- before we buffered whole content before uploading/download
- this could lead to OOM errors on big files
- now instead we stream when PUT/GET are sent from SVN resulting in faster more efficient data transfer
- use new exceptions for branch permissions
- detect if we need to check permissions
- calculate super-admin the same way as regular users for override capabilities
- added new tests for branch permissions
- fixes Mercurial CVE for lack of permissions checking on mercurial batch commands
- more strict checks for permissions, now default to push instead of pull
to be always on safe side.
- decypher batch commands and pick top-most permission to be used
- This change will allow SVN to execute rhodecode integration framework
via executing the builting hooks.
- uses txn_id for svn to provide metadata required via Apache
- Interception of SVN transaction is required to use temporary stores for metadata
- hooks are now part of vcsserver where the install/control logic is stored
- especially for mysql in case large number of data in caches there could be critical errors storing cache, and thus preventing
users from authentication. This is caused by the fact that we used single namespace for ALL users. It means it grew as number of users grew
reaching mysql single column limit.
This changes the behaviour and now we use namespace per-user it means that each user-id will have it's own cache namespace fragmenting maximum column data to
a single user cache. Which we should never reach.
- in some cases like usage of tortoisesvn non of the available SVN
protocol detection methods are available
- with extended method we're able to catch directory delete/create
- this solves a case when SVN repository needs to find out a repo without
per-repo settings available
- we need to get the global config first to be able to extract repo name
- dependencies: bumped:
-setuptools-scm==1.15.6
-celery==4.1.0
-kombu==4.1.0
-pytz==2017.3
-Routes==2.4.1
Removed all pylons components. Refactored the code to use only pyramid
parts
- this gives a 30% speed increase in operations like svn commit
- generally uses the same mechanismy like auth cache to cache permissions for faster
access to vcs commands
This should solve a problem of uploading large files to rhodecode. In case of
git with small postBuffers GIT client streams data to the server. In such
case we want to stream the data back again to vcsserver without reading it fully
inside RhodeCode.
We we to know to detect vcs type request as early as possible. IF we know
we're handling an VCS request, we can skip some of the pylons stack initialization.
- This fixes issues with leaking sessions and other object from pylons into
a purely VCS type request.
