Headfi and safari = slow as *****!
May 3, 2012 at 4:08 PM Post #16 of 65
Reproducibility: How reproducible is this issue?
Steps to Reproduce: What steps should we take to see the problem occur?
Expected: What action did you expect to occur when this bug appeared?
Notes: Any additional information that will help us diagnose the issue
 
Browser/OS: It's sometimes necessary to reproduce the same bug on different browsers and operating systems to help narrow down the nature of the problem. Whenever possible, please include this info in your bug report. 
 
URL: This will be the URL where the error appears, or the URL where the test steps begin. Descriptions like "when I clicked from this page to that page" are helpful, but it's always best to copy/paste the exact URL in question.
 
If you can, upload a screenshot of unexpected behavior. If you're describing a detailed problem, or if you're unsure how to describe a bug, use Jing to send us a screenshot or a video of what you see on your screen. (http://www.techsmith.com/jing/) -- Note: this is not available for mobile devices.



 
1) extremely reproducible. Happens virtually 99.9% of the time (except for periods of time like the past several days, when it ceased happening at all, but has now resumed happening, since approximately 24 hours ago. The earlier posts in this thread document when it was happening, when it stopped, and when it resumed).
 
2) To make it occur, simply enable javascript in Safari and visit Head-fi forums. It's that simple. It ONLY seems to happen in Head-fi, not in other websites that use javascript. If I disable javascript, I can navigate Head-fi without Safari/OSX slowing down but of course there is some reduction in functionality. If I use Opera, it doesn't seem to slow down with javascript enabled, when visiting Head-fi, but Opera has some formatting issues when quoting people in replies, so it's a PITA to use (Opera 10.63, build 8450).
 
3) I expected Safari to load a Head-fi forum page, but instead it took several minutes to load the page, during which time the entire system (not just the browser) ground to a complete halt, with the hard-drive being furiously-accessed, as though some kind of 'swap-file' was struggling to keep up with demand.
 
4) additional information - nothing other than that it only seems to occur in Safari, and there is probably a major clue if you look at the timeline, in this thread,  of when it began, when it stopped, and when it resumed. If it is a piece of malicious code on my machine, then I wonder why it only seems to pose a problem with Head-fi, and not other javascript-heavy websites. I have looked for a checker/removal tool for malicious code but my older version of OSX does not seem to be supported (I welcome a link to something which will work with OSX 10.4.11). I still suspect the issue lies with something specific in the javascript code of Head-fi, though.
 
5) URL - it happens with ANY page on Head-fi forums - take your pick.
 
6) Screenshot - this isn't relevant for this particular bug - the system just slows to a halt, there's nothing visual on the screen that's out of the ordinary.
 
May 3, 2012 at 4:22 PM Post #17 of 65
I might be off-target about this, since I'm not a programmer, but, subjectively, it seems as though the computer is attempting to parse thousands upon thousands of lines of javascript code, or perhaps getting stuck in an (almost)-infinite loop. It can't be any coincidence that others are reporting similar troubles loading the site with Android (http://www.head-fi.org/t/605060/4-10-crashing-on-android-fixed-then-5-2-again) during about the same period of time I've experienced difficulties with Safari. I also access Head-fi with Android and although my phone (SG2 using stock browser) has remained capable of loading Head-fi pages OK, I did notice a change in the way the javascript adverts load, when the problems began, a few weeks ago. Whereas the Android browser used to load the forum pages fairly smoothly, and one could begin scrolling before it had completed loading all adverts, it now repeatedly jumps to the top of the page, multiple times, when loading each individual advert tile, making scrolling the page totally futile until all adverts have loaded. I understand that Head-fi exists because of advertising revenue, but IMHO it has become top-heavy with advertising content and the user browsing experience is beginning to suffer for it; the parable of the Golden Goose springs to mind...
 
May 3, 2012 at 4:43 PM Post #19 of 65
It also seems reminiscent of my old Windows 98 machine when I used to get 'memory stack overflow' messages. The HDD used to take a hammering then, with the system grinding to a halt, although in this particular instance, owing to the very poor memory handling in Win98, it used to culminate in a blue screen of death. Interesting, though, that the behaviour I described earlier, of the webpage being (seemingly) repeatedly re-drawn as each individual advert tile is loaded might perhaps lead to memory handling errors, and, perhaps, also to increased reliance upon the HDD to compensate through use of a 'swap-file'. Again, I'm just a layman, not a programmer, so I could be talking nonsense here - I accept that. I'm not speculating in order to look stupid (which I'm content to accept I may do); I'm speculating merely because I'm trying to describe how the current Head-fi/Safari-javascript experience seems, subjectively. It might provide a clue, however indirectly or tangentially, so it's worth risking looking stupid for.
 
May 3, 2012 at 5:05 PM Post #20 of 65
This just happened to me maybe about 45 minutes ago:
 
Was looking at one of my head-fi gallery pages. Page took forever to load, browser became unresponsive, and all my free RAM got used up. Eventually stuff did appear on the screen, and I noticed that all the ad boxes showed the Head-fi logo. Turns out every ad box was like a frame with a whole new Head-fi page loading inside. And maybe the effect was recursive? With no RAM left and Safari not responding again, I did not stick around to find out more.
 
This only happened to me once so far. Things seem back to normal now. Hope it helps with the debugging.
 
Safari 5.1.5, OS X 10.7.3
 
May 3, 2012 at 5:13 PM Post #21 of 65
That kind of thing might increase ad revenues, mightn't it? (You may call me a cynic! :wink: )
 
May 4, 2012 at 5:14 PM Post #27 of 65
At the risk of tempting fate, I'll dare to venture that it appears to have been fixed..
 

Users who are viewing this thread

Back
Top