Closed Bug 569486 Opened 14 years ago Closed 12 years ago

Pollers should respect etags and last-modified headers (if they exist)

Categories

(Release Engineering :: General, enhancement, P5)

enhancement

Tracking

(Not tracked)

RESOLVED WONTFIX

People

(Reporter: christian, Unassigned)

References

Details

(Whiteboard: [automation])

The pollers in http://hg.mozilla.org/build/buildbotcustom/file/8b1380cbf092/changes should respect etags and last-modified headers if they exist.

Note that hg's rss feeds don't currently send this info (bug 566673).
Depends on: 566673
Priority: -- → P5
Whiteboard: [automation]
Is the idea here to use a HEAD request to avoid fetching the whole log? I'm not sure how etag helps otherwise.
I don't see how this helps - we poll with &fromchange=XXXXX so usually it's an empty reply.
Maybe that wasn't the case in the past. Sounds like this is WONTFIX now, given that.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → WONTFIX
How they help on your end is if it's faster to say "if status == "304" return" (roughly, I don't know what Twisted actually puts in status) instead of having minidom parse the shell of a feed and do an empty for loop and reverse an empty list.

But typically you do etags on the client side to be a good citizen, and the benefit is on the other end: if it's an expensive db operation to look at whether there's anything newer than XXXXX, but a nearly free filesystem operation to see whether a file named etags/fromchange=XXXXX-123456 exists and just return a 304 if it does, then yay etags. From a few minutes glance at the patch, it doesn't look like it's a huge load reducer on hg.m.o's side, but djc would probably have a better idea than me about that.
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.