Open Bug 554190 Opened 15 years ago Updated 2 years ago

gloda indexer should prioritize indexing for real-time responsiveness over batch indexing

Categories

(MailNews Core :: Database, defect)

defect

Tracking

(Not tracked)

People

(Reporter: asuth, Unassigned)

Details

(Keywords: perf, Whiteboard: [gloda key])

Gloda should interrupt long-running batch jobs in favor of indexing newly received messages, messages the user has modified, and unindexed messages that the user's current actions suggests they are interested in. Gloda gained the ability in the 3.0 correctness push to interrupt jobs cleanly. It is now quite feasible to interrupt batch jobs when new event driven indexing events come in or other high priorities suggest we should do so.
Whiteboard: [gloda key]
Thanks for creating this. As a side remark, what would be useful to me would be for getMessageCollectionForHeaders to automatically trigger an indexing request with a high priority for the headers I gave it that are not indexed yet. This may return all the headers it knows of at the time of the request, and then call onItemsAdded when the missing headers are indexed. I don't know if that's what you were thinking of, just my 2 eurocents.
That sounds like a great idea, and indeed the query itself would already behave as you suggest once the indexing occurs.
should bug 534449 be marked a duplicate of this one ?
Blocks: 522908
No, that bug is a failure of MailNews to generate an event, not gloda to act on it in a timely fashion.
Removing myslef on all the bugs I'm cced on. Please NI me if you need something on MailNews Core bugs from me.
Keywords: perf
No longer blocks: 522908
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.