Closed
Bug 309982
Opened 19 years ago
Closed 19 years ago
Site never finishes loading, nothing displayed
Categories
(Camino Graveyard :: Page Layout, defect, P1)
Tracking
(Not tracked)
RESOLVED
INVALID
Camino1.0
People
(Reporter: sfraser_bugs, Assigned: mikepinkerton)
References
()
Details
Someone on the feedback list reported that this site: http://dna.annonces.net/petites-annonces-informatique.php is very slow to load in camino. It works fine in Firefox and Safari. Related to bug 309873?
| Reporter | ||
Updated•19 years ago
|
Severity: normal → critical
Flags: camino1.0+
Priority: -- → P1
Target Milestone: --- → Camino1.0
Well, it's at least JS related again. Removing/commenting out the middle script, <script language="Javascript" src="oas.js"></script> gets the page to load. That looks to me like it's mostly an ad script, but....
Comment 2•19 years ago
|
||
It loads slowly in both Camino and Firefox on the branch. Maybe it doesn't even load at all, I don't know, I got impatient and gave up. No trouble with either on the trunk. -> core.js_engine?
One more wierd bit: I turned off ad-blocking to test that hypothesis. The page was loading much more quickly and I was asked about accepting a cookie for memorix.sdv.fr. I said "deny" and the page continued to load. Thereafter, even after emptying cache and turning ad-blocking back on, the page loads fairly quickly (memorix.sdv.fr still set to "deny"). A wierd interaction between ad-blocking and cookies and JavaScript?
Comment 4•19 years ago
|
||
Look at oas.js, there's nothing to finger core JS in it:
<!--
///configuration
OAS_url = 'http://memorix.sdv.fr/RealMedia/ads/';
OAS_sitepage = 'www.dna.fr/paweb';
OAS_listpos = 'Top';
OAS_query = '';
OAS_target = '_top';
//end of configuration
OAS_version = 10;
OAS_rn = '001234567890'; OAS_rns = '1234567890';
OAS_rn = new String (Math.random());OAS_rn.substring (2,11);
function OAS_NORMAL(pos) {
document.write('<A HREF="' + OAS_url+'click_nx.ads/' + OAS_sitepage +
'/1' + OAS_rns + '@' + OAS_listpos + '!' + pos + '?' + OAS_query + '" TARGET=' +
OAS_target + '>');
document.write('<IMG SRC="' + OAS_url + 'adstream_nx.ads/' +
OAS_sitepage + '/1' + OAS_rns + '@' + OAS_listpos + '!' + pos + '?' + OAS_query
+ '" BORDER=0></a>');
}
//-->
<!--
OAS_version = 11;
if(navigator.userAgent.indexOf('Mozilla/3')!=-1 ||
navigator.userAgent.indexOf('Mozilla/4.0 WebTV')!=-1)
OAS_version = 10;
if(OAS_version >= 11)
document.write('<SCR' + 'IPT LANGUAGE=JavaScript1.1 SRC="' + OAS_url +
'adstream_mjx.ads/' + OAS_sitepage + '/1' + OAS_rns + '@' + OAS_listpos + '?' +
OAS_query + '"></scr' + 'ipt>');//-->
<!--
document.write('');
function OAS_AD(pos) {
/*if (OAS_version >= 11)
OAS_RICH(pos);
else*/
OAS_NORMAL(pos);
}
//-->
With up-to-date trunk or branch, all the comment hiding hacks (which serve no
purpose in a .js file, only in inline script content in HTML) are handled correctly.
The only thing that might affect perf is DOM, not JS: the document.write.
Cc'ing mrbkap.
Someone should profile, or do poor-man's profiling in a debugger (just break or
ctrl-C frequently and note stacks). This is not going to go anywhere without
more diagnosis.
/be| Reporter | ||
Comment 6•19 years ago
|
||
Hrm, once I've managed to get the site to load once (Stopping the first load, then doing a reload), it then loads fine, even after killing the cache and relaunching Camino. I wonder if this is some kind of DNS problem?
| Assignee | ||
Comment 7•19 years ago
|
||
regarless, shouldn't someone else own this bug? Is it ok on the 1.7branch?
(In reply to comment #6) > I wonder if this is some kind of DNS problem? I just did this series of tests: Cm 0.8.4 fails Cm 1.0a1+ fails Fx 1.5b2+ fails Fx 1.0.7 loads Fx 1.5b2+ loads Cm 0.8.4 loads Cm 1.0a1+ loads This certainly lends credence to the DNS theory. (Re: comment 3, I still have "memorix.sdv.fr" set to "deny" cookies in my Camino profile; Fx is set to accept cookies from the originating site only, which should be an implicit "deny" for "memorix.sdv.fr", too.)
Comment 11•19 years ago
|
||
Simon, are we going to call this a DNS problem or do you have a better idea of what this might be?
I just ran traceroute (in Network Utility) on the site, and there are a number of pauses even early on, but once it gets into the sdv.fr domain (the host of dna.annonces.net), things completely die. So it really seems like some sort of DNS/routing problem. OTOH, while traceroute was plugging away, never getting to the site, Safari loaded it right up, and curl didn't have any problem. And finally, as I'm done with these tests, Camino loaded the page. Are we (and traceroute/Network Utility) interfacing with Mac OS X DNS stuff differently from curl and Safari? This reminds me of the issue that was a big deal for a while in one of the versions of Mac OS X this spring/summer; Safari always timed out and failed to load sites but Camino et al. never had any problems....
Torben, you're in Europe, right? Maybe there are fewer hops or you have a "friendlier" DNS? Or maybe it's 10.2.8 :-) I tried this site in Camino with my hotel's provider last week and nothing but the tab title loaded after about five minutes (unfortunately, I didn't have time to run a full suite of tests).
Comment 15•19 years ago
|
||
(In reply to comment #14) > Torben, you're in Europe, right? That's right. > Maybe there are fewer hops or you have a > "friendlier" DNS? Or maybe it's 10.2.8 :-) Well, I'll try with 10.3.9 tonight or tomorrow, but it certainly looks like a DNS problem.
Comment 16•19 years ago
|
||
Yeah, this is weird. I just tried to open in Camino, it failed (or never stopped loading), then refreshed and it worked fine. I'm going to say this is DNS issues. Does anyone disagree with that now?
| Assignee | ||
Comment 17•19 years ago
|
||
just tried this again with 1.0b1 and it never loads. where are we with this?
Comment 18•19 years ago
|
||
Since this is reproducible on Windows in both FF 1.5 and IE, I'm going to say that it's DNS issues and, thus, is INVALID. That would explain various people having problems and others not having problems. It seems to switch fairly randomly and I doubt it's something we can control.
Status: NEW → RESOLVED
Closed: 19 years ago
Resolution: --- → INVALID
You need to log in
before you can comment on or make changes to this bug.
Description
•