This is within the failed lock acquiring branch. So the lock is free by another process and should not be removed because the cached file (that was created by the process having the lock) appeared on the filesystem.
Signed-off-by: Morris Jobke <hey@morrisjobke.de>
On some systems with a lot of users this creates a lot of extra DB
writes.
Being able to increase this interval helps there.
Signed-off-by: Roeland Jago Douma <roeland@famdouma.nl>
When asking for sharees we compare not only UID and displayname but also
the email address. And if that matches we return the sharee as an exact
match. This logic had a flaw as in that it also matched the empty string
to users with no email address.
This is most noticeable when you disable sharee enumeration and open the
ownership transfer dialog. It suggested other users of the instance
before. This has stopped now.
Signed-off-by: Christoph Wurst <christoph@winzerhof-wurst.at>
Before when you have a user "smith" and a federated user "smith@example.com"
you could see the federation result with "smit" but not with "smith" anymore.
With most LDAP configurations and local backend setups this is disturbing and
causes issues.
The idea of not showing the email and federation on a matching user was with:
Local user registered with "smith@example.com" user id and having that same
email / cloud id in your contacts addressbook. So we now only hide those
"side results" when the search does contain an @
Signed-off-by: Joas Schilling <coding@schilljs.com>
Since we try to do range requests this will fail hard.
However since empty files are not that interesting to read anyways we
just read from an emptry memory stream.
Signed-off-by: Roeland Jago Douma <roeland@famdouma.nl>
Since all the compiled routes are based on the server webroot,
we have to use this, independent from which app this belongs to.
Signed-off-by: Joas Schilling <coding@schilljs.com>
When your password changes out of bounds your Nextcloud tokens will
become invalid. There is no real way around that. However we should make
sure that if you successfully log in again your passwords are all
updates
* Added event listener to the PostLoggedInEvent so that we can act on it
- Only if it is not a token login
* Make sure that we actually reset the invalid state when we update a
token. Else it keeps being marked invalid and thus not used.
Signed-off-by: Roeland Jago Douma <roeland@famdouma.nl>
Else we might run into the issue that for an external storage where the
size is not known yet we do not use zip64. Which then of course fails on
large zip files.
Signed-off-by: Roeland Jago Douma <roeland@famdouma.nl>
If the object store errors we should not always delete the filecache
entry. As this might lead to people losing access to their files.
Signed-off-by: Roeland Jago Douma <roeland@famdouma.nl>
instead of showing the generic 'Your storage is full' message, better explain that it's the group folder/external storage that is full
Signed-off-by: Robin Appelman <robin@icewind.nl>
* Systems that upgrade have this enabled by default
* New systems disable it
* We'll have to add some wargning in the setup checks if this is enabled
Signed-off-by: Roeland Jago Douma <roeland@famdouma.nl>
Morris Jobke
Roeland Jago Douma
Joas Schilling
Roeland Jago Douma
Julius Härtl
Christoph Wurst
korelstar
Arthur Schiwon
Robin Appelman
J0WI
Elijah Martin-Merrill