry <pvince81@owncloud.com>
2017-01-11 14:53:42 +01:00
447e388ce9
Merge pull request #2744 from nextcloud/oc_26324
...
[downstream] Fix DAV stat cache to properly cache 404
2017-01-11 08:16:40 +01:00
00b17c3af5
Only check for local shared storages
...
Signed-off-by: Roeland Jago Douma <roeland@famdouma.nl>
2017-01-10 16:14:21 +01:00
706655ef20
Skip local shares in bkg scan and occ files:scan ( #26590 )
...
Local shares should only be scanned when doing it for the owner to
avoid repeatedly rescanning the same shared storage over and over again
for every recipient.
2017-01-10 16:11:45 +01:00
5774d3e82c
replace close:// streamwrapper with CallBackWrapper
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2017-01-06 15:33:32 +01:00
73fcb69cad
Merge pull request #2934 from nextcloud/quota-stream-non-global
...
remove the need to register the quota streamwrapper globally
2017-01-06 15:25:20 +01:00
6d4abdab8f
Fix DAV stat cache to properly cache 404
...
404 errors were not properly cached due to catching the wrong
exception. Now catching ClientHttpException and checking the error
code. In case of 404, adjust the stat cache accordingly.
2017-01-05 09:39:07 +01:00
72a3ea6073
log a warning when trying to use a non basic fopen mode
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2017-01-05 08:49:04 +01:00
bca91f2216
remove static:// stream wrapper
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2017-01-04 15:57:02 +01:00
968de70bc5
remove the need to register the quota streamwrapper globally
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2017-01-04 15:15:47 +01:00
5b09565594
remove no longer used fakedir stream wrapper
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2017-01-03 17:06:03 +01:00
4744dce4df
Nuke the legacy storage fallback from orbit
...
Signed-off-by: Lukas Reschke <lukas@statuscode.ch>
2016-12-23 12:49:46 +01:00
14a561ddad
always use numeric storage id in cacheentry
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-12-13 12:53:38 +01:00
86473306f3
fix cacheentry return wrong storage id
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-12-13 12:50:17 +01:00
6f46a8bd5b
also compare storage ids when checking for changed mounts
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-12-13 12:11:10 +01:00
1253d1008a
Merge pull request #2411 from nextcloud/fix-encryption-home-storage
...
check if the file should really be encrypted before we update the file cache
2016-12-05 15:38:12 +01:00
1131338562
Merge pull request #2486 from nextcloud/shared-storage-mask
...
apply permissions mask for shared storage
2016-12-02 23:39:48 +01:00
02ee4f958c
apply permissions mask for shared storage
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-12-02 17:59:02 +01:00
efbe26f9c3
only update the cache if the storage operation succeeds
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-12-02 16:20:49 +01:00
b7f866988e
Prevent endless loop in \OC\Files\View::createParentDirectories
...
\OC\Files\View::createParentDirectories was previously prone to an endless loop. If a path such as /foo/existingfile.txt/bar/foo was passed and existingfile.txt existed in foo the loop was never left and running until the PHP process timed out.
This commit changes the logic to a foreach loop over an array and additionally additional error handling using is_file.
Signed-off-by: Lukas Reschke <lukas@statuscode.ch>
2016-12-01 16:57:12 +01:00
0f8fe77b3a
check if the file should really be encrypted before we update the file cache
...
Signed-off-by: Bjoern Schiessle <bjoern@schiessle.org>
2016-11-29 20:34:45 +01:00
ce2f9493a7
filter out oc_mounts results from non existing users
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-24 11:44:18 +01:00
558f169671
Move the validation into one place only
...
Signed-off-by: Joas Schilling <coding@schilljs.com>
2016-11-21 09:23:37 +01:00
4652d203e3
Make sure we don't scan files that can not be accessed
...
Signed-off-by: Joas Schilling <coding@schilljs.com>
2016-11-21 09:23:32 +01:00
cc500c2810
Merge pull request #2199 from nextcloud/readfile-close-handle
...
close the file handle after readfile
2016-11-18 14:58:03 +01:00
8b9ad46ba3
Merge pull request #768 from nextcloud/s3-objectstore
...
Add S3 objectstore backend
2016-11-18 14:55:07 +01:00
89d37575a3
close the file handle after readfile
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-18 12:19:49 +01:00
2f03fcab4a
let the share backend get the node cacheentry to save queries
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-17 18:48:38 +01:00
14cdeafc3e
make source cache injectable in cache wrappers
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-17 18:48:38 +01:00
02bfd91553
Merge pull request #2176 from nextcloud/share-cache-root-info
...
reuse the known rootinfo in the shared cache when possible
2016-11-17 18:47:39 +01:00
54ca411ff0
Merge pull request #2167 from nextcloud/mountmanager-lockdown
...
ensure we setup the mount manager when in lockdown
2016-11-17 16:56:55 +01:00
30e7c27d86
reuse the known rootinfo in the shared cache when possible
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-17 14:59:31 +01:00
b373f52fa8
only use one kind of hack for the storage wrappers of the sharing code
...
Signed-off-by: Morris Jobke <hey@morrisjobke.de>
2016-11-16 19:04:47 +01:00
e7b1196dd6
Hack to prevent warning for read-only wrapper in public links
...
Storage wrappers should normally always be registered inside a presetup
hook. However in the public link scenario this is not possible and
there is currently no good alternative with the current architecture.
Instead of logging a warning every time, this fix prevents the warning
to be shown but also adds a FIXME in the code for the future. This is
ok because this app is already using private/internal APIs at the
moment and should be reworked properly in the future.
2016-11-16 19:04:42 +01:00
a7c0d99705
ensure we setup the mount manager when in lockdown
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-16 17:24:37 +01:00
3ab160dd5a
fopen s3 objects directly to work around unexplainable guzzle bug
...
For some reason when a text file started with a valid hex character ([0-9a-f]) it would eat the text untill the first newline
The new code does basically the same thing as guzzle/s3-sdk did only without wrapping everything in a guzzle stream
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-16 15:30:37 +01:00
0ee958595e
Improve isMounted to handle primary storage object store
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-16 15:30:36 +01:00
eefd059716
add amazon s3 objectstore backend
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-16 15:30:36 +01:00
64e896cc0d
split testing of objectstoragestorage and objectstore implementations
...
Signed-off-by: Robin Appelman <robin@icewind.nl>
2016-11-16 15:30:36 +01:00
Morris Jobke
Jörn Friedrich Dreyer
Robin Appelman
Lukas Reschke
Magnus Walbeck
Vincent Petry
Roeland Jago Douma
Roeland Jago Douma
Bjoern Schiessle
Joas Schilling