From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.4) Gecko/20010913 BuildID: 2001091323 If a meta http-equiv line is malformed, mozilla can be made to loop and load the page an infinite number of times. mozilla should not do anything in this situation. (Netscape 4.7 has the same behaviour.) Reproducible: Always Steps to Reproduce: Malform a refresh line in the head of a page: <META HTTP-EQUIV="refresh" CONTENT="URL=http://www.mozilla.org"> instead of <META HTTP-EQUIV="refresh" CONTENT="0; URL=http://www.mozilla.org"> to cause an infinite loop. Actual Results: Browser loops and reloads the source page (before displaying it!) an infinite number of times instead of going to the specified destination. Expected Results: Either nothing, or go to destination while complaining about syntax. Here's a minimal example: <CODE> <HTML> <HEAD> <TITLE>infinite loop</TITLE> <META HTTP-EQUIV="refresh" CONTENT="URL=http://www.mozilla.org"> <!-- notice not 0; URL=... --> <BODY> This demonstrates how a broken refresh command leads to an infinite loop. </BODY> </HTML> </CODE>
jag, this is a dup of one of yours, right?
related to bug 62742?
I can't find the bug this is a dupe of, so I'll just take it for now. Btw, NS4 doesn't deal with this all too nicely either.
Assignee: neeti → jaggernaut
Status: UNCONFIRMED → NEW
Component: Networking → Browser-General
Ever confirmed: true
Reducing severity, targetting for 0.9.9. And yes, it's related to but not the same as bug 62742.
Severity: critical → normal
Target Milestone: --- → mozilla0.9.9
I just went to this site: http://www.92ndsty.org/ It refreshes infinitely and never finishes loading. The meta tag didn't look obviously invalid, but had a very large number of seconds in it; maybe it overflowed Mozilla. Can a preference be added to disable meta refreshes?
The line in question: <meta HTTP-EQUIV="refresh" CONTENT="432000;URL=http://www.92ndsty.org"> Tested on my linux and w2k boxen, WFM.
I have this line on a web page: <META HTTP-EQUIV="Refresh" CONTENT="1800"> And it works correctly most of the times. Sometimes though it's like having CONTENT="0" and it doesn't get fixed even after quitting/restarting mozilla. I think this happened on the latest nightlys..
The "infinite zero-time reload" problem on meta refresh tags seems to be a sporadic, hard to predict thing that comes and goes -- I reported it earlier with the http://www.92ndsty.org/ site, but others said it "WFM"... while it kept doing the problem for me through several builds. But now I tried that site and it works for *me*... but meanwhile somebody else has reported the same problem happening on a different site. There must be some hidden factor that's causing this to happen, and it won't be solved until that factor is identified.
I'm experiencing something similar using Build 2002020208 on Windows 98 (4.10.1998). When I load http://sports.yahoo.com/ncaab/scores/20020202/sbfkaf.html the page will halfway load and then start a refresh. The refresh tag looks like: <meta http-equiv=refresh content=1200>. I know that there's no standard for the meta refresh tag, but it seems that a content of 1200 should mean to refresh after 20 minutes.
Running Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.8) Gecko/20020204 the infinite refresh (without displaying) seems to occur on this page when the number of seconds between refreshes is greater than aproximately 1 hour (3600): http://www1.burtonsys.com/mozilla_bug.html
Bug 116361 deals with infinite reloading on what seem to be valid meta refresh tags. So, I guess, if this one isn't considered a dupe, it should be regarded as dealing more specifically with the case of infinitely reloading on *invalid* meta refresh tags, a related but distinct issue.
nsbeta1+ per ADT triage team
Keywords: nsbeta1 → nsbeta1+
On content="URL=...": NS4 and current Mozilla builds do the same thing: reload the current page infinitely. IE6 does nothing at all. The expected behaviour is to load the url imediately, but do realize that this will only work in Mozilla, if we decide to fix this, instead of following IE6's lead and doing nothing. On content="1200": As far as I can tell, my code is working fine there. When we don't find anything after the 1200, we assume the current uri as the one to reload. We could be parsing the 1200 incorrectly, or somehow timers overflow on your system, if it's still happening, please file a new bug on that.
Target Milestone: mozilla0.9.9 → mozilla1.0
nsbeta1- per ADT
Duping. *** This bug has been marked as a duplicate of 170021 ***
Status: NEW → RESOLVED
Last Resolved: 16 years ago
Resolution: --- → DUPLICATE
Target Milestone: mozilla1.1alpha → ---
The dupe is in docshell, but I'm sending this to http for routing to the correct component. Components have changed a lot since I routed my last meta tag bug....
Component: Browser-General → Networking: HTTP
QA Contact: benc → httpqa
You need to log in before you can comment on or make changes to this bug.