summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* Fix up set_head set_description and set_readme to thread committerDaniel Silverstone2015-11-051-3/+3
|
* Fix up calls to set_ownerDaniel Silverstone2015-11-052-3/+3
|
* Ensure uses of repo:save_admin() propagate author and committerDaniel Silverstone2015-11-052-10/+14
|
* Support committing to adminrefs with author/committer differentDaniel Silverstone2015-11-051-7/+11
|
* Thread the old user context through cleanlyDaniel Silverstone2015-11-051-1/+2
|
* Support passing the committer for user and group and sshkey commandsDaniel Silverstone2015-11-052-3/+4
|
* Support author and committerDaniel Silverstone2015-11-051-7/+12
|
* Support dots in users and groupsDaniel Silverstone2015-11-052-4/+4
|
* Support renaming the user calling user renameDaniel Silverstone2015-11-051-0/+4
|
* Fix message on group renameDaniel Silverstone2015-11-051-1/+1
|
* Add support for auto_user_* and auto_group_*Daniel Silverstone2015-11-051-1/+36
|
* Change the http interface to use cURLDaniel Silverstone2015-11-052-49/+44
|
* Ensure the archive plugin is installed (and named reasonably)Daniel Silverstone2015-11-012-1/+1
|
* yarns: Add test for git-upload-archive commandRichard Maw2015-10-302-1/+13
| | | | This is used by git-archive to get the server to generate a tarball.
* tests: add gitarchive subcommand to GTT and yarn implementationRichard Maw2015-10-302-0/+11
|
* plugins: Don't check for the file typeRichard Maw2015-10-301-1/+1
| | | | | | | dirp:iterate()'s fileinfo may not have d_type initialised, do the check could fail even if it was a valid plugin file. I think the filename check is sufficient.
* git-archive command as pluginGavin Lee2015-10-301-0/+33
|
* yarns: Check cgit hide=1 set when project.archivedRichard Maw2015-10-302-1/+18
|
* yarns: Test behaviour when no commands givenRichard Maw2015-10-301-0/+13
|
* yarnlib: Allow empty "WHEN foo bar runs" commandsRichard Maw2015-10-301-2/+2
| | | | | Rather than specialising a 0-ary version, it's less effort to make the existing implementation flexible enough.
* Fix empty command bugRichard Ipsum2015-10-304-8/+25
| | | | | | | | | This fixes a bug that causes gitano to crash when it's given an empty command list Opening and closing of the log should be done in the application, so it can log before checking for authorization, log after, and clean up at the end
* Set hide=1 in cgitrc when project.archivedDaniel Silverstone2015-10-271-0/+4
|
* Merge archiverepo workDaniel Silverstone2015-10-133-10/+45
|
* Change gfind to gmatch so script works with both 5.1 and 5.2Daniel Silverstone2015-02-031-1/+1
|
* Correct false assumption, require luxio explicitlyDaniel Silverstone2015-02-032-0/+2
|
* Support indicating the object type (and tagged object and type) on updateDaniel Silverstone2014-08-251-1/+26
|
* Clean up copy routines a little to improve repository copy functionalityDaniel Silverstone2014-08-241-7/+33
|
* Update gitano-test-tool because git fsck doesn't always accept a pathDaniel Silverstone2014-08-231-1/+1
|
* Support paths.skel in gitano-setup to change skeleton pathbaserock/pedroalvarez/trove-ansible3Daniel Silverstone2014-05-211-1/+4
|
* Add support for cached ssh keys for testing. Shaves 20% off test runtimeDaniel Silverstone2014-05-177-6/+106
|
* Add tests for configDaniel Silverstone2014-05-171-0/+91
|
* Add support for server-side repo checkingDaniel Silverstone2014-05-172-0/+14
|
* Short-circuiting hooks if repository is accessed directlyGavin Lee2014-05-164-0/+19
| | | | Signed-off-by: Vanush "Misha" Paturyan <misha@cs.nuim.ie>
* Merge remote-tracking branch 'ri/richardipsum/refactor'Daniel Silverstone2014-05-165-158/+38
|\
| * Make runcmd output "Unable to continue"Richard Ipsum2014-05-041-0/+1
| | | | | | | | | | At the moment the test suite requires that certain errors produce "Unable to continue" on stderr
| * Make gitano-command cgi use runcmdRichard Ipsum2014-05-041-21/+1
| |
| * Make gitano-auth use is_authorized and runcmdRichard Ipsum2014-05-041-133/+9
| |
| * Make gitano-smart-http pass repo_root to is_authorizedRichard Ipsum2014-05-041-1/+2
| |
| * Make gitano-command pass repo_root to is_authorizedRichard Ipsum2014-05-041-1/+1
| |
| * Add repo_root parameter to is_authorizedRichard Ipsum2014-05-041-2/+1
| |
| * Move runcmd into utilRichard Ipsum2014-05-041-0/+23
|/
* Ensure we pass the repo through otherwise HTTP commands might not workDaniel Silverstone2014-04-092-4/+5
|
* UPDATE-HOOK: Extra audit message informationDaniel Silverstone2014-04-041-1/+8
| | | | | | | Increase the audit log message data to include the update action type, and the old and new SHA. Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Support rsync directly to subdirectoriesDaniel Silverstone2014-04-041-2/+3
|
* Cache expansionsDaniel Silverstone2014-03-221-0/+8
|
* Allow tests to run even when not inside gitano-allDaniel Silverstone2014-03-141-1/+9
|
* RSYNC: Update rsync plugin to fix hiccoughsDaniel Silverstone2014-03-121-5/+2
| | | | | | | | This removes the rsync path reporting since I'm now confident it works, also we fix a bug in the lace running so that it tests against the right repo and we hide the rsync command since it's not a normal user command. Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Merge in rsync work from CodethinkDaniel Silverstone2014-03-1117-39/+291
|\ | | | | | | | | | | | | This adds support for Codethink's Trove server's rsync command and also adds a generic plugin interface and repository detection support. Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
| * Support for rsync in Trovesbaserock/danielsilverstone/S10634-rsync-supportDaniel Silverstone2014-03-112-1/+107
| | | | | | | | | | | | | | | | | | | | | | | | In order to support ingesting of binary artifacts, we are proposing the use of `git-fat` which is a content filter which uses rsync to store the binary artifacts. That requires an access controlled binary artifact repository, so we take advantage of Gitano's ACLs and associate an optional rsync repository with every git repository on the server. By placing it inside the git repository, all of gitano's destroy and graveyard behaviour automatically works with the new rsync content. Signed-off-by: Daniel Silverstone <daniel.silverstone@codethink.co.uk>
| * Use new configurable repository detectionbaserock/danielsilverstone/S10633-repo-detectionDaniel Silverstone2014-03-113-31/+12
| | | | | | | | | | | | | | | | | | Each of these callsites independently detected repositories before. Now they use the configurable repository detection which means (a) they are unified in their behaviour and (b) they can take advantage of new commands which might not have repositories represented in the same way. Signed-off-by: Daniel Silverstone <daniel.silverstone@codethink.co.uk>