384091 - possible to stop background checking livemarks for updates (until you restart the browser)

Status: RESOLVED WONTFIX

(Firefox :: Bookmarks & History, defect)

Description

[(not reading, please use [e-mail] instead)]

possible to stop checking livemarks for updates (until you restart the browser)

this firefox 2 bug comes from the work Christine did for bug #357236.

The bug is in nsLivemarkLoadListener::OnStopRequest()


The intent of that code is:

if the expires time of the feed page (in the cache) is in the future, further than the default time to reload the feed, then use cache expiration (as otherwise, we'll just hit the cache.)

But the current Firefox 2 code does this:

272             if (nowsec >= expiresTime) {
273               expiresTime -= nowsec;
274               if (ttl < (PRInt32) expiresTime)
275                 ttl = (PRInt32) expiresTime;
276             }

We should be doing:

272             if (expiresTime >= nowsec) {

With the current code, you can end up with a huge ttl, which would result in us not checking the feed for updates (until you restart the browser).

I couldn't find a known bug in bugzilla about this, but one might exist.

This one liner might be good for fx 2.0.0.x.

Note, the trunk doesn't use this code anymore so it would be branch only.

Comment 1

[(not reading, please use [e-mail] instead)]

Attachment: patch for branch

Comment 2

[(not reading, please use [e-mail] instead)]

note, christine and I saw this with http://rss.cnn.com/rss/cnn_latest.rss

See also bug #383538

Comment 3

[(not reading, please use [e-mail] instead)]

note, the trunk (Firefox 3) had this same bug lurking in LLL_onStopRequest(), but that has been fixed by Christine in her patch for bug #357236

Comment 4

[Daniel Veditz [:dveditz]]

-> Seth since he added the patch (hot-potato theory)

Not a blocker, but if you get reviews and ask for approval we'll consider it.

Comment 5

[(not reading, please use [e-mail] instead)]

I think it would be doable to reproduce the bug, and verify the patch fixes the problem, but I'm not actively working on this as it doesn't block the 2.0.0.x release (and it is fixed on the trunk.)

Comment 6

[Marco Bonardo [:mak] (Away Apr 25 - May 5)]

not going to take further fixes on 2.x, feel free to reopen if you can reproduce on 3.x or current trunk