BMO/Meetings/2013-09-10

mcote has been converting the API docs into a REST-specific ReST doc. Not sure as to the exact value since the REST interface (e.g. URLs) is better documented than he suspected. Still appears to be value in having one unified doc for the REST API, though, in particular mixing in docs from API extensions. Will investigate a POD-to-RST converter to do some of the boring conversion work.

Confirmed fubar as our main point of contact for IT issues related to BMO. Upcoming work that will require IT support includes - memcached - change notification system - ElasticSearch ETL

Doing a push early this week (Wednesday night ET/PT) to clear out a backlog while ashish was on PTO. That will leave more time in the week to fix things in case there is a problem in any of the pushed commits.

New BMO newsgroup has been approve: mozilla.tools.bmo. Should be going live soon.

Persona difficulties vis-à-vis disabled accounts. With normal accounts, when an employee leaves but wants to remain involved, we switch their email address to their personal address but leave the password the same. With persona, there is no password (it's handled by the persona service), so there's no way to ensure that the given personal address is the same as the employee's old address, that we're not accidentally giving access to the account to someone else. Solution is to have their former manager vouch for them. Official docs on the process have been updated.

The Get Help page, linked from the first large icon on the front page, is pretty useless. Asking lizzard for some help in creating better content; will take it down until we can make something useful.

Display of group membership flags (on the right, beside attachments) is ugly. Patch was created some time ago to fix it but stalled pending discussion of bigger UI changes (e.g. BugzillaPretty contest). Decided to apply it upstream and back port it regardless, since it helps now and can be change later if we revamp the UI.

Posting for new BMO position is up.

ekyle has noticed some inaccuracies (mostly in the history) in the existing ElasticSearch ETL, also in the new Python-based ETL. Fixing obvious problems, but bigger problems will wait pending more important features, namely, restricting ETL to public bugs only, and creating a scrubber to remove bugs, attachments, and comments that move from being public to private.