Files
@ 3505a6be2988
Branch filter:
Location: kallithea/docs/usage/locking.rst - annotation
3505a6be2988
1.1 KiB
text/prs.fallenstein.rst
repogroups: fix private repo recursion check
The purpose of this check is to ensure that we don't recursively assign
"default" user perms for a repo with the "private" flag set (because in
that case, the "default" user perms should always be "no access").
(The check, and this fix, is of course only applicable to Kallithea
instances that have anonymous access enabled to begin with.)
However, the check was only functional if the user was specified as a
username. This is apparently always the case when Kallithea is running,
but was not e.g. the case in the test suite, which consistently passed
a user ID instead of a username.
This commit ensures that the user is always resolved before the check is
made. There's no significant overhead to this, as the code immediately
calls RepoModel().grant_user_permission, which resolved the user anyway.
This change just moves the database lookup a bit earlier.
Fixing this revealed the matching test case to be broken, so it has been
fixed as well.
Down the road, we should eliminate Kallithea's bizarre practice of
passing around usernames and user IDs, in favor of passing actual User
objects. That'll get rid of mistakes like these, as well as repeated
needless database lookups.
The purpose of this check is to ensure that we don't recursively assign
"default" user perms for a repo with the "private" flag set (because in
that case, the "default" user perms should always be "no access").
(The check, and this fix, is of course only applicable to Kallithea
instances that have anonymous access enabled to begin with.)
However, the check was only functional if the user was specified as a
username. This is apparently always the case when Kallithea is running,
but was not e.g. the case in the test suite, which consistently passed
a user ID instead of a username.
This commit ensures that the user is always resolved before the check is
made. There's no significant overhead to this, as the code immediately
calls RepoModel().grant_user_permission, which resolved the user anyway.
This change just moves the database lookup a bit earlier.
Fixing this revealed the matching test case to be broken, so it has been
fixed as well.
Down the road, we should eliminate Kallithea's bizarre practice of
passing around usernames and user IDs, in favor of passing actual User
objects. That'll get rid of mistakes like these, as well as repeated
needless database lookups.
aa17c7a1b8a5 aa17c7a1b8a5 8d065db04909 8d065db04909 8d065db04909 aa17c7a1b8a5 5ae8e644aa88 8d065db04909 03bbd33bc084 03bbd33bc084 5ae8e644aa88 5ae8e644aa88 8d065db04909 aa17c7a1b8a5 5ae8e644aa88 8d065db04909 aa17c7a1b8a5 8d065db04909 5ae8e644aa88 8d065db04909 aa17c7a1b8a5 8d065db04909 8d065db04909 5ae8e644aa88 8d065db04909 aa17c7a1b8a5 8d065db04909 8d065db04909 | .. _locking:
==================
Repository locking
==================
Kallithea has a *repository locking* feature, disabled by default. When
enabled, every initial clone and every pull gives users (with write permission)
the exclusive right to do a push.
When repository locking is enabled, repositories get a ``locked`` flag.
The hg/git commands ``hg/git clone``, ``hg/git pull``,
and ``hg/git push`` influence this state:
- A ``clone`` or ``pull`` action locks the target repository
if the user has write/admin permissions on this repository.
- Kallithea will remember the user who locked the repository so only this
specific user can unlock the repo by performing a ``push``
command.
- Every other command on a locked repository from this user and every command
from any other user will result in an HTTP return code 423 (Locked).
Additionally, the HTTP error will mention the user that locked the repository
(e.g., “repository <repo> locked by user <user>”).
Each repository can be manually unlocked by an administrator from the
repository settings menu.
|