Expire page data in memory when no longer needed
Categories
(Toolkit :: Places, enhancement, P3)
Tracking
()
Tracking | Status | |
---|---|---|
firefox100 | --- | fixed |
People
(Reporter: mossop, Assigned: mossop)
Details
Attachments
(1 file)
Assignee | ||
Comment 1•3 years ago
|
||
Currently we cache page data in memory until the browser is closed but this is
too long. We know there could be a couple of consumers of page data, snapshots
and AVM being two so we want to keep any discovered data in memory until we
know that no-one is interested anymore.
This adds a cache and a way for an "actor" (ugh!) to register interest in a url.
As long as there is an actor interested in a url any data for that url will be
cached in memory.
The idea is that when we start tracking a new interaction we start caching any
data for that url. When interactions have been flushed to disk and we've made
any decision about snapshotting we allow the data to expire.
We also by default keep data in the cache until the browser it came from is
destroyed.
Later the AVM can keep a page's data alive until it no longer exists in the
river.
Updated•3 years ago
|
Updated•3 years ago
|
Comment 3•3 years ago
|
||
bugherder |
Description
•