This is a process that listens to activity on upstream repos on pagure and github via fedmsg, and syncs new issues there to a Jira instance elsewhere.
Configuration is in fedmsg.d/.
You can maintain a mapping there that allows you to match one upstream repo (say, 'pungi' on pagure) to a downstream project/component pair in Jira (say, 'COMPOSE', and 'Pungi').
On startup, if the initialize option is set to True in the
fedmsg.d/ config, then all open issues from all upstream repos will be
scraped and added to Jira if they are absent.
If the testing option is set to True, then the script will perform a
"dry run" and not actually add any new issues to Jira.
Each project is accompanied by an 'updates' array as seen below:
'Demo_project': {'project': 'PROJECT', 'component': 'COMP',
'updates': [...], 'owner': 'project_owner_username'},
The following can be added to the updates array to specify what to sync with downstream JIRA issues:
'comments' :: Sync comments and comment edits
{'tags': {'overwrite': True/False}} :: Sync tags, do/don't overwrite downstream tags
{'fixVersion'; {'overwrite': True/False}} :: Sync fixVersion (downstream milestone),
do/don't overwrite downstream fixVersion
'assignee' :: Sync assignee (for Github only the first assignee will sync)
'description' :: Sync description
'title' :: Sync title
{'transition': True/'CUSTOM_TRANSITION'} :: Sync status (open/closed), Sync only status/
Attempt to transition JIRA ticket to
CUSTOM_TRANSITION on upstream closure
Note: Overwrite set to True will ensure that upstream issue fields will clear downstream issue fields, overwrite set to False will never delete downstream issue fields only append.
The optional owner field can be used to specify a username that should be used if the program cannot find a matching downstream user to assignee an issue too. The owner field will also be used to alert users if duplicate downstream issues exist.
To set up the mailer you need to set the following environmental variables:
To run it yourself, you're going to need to put in some jira authentication
options into fedmsg.d/ including the JIRA url and your username and
password. Furthermore, you need the fedmsg configuration from the Fedora
rpm. Get that with dnf install fedmsg.
You will also need to get a github API token and store it in the
config as github_token to avoid being rate-limited.
If you run into SSL errors, you may need to tell python-requests to use the system CA cert:
$ export REQUESTS_CA_BUNDLE=/etc/pki/tls/certs/ca-bundle.crt
Here are some things that this program does not do:
If there is interest in having those things, please file an RFE or comment on an existing one here and we'll see about prioritizing it.
We have decent test coverage.
Run the tests with:
$ sudo dnf install detox $ detox