In the course of looking at the no-store fix by darin (last minute branch checkin), I happened to notice the following glitch in the server logs. (Note: this doesn't depend on the no-store fix in any way; that's just how I happened to notice this). If you have not previously visited this URL (i.e., you don't have this content cached), then when you go to the URL: http://cowtools.mcom.com/perf/content/base/www.yahoo.com/index.html then you get this output in the access_log (edited for clarity and 78 columns) [22:57:14] "GET .../www.yahoo.com/index.html HTTP/1.1" 200 14340 [22:57:15] "GET .../www.yahoo.com/res/m5v2.gif HTTP/1.1" 200 5776 [22:57:15] "GET .../www.yahoo.com/res/hp_auctions1.gif HTTP/1.1" 200 4706 [22:57:15] "GET .../www.yahoo.com/res/m5v2.gif HTTP/1.1" 200 5776 [22:57:15] "GET .../www.yahoo.com/res/hp_auctions1.gif HTTP/1.1" 200 4706 So, there are two fetches of the two images in that page. Hmmm. I haven't tried this on Mac and Linux (noticed it on win32 and thought I would just file it before I forgot).

John this looks like Bug 91738 - " mozilla geting file twice"

And both this bug and bug 91738 are likely what's causing bug 76565...

This could be a dupe, but you're not reloading, are you?

No, this is not on a reload. To be specific, delete Cache, cookies.txt, and history.dat, start/shutdown mozilla a couple of times, and then start mozilla. Enter http://cowtools.mcom.com/perf/content/base/www.yahoo.com/index.html in the urlbar and hit Enter. Mozilla will GET index.html, and then each of the two images _twice_ (200 OK on all GET's).

*** Bug 101165 has been marked as a duplicate of this bug. ***

Hey, isn't bug 103432 a dup of this as well?

From my bug report, I see this on http://news.bbc.co.uk/ Thought I'd mention the URL here, since this bug's URL is internal to NSCP. Happening on Linux for me, so marking OS All.

fixed

Hey jrgm@netscape.com, would you mind verifying this? Thanks, Terri

actually not fixed. Still fetches each image of those two images twice (and now doesn't fire onload [bug 120041]). current trunk at 4pm/Jan15 win2k. Steps to reproduce and results (from above): > To be specific, delete Cache, cookies.txt, and history.dat, start/shutdown > mozilla a couple of times, and then start mozilla. > > Enter http://cowtools.mcom.com/perf/content/base/www.yahoo.com/index.html > in the urlbar and hit Enter. Mozilla will GET index.html, and then each of > the two images _twice_ (200 OK on all GET's).

Similar thing on http://news.bbc.co.uk/ which I used to have a problem with too. Page now never finished loading (it looks done, but status bar still says it's retrieving stuff), which is like your onlond issue, plus I notice in network trace it requests things like "v.gif?" (with the question mark) though view source didn't show a v.gif or v.gif? in the source. When I went to investigate further, I got caught up filing a bunch of save bugs.

*** Bug 104827 has been marked as a duplicate of this bug. ***

moving this to future. The problem described here is due to the page getting reframed. This ends up causing images to get destroyed and refetched. I'm cc'ing waterson and marc incase they have any ideas.

In bug 133217 Mozilla (build 2002040503) is loading every image (for the rollover) again from network.

It also happens when the image url is typed in directly: for http://www.google.com/images/logo.gif on a windows 2000 PC with Mozilla build 2002051006, this image is downloaded twice. A second request is sent from the browser right after the response to the first one has been fully received. I repeated it with other image URLS too.

Who is willing to take this - deserves really some attention.

Actually, for the specific case for which this bug was filed, the yahoo example, this is no longer observable in a current trunk build. WFM.

regarding comment #15 we should file a new bug if this problem is still there.

Comment 15 is (or was) a known problem with a bug filed somewhere.