Closed Bug 44282 Opened 25 years ago Closed 23 years ago

bookmark scheduled update fails to trigger

Categories

(SeaMonkey :: Bookmarks & History, defect, P3)

x86
Windows 2000
defect

Tracking

(Not tracked)

RESOLVED DUPLICATE of bug 107268

People

(Reporter: danm.moz, Assigned: bugs)

Details

- Bookmark some page that's served up by HTTP - Edit the bookmark's properties to check for changes. Tell it to open the URL in a new window when a change is noticed. - Edit the HTTP page - Wait for Mozilla to notice. Keep waiting. I'm not quite sure what the pattern is. Multiple edits on the source don't seem to help. I believe Mozilla begins to notice only after editing the bookmark's schedule property a second time, turning on all the notification methods. (Also, quitting and relaunching seems to make it not notice that the page has been changed since last it checked while running.) - Repeat with a different page. (set up two bookmarked pages to simultaneously check for changes.) Generally it seems only the new page is checked. Sometimes the first page is also checked for a while. It's hard to be certain, because the act of making the new bookmark to check seems to turn off all checking until the bug mentioned above has been somehow circumvented.
Here's a description of what's going on. To tell whether a remote HTTP URL has changed or not, we track what the remote server indicates as the "Last-Modified" value. When you first mark a bookmark with a schedule, we have no idea what the "Last-Modified" value is at that point, so the URL starts out marked in an "unknown" state. When the schedule fires for the first time after that, we get the "Last-Modified" value from the server and merely remember it, changing the URL to a "current" state. From that point on, whenever the schedule fires, we continue getting the "Last-Modified" value from the server and comparing it to the last well-known value we have.
OK, that makes sense. But it doesn't quite cover all my worries. I didn't explain this well, but part of my testing above was setting the update period to one minute and then changing the source HTML file once every minute or so for several minutes. This wasn't guaranteed to make the file reload.
Yeah, that's what I usually do... set the duration to 1 minute... makes for quick/easy testing. If you want to test it some more, I'll stop by and show you how to turn on some stronger logging, and we can try and get a feel for what's happening with it.
Marking this REMIND to see if we can come up with a faster resolution of detecting changes of remote URLs. (I have some ideas.)
Status: NEW → RESOLVED
Closed: 25 years ago
Resolution: --- → REMIND
REMIND is deprecated.
Status: RESOLVED → REOPENED
Resolution: REMIND → ---
->default
Assignee: rjc → ben
Status: REOPENED → NEW
This is the same issue as bug 107268, but that bug is ASSIGNED and has been getting more attention than this one. *** This bug has been marked as a duplicate of 107268 ***
Status: NEW → RESOLVED
Closed: 25 years ago23 years ago
Resolution: --- → DUPLICATE
An update: This is still an issue. A new bug has been openen, bug 207814, to track these problems. Nobody seems to work on bug 107268 anymore, the owner wasn't even able to update the status fields. So, if anyone knows anything about the bookmark schedule system, please take a look at that one.
Product: Browser → Seamonkey
You need to log in before you can comment on or make changes to this bug.