There might be the case that `fopen($file, 'r')` returns false and thus ownCloud might believe that the config file is empty and thus potentially leading to an overwrite of the config file.
This changeset introduces `file_exists` again which was used in ownCloud 5 where no such problems where reported and should not be affected by such problems.
Ref https://github.com/owncloud/core/issues/12785#issuecomment-71548720
Whenever an exception occurs during scan of a remote share, the share is
checked for availability. If the storage is gone, it will be removed
automatically.
Also, getDirectoryContent() will now skip unavailable storages.
HHVM seems to have problems with this at the moment (even setting those values in the php.ini of HHVM doesn't have helped much) and thus the unit test execution failed.
So it's better if we disable this check for now for HHVM.
A recent change that prevents reshare permission changes to delete group
share children had the side-effect of also preventing group share
children deletion when it needed to be done.
This fix adds an extra flag to isolate the "reshare permission change"
deletion case and keep the other ones as they were before, not only to
fix the regression but also fix other potential regressions in code that
uses this method.
Also updated the comment because now Helper::delete() is no longer
limited to reshares but also applies to group share children.
\Sabre\DAV\Auth\Backend\AbstractBasic::authenticate was only calling \OC_Connector_Sabre_Auth::validateUserPass when the response of \Sabre\HTTP\BasicAuth::getUserPass was not null.
However, there is a case where the value can be null and the user could be authenticated anyways: The authentication via ownCloud web-interface and then accessing WebDAV resources. This was not possible anymore with this patch because it never reached the code path in this scenario.
This patchs allows authenticating with a session without isDavAuthenticated value stored (this is for ugly WebDAV clients that send the cookie in any case) and thus the functionality should work again.
To test this go to the admin settings and test if the WebDAV check works fine. Furthermore all the usual stuff (WebDAV / Shibboleth / etc...) needs testing as well.
This function is called a lot of times and was really slow before due to not reusing the same array.
Previously when it was called 500'000 times it took about 2seconds, now we're down to 0.2 seconds on my local machine.
Ref https://github.com/owncloud/core/issues/13434
There are a lot of clients that support multiple WebDAV accounts in the same application. However, they resent all the cookies they received from one of the accounts also to the other one. In the case of ownCloud this means that we will always show the user from the session and not the user that is specified in the basic authentication header.
This patch adds a workaround the following way:
1. If the user authenticates via the Sabre Auth Connector add a hint to the session that this was authorized via Basic Auth (this is to prevent logout CSRF)
2. If the request contains this hint and the username specified in the basic auth header differs from the one in the session relogin the user using basic auth
Fixes https://github.com/owncloud/core/issues/11400 and https://github.com/owncloud/core/issues/13245 and probably some other issues as well.
This requires proper testing also considering LDAP / Shibboleth and whatever instances.
It's better to encode the string to prevent possible (yet unknown) bugs in combination with PHP's type juggling.
Previously the boolean statements evaluated to either an empty string (false) or a not empty one (true, then it was 1). Not it always evaluates to false or true.
This also removes a stray - that was not intended there but shouldn't have produced any bugs. Just to increase readability.
Thanks @nickvergessen for spotting.
Addresses https://github.com/owncloud/core/pull/13235/files#r22852319
The internal path was matched without the last "/" which caused
"files_trashbin" to also match when the internal path was "files".
This adds the missing slash for the comparison.
Apparently `normalizer_normalize` is not verifying itself whether the string needs to be converted or not. Or does it at least not very performantly.
This simple change leads to a 4% performance gain on the processing of normalizeUnicode. Since this method is called quite often (i.e. for every file path) this has actually a measurable impact. For examples searches are now 200ms faster on my machine. Still not perfect but way to go.
Part of https://github.com/owncloud/core/issues/13221
Isset is a native language construct and thus A LOT faster than using strlen()
On my local machine this leads to a 1s performance gain for about 1 million paths. Considering that this function will be called a lot for every file operation this makes a noticable difference.
`normalizePath` is a rather expensive operation and called multiple times for a single path for every file related operation.
In my development installation with about 9GB of data and 60k files this leads to a performance boost of 24% - in seconds that are 1.86s (!) - for simple searches. With more files the impact will be even more noticeable. Obviously this affects every operation that has in any regard something to do with using OC\Files\Filesystem.
Part of https://github.com/owncloud/core/issues/13221
The check for invalid paths is actually over-complicated and performed twice resulting in a performance penalty. Additionally, I decided to add unit-tests to that function.
Part of https://github.com/owncloud/core/issues/13221
Otherwise every time the AppStore was opened a lot of connections to the AppStore server were made which resulted in a terrible performance.
This changeset will cache the response for a sensible time so that only the first request will be somewhat slow.
Performance changes:
- Loading a category took previously more than 3 seconds on my machine. Now for every follow-up request it takes less than 200ms, resulting in a performance gain of 1950%
- Loading the category list took previously about 750ms - now it takes 154ms, a total performance gain of 395%
