Fortunately, I don't think anyone has to go through that timesome 
process of adding the profile dir file by file, because I found the 
reason for the problem: corrupted cache!

Thanks for your time though, read my reply on Travis Crump for more info.

/ David

Jason Johnston wrote:
> David Tenser wrote:
> <snip/>
> 
>> Thanks for the tip. I tried to create a new profile, and, very 
>> surprising for me, it does make a big difference! It's still pretty 
>> slow when initially loading the page (the address previously 
>> mentioned), but when alt-tabbing between windows, it repaints the page 
>> instantly now.
>>
>> How come?!
>>
>> What exactly is stored in the profile? Shouldn't the page renderer 
>> (Necko?) render a page the same way, regardless of what email settings 
>> that I have for instance? How can the profile affect the performance 
>> on the screen painting of websites? And an even harder question to 
>> ask, why is it just this page, and not every page?
>>
>> Of course, I'm sure there are other pages that has the same problem on 
>> this machine, but still... why?
>>
>> / David
>>
> 
> 
> Yes, this kind of thing shouldn't happen, but that's why we're all 
> running Mozilla test builds, isn't it, to help find these wacky things 
> and report them as bugs?
> 
> Ideally the profile shouldn't affect rendering speed, but occasionally a 
> bad nightly build will slightly corrupt some file or add a bogus pref or 
> something else that causes a conflict in some obscure part of the 
> rendering code.  Last month I found a bogus line in my prefs.js that 
> caused external javascript files to not load! :-0 (It's since been fixed.)
> 
> If you have the time and the patience, what would be very helpful is to 
> try copying your old profile file-by-file into the new one, testing Moz 
> after each step, and see at what point the bad behavior returns.  If you 
> can narrow it down to a specific file, you can file a bug report and 
> attach the file to it (provided it doesn't contain any sensitive 
> personal information ;-)).  Then the developers will have a testcase to 
> go by, which often is the most important step.
> 
> If you don't have time to do this, then I'm afraid whatever bug it is 
> will probably go unfixed, unless someone else happens to have the same 
> problem and does the legwork.
> 
> --J
> 



Reply via email to