summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* testing: clone/push over http in http test modeRichard Maw2016-12-181-7/+86
| | | | | | | | | This sets up password auth for the gitano-smart-http wrapper so we can finally test commands over http. We wouldn't normally bother with auth during test mode, but lighttpd does not pass REMOTE_USER through to the CGIs unless we have authenticated.
* testing: In http mode add passwords for usersRichard Maw2016-12-181-0/+13
| | | | | | | These use the username as the password as a hack. We don't have a fully password-self-hosted setup, so this is a bit of a hack with a bootstrapping problem. This would be fixed by having gitano-setup accept passwords too.
* testing: Record stdout and stderr from cloneRichard Maw2016-12-181-1/+2
|
* Fix updating htpasswd filesRichard Maw2016-12-181-2/+2
| | | | | | | | | | | | | | | | This would previously break because if it's updating htpasswd it passes an empty argument rather than no argument. Rather than stop it passing the empty argument this patch instead passes -i flags unconditionally and merges the flags argument. This is to make it use the batch-input code-path which doesn't attempt to use a tty for its input, since it caused issues in the test suite which fakes the server. This requires changing it to only pass the password as input once.
* smart-http: Only generate Status 500 on non-exit of http-backendRichard Maw2016-12-181-3/+3
| | | | | If smart-http succeeds or fails it will produce its own message, we only need to produce our own code if we don't even get that far.
* testing: Rename gtt {push,pull}viassh to {push,pull}Richard Maw2016-12-182-7/+7
| | | | | This is going to support smart HTTP {push,pull} so it's incorrect to be named "via ssh".
* Remove executable bit from cgi inputsDaniel Silverstone2016-12-182-0/+0
|
* testing: Run lighttpd during http testsRichard Maw2016-12-183-1/+47
|
* Fix accidental duplicate FINALLYRichard Maw2016-12-181-1/+0
|
* testing: Add teardown step to all scenariosRichard Maw2016-12-1824-0/+68
|
* testing: Split out current ssh key testRichard Maw2016-12-182-19/+21
| | | | | | | This was disabled because when accessing Gitano over HTTP there is no current key to display or prevent you from deleting. By splitting it out we can add an ASSUMING to conditionally activate it.
* More basic tests, checking syntax and loadability of modulesDaniel Silverstone2016-12-181-2/+8
|
* json is experimental and unsupported and incompleteDaniel Silverstone2016-12-181-0/+4
|
* Rename rules-patches to admin-patches since that's what they areDaniel Silverstone2016-12-1810-15/+15
|
* Tweak ordering for gitano-setup and update manpageDaniel Silverstone2016-12-182-14/+37
|
* Update README a littleDaniel Silverstone2016-12-181-9/+11
|
* Remove obsolete CIA hook (CIA has been dead for a while now)Daniel Silverstone2016-12-181-45/+0
|
* Remove now entirely obsolete notes about how 'Legit' would workDaniel Silverstone2016-12-1812-568/+0
|
* More testing notesDaniel Silverstone2016-12-171-0/+35
|
* Ensure the tests actually run under HTTP modeDaniel Silverstone2016-12-165-40/+52
|
* Very basic git-annex-shell support, not checked repo copies yetDaniel Silverstone2016-12-132-1/+91
|
* Honour suppression of error messages from running comandsDaniel Silverstone2016-12-132-7/+11
|
* Support suppression of error messages from commandsDaniel Silverstone2016-12-131-2/+6
|
* Explain error codes when a command failsDaniel Silverstone2016-12-132-2/+2
|
* Ensure we propagate exit codesDaniel Silverstone2016-12-131-1/+2
|
* Fix braino in bypass user expansion nameDaniel Silverstone2016-12-131-1/+1
|
* Missing keyDaniel Silverstone2016-12-122-0/+28
|
* Merge remote-tracking branch 'benjamb/benbrown/broken-pattern'Daniel Silverstone2016-12-122-1/+13
|\
| * Update test to verify tag names with dashes are acceptedBen Brown2016-12-111-0/+12
| |
| * Fix broken match by moving dash to end of setBen Brown2016-12-111-1/+1
|/ | | | The '-' fails to represent a range between '.' and '_'.
* testing: Add smoke testing for help outputv0.2Richard Maw2016-12-032-0/+60
|
* Add the newkey instancesDaniel Silverstone2016-12-032-0/+28
|
* testing: Add yarns for whoami outputRichard Maw2016-12-031-0/+11
|
* testing: Test user details change and renameRichard Maw2016-12-031-2/+29
|
* testing: Add user creation/deletion yarnsRichard Maw2016-12-031-2/+27
|
* Set push.default to simple to quiet down git during tests.Daniel Silverstone2016-12-031-0/+1
|
* If you try to sshkey del the key you're using, you get told you can'tDaniel Silverstone2016-12-032-4/+6
|
* testing: Add tests for add/del/list ssh keysRichard Maw2016-12-032-0/+53
| | | | | This does not test that we can't delete our own key since that functionality was never implemented.
* Finalise lace context simple matchesDaniel Silverstone2016-12-031-0/+4
|
* No BANANA todayDaniel Silverstone2016-12-031-1/+0
|
* Ensure gitano-auth passes the keytag through and that ↵Daniel Silverstone2016-12-032-3/+2
| | | | gitano.auth.is_authorized uses it
* Remove a failure-ensues which shouldn't be thereDaniel Silverstone2016-12-031-1/+0
|
* Support whining for supple, and clean up a few bitsDaniel Silverstone2016-12-038-16/+56
|
* Ensure that lace warns if an ACL check is being bypassedDaniel Silverstone2016-12-033-0/+19
|
* Ensure that if the gitano-bypass user is invoked, Gitano whines loudlyDaniel Silverstone2016-12-033-0/+22
|
* Ensure config parsing fails if gitano-bypass is not viableDaniel Silverstone2016-12-032-0/+19
|
* Support the bypass key need in the test suiteDaniel Silverstone2016-12-032-2/+4
|
* Add the bypass key for the test suiteDaniel Silverstone2016-12-032-0/+28
|
* Support a gitano-bypass user in gitano-setupDaniel Silverstone2016-12-032-9/+28
|
* yarns: Amend rename text for new error messageRichard Maw2016-12-031-3/+1
| | | | | | | | | | | | | | | This was previously commented out because the given error message was that the ruleset denied action. This was because the only way reading was permitted was by repository ownership, and since the repository did not exist it could not have an owner. Since the rules skeleton was updated to allow read in the user's prefix it can have a more useful error message if the repository does not exist. The message produced when a repository is missing was also changed since the code was commented out, due to localisation changes, so as well as re-enabling these commands the error message has been updated to reflect its new wording.
View Lorry Controller Status