|Some FF bugs seem never to get fixed|
| 12:43 am on Jan 28, 2010 (gmt 0)|
I'm sticking to 3.5.2 due to the now infamous scrollbar/pause bug in releases thereafter.
To anyone whose copy of Firefox freezes for 1-2 seconds every 10-20 seconds, just revert back to the older version.
One of the problems with freeware, even if it is open-source, is that programmers often refuse to accept that bugs even exist. Firefox has the biggest and most predictable memory leak that I have ever seen, and all the Mozilla team have done is to offer two recommendations - restart it if necessary and don't switch off virtual memory. This should be one of the easiest bugs ever to find but it has existed since version 2.0 or thereabouts (maybe earlier).
I could happily switch to Opera if they would only implement Ctrl-click to open a page in a new tab - I am genuinely baffled as to why Opera haven't done this!
Yes, I do know that Shift-click works instead but with smallish hands on a laptop, that extra reach makes a big difference - plus it's what I'm used to.
[edited by: tedster at 2:58 am (utc) on Jan. 31, 2010]
[edit reason] moved from another location [/edit]
| 1:53 am on Jan 28, 2010 (gmt 0)|
I don't find OSS developers to be any worse than anyone else for refusing to accept that bugs even exist. MSFT is a case in point. Just look at how bad MSIE's rendering engine was for so long and how MSFT refused to fix it. Because of MSIE's broken box model, web development has at times been a real PITA for years.
The best way to get things fixed with Firefox is to report bugs through https://bugzilla.mozilla.org/ If you create an account on that site you can also track your bugs, vote on existing bugs and provide feedback. How quickly a bug gets fixed depends upon its nature. There was a tooltip bug that resulted in long tooltips getting cropped, that took like six years to get fixed, but last summer my Firefox theme uncovered a bug that caused Firefox to crash under certain circumstances. Because it was considered a security issue, that bug was fixed very quickly.
Security issues, browser crashes, etc. are high priority issues and get fixed quickly. Cosmetic issues like the tooltips issue can languish until someone decides to adopt the bug and fix it.
The real key with any bug for any browser is to make sure to document it as completely as possible and report it through proper channels.
| 2:16 am on Jan 28, 2010 (gmt 0)|
I have yet to discover a bug that hasn't been reported and discussed at length. Also, right now, I can't think of a single bug that's ever been fixed, but maybe that's because it's 2:10 am and I'm about to go to bed.
Granted, Microsoft can be very slow to fix bugs, but Mozilla is pretty hopeless too. Many people have commented that they should put more effort into fixing what they have instead of adding new features and I concur wholeheartedly. The problem is, I guess, fixing bugs just isn't sexy.
| 8:23 pm on Jan 29, 2010 (gmt 0)|
i had the same problem but if u give it a bit it will stablize or atleast it did on mine but also have u speed up ur firefox yet ? if u know the trick
| 10:18 pm on Jan 29, 2010 (gmt 0)|
|I can't think of a single bug that's ever been fixed, |
I find most bugs I have reported and/or tracked via bugzilla do get fixed in time.
There have been a few that I really wanted to get fixed that didn't but, in general all the major ones did.
| 10:36 pm on Jan 29, 2010 (gmt 0)|
Have you seen how Moz Dev deals with some of those bug reports? Or at least used to? I got too frustrated to continue to see if they still do this. But...
They would spew venom at the reporter of bugs and often rail them for their poor grammar and other ad hominem type attacks without addressing the actual bug being reported... then they close the incident report as if insulting the guy has resolved the issue.
| 11:28 pm on Jan 29, 2010 (gmt 0)|
I didn't realise it was that bad!
I once reported a bug to Opera, explaining it in some detail and giving them a simple example page of HTML that demonstrated it. To their credit, they replied, but denied that the bug could possibly exist because Opera ignored some attribute or other - completely nuts!
Ultimately, Mozilla will have to start taking bug reports seriously, if they don't people will dump Firefox - I have very nearly done so several times and would switch to Opera now but for the Ctrl-click issue I mentioned above.
| 1:34 am on Jan 30, 2010 (gmt 0)|
|Ultimately, Mozilla will have to start taking bug reports seriously, if they don't people will dump Firefox - I have very nearly done so several times and would switch to Opera now but for the Ctrl-click issue I mentioned above. |
Just because you say they don't take bug reports seriously doesn't mean its so.
From first hand experience, I know that if a bug is well documented such that Mozilla developers can replicate the bug and thus figure out what is going on and the bug is of a serious nature (e.g. causes Firefox to crash), they will get the bug fixed very quickly.
Like I said earlier when Firefox 3.5 came out last summer there was an issue with the theme I maintain crashing Firefox under very specific circumstances. Even the most screwed up theme should never cause Firefox to crash and when this issue got reported to Bugzilla it was immediately flagged as a security issue and it was patched within a matter of weeks.
I am not a Firefox developer, I only maintain a Firefox theme. There are times I am frustrated with Firefox development from a theme developer standpoint, and there are a few "bugs" in Firefox that cause headaches with theme development. HOWEVER, overall from an end user standpoint, Mozilla developers do a very good job of getting issues resolved in comparison to other browsers.
All browsers have bugs that languish and don't get fixed in reasonable manner. I have even documented one very blatant rendering bug in Opera since v7.23 which has never been fixed. Every year or so when a major new version of Opera is in the works, I go back to the Opera forums and bump my thread on the issue in hopes that maybe finally someone will fix the issue.
How about this issue in Internet Explorer: If you have a page that validates to W3C HTML and CSS specifications, but the underlying text files are save in UTF-8 such that there is a BOF character in the file, IE will flip over into quirks mode rather than standards compliance mode because the DOCTYPE tag doesn't start at the very first character in the file.
| 1:38 pm on Jan 30, 2010 (gmt 0)|
|Just because you say they don't take bug reports seriously doesn't mean its so. |
Very true, but it's not just my view.
Here are three big fat, juicy, reproducible bugs that haven't been fixed despite being present from months to years.
1) Memory leak - just keep Firefox open for a few hours of browsing and watch that memory get eaten (reported and widely discussed and present since about version 2.0 I think).
2) Frequent pauses (for 1-2 seconds every 10-20 seconds) introduced in 3.5.3 I believe (as reported by another user and widely discussed). Not quite as reproducible as the memory leak but I have experienced it on multiple computers and operating systems. You can watch this happen with the task mananger. On my dual-core machine cpu usage spikes at 50% : on a single cpu it would spike at about 100%. If they maintain a change-log, this ought to be real easy to fix.
3) Internet freeze under Vista/ZoneAlarm - they blame ZoneAlarm but other browsers work fine with it. This bug requires a full system reset. It's interesting they don't blame Vista - such an assignment of blame would be equally valid, but I guess Microsoft has a bigger legal dept.
I haven't tried 3.6 yet so maybe one or more of these bugs have been fixed but I doubt it.
| 4:20 pm on Jan 30, 2010 (gmt 0)|
Memory leaks are evil issues. My experience is that almost all software has them. ZoneAlarm is worst than most. With a browser, they are especially evil to fix because of plugins, extensions, etc. all can cause memory leaks that appear to be the blame of the browser. This isn't saying Firefox doesn't have memory leaks, its just that they are probably the hardest issue to eliminate and I don't find Firefox any worse than any other browser.
I almost never reboot my computer (WinXP Pro) and over time performance degrades to a point I'm forced to reboot. To mitigate performance issues I always keep Windows Task Manager opened and I sort processes by their virtual memory size to see if an application is getting too greedy with virtual memory. When an app does get too greedy I shut it down then reopen it.
Firefox does tend to be one of my biggest user of memory, but I also keep between 20-30 tabs open. What is interesting to watch is how different webpages can impact memory usage. For instance, yesterday I noticed that Firefox was using over 500mb of memory. I was about to shut down Firefox, but first navigated my Google AdSense tab away from the AdSense Ad Review Center page back to the main Reports page. When I did this, Firefox's virtual memory usage fell by over 200mb. The cause? Lots of Flash objects on the Ad Review Center. Was the memory issue a Firefox issue? No, not really, the problem was an Adobe Flash issue and memory greedy Flash ads. The thing is, Firefox is taking the blame.
I find in general that Flash objects are the biggest memory issue because they are so frequently badly designed. The best way to resolve the memory Flash object issues is to use the Flashblock extension so that Flash objects don't run unless I want them to. In the case of the AdSense Ad Review Center, I had activated several flash ads to see what they were promoting. As a result they started chewing up my memory. As soon as I navigated away from that page, those objects freed up the memory they were using.
Is it fair to blame the Flash memory issue on Firefox when the problem really lays with Adobe Flash and badly written Flash objects? No it isn't. Adobe should be taking the lead to resolve this issue.
Even with Opera, which has a reputation of being very easy on memory, I can quickly drive it to a point of consuming too much memory with only one or two tabs open. All I need to do is open Opera's Developer tools and then reload a page a whole bunch of times over a few hours while testing design changes to a website. After this, Opera's memory usage makes Firefox's memory usage look down right reasonable. If I leave Opera open overnight with its developer tools also open, by morning Opera will be on the verge of crashing and it will frequently crash when I then try to shut it down.
In regards to Firefox 3.6, so far I have found that it is a lot easier on memory than FF3.5 was and it is faster in general.
| 6:55 pm on Jan 30, 2010 (gmt 0)|
Flash is certainly memory-hungry - but I do have flashblock installed. I haven't done proper tests, but I don't think flash is to blame for the leak.
All memory allocation/deallocation should unltimately pass through a handful of functions - debugging should be straightforward. Either the bug is in the memory management code, or it is caused by an imbalance in allocation/deallocation requests. Either way, basic debugging should find the culprit pretty quickly. As for plugins, etc. - these can be disabled and so can be quickly eliminated.
| 6:56 pm on Jan 30, 2010 (gmt 0)|
I'm a little sensitive about this term, because it's often thrown at a problem when support can't figure out actually what's wrong. :-) (Not saying this is what you are doing here.) Do you have docs on this actually being a memory leak, or is it just a memory hog? Completely agree on the latter, it's always been this way with FF, not sure on the former.
The Wiki [en.wikipedia.org] describes my sentiment best, and is why I ask:
|A memory leak has symptoms similar to a number of other problems (see below) and generally can only be diagnosed by a programmer with access to the program source code; however, many people refer to any unwanted increase in memory usage as a memory leak, even if this is not strictly accurate. |
| 6:59 pm on Jan 30, 2010 (gmt 0)|
I believe Mozilla have accepted this as a leak and offered two pieces of advice...
1) Restart Firefox periodically.
2) Don't disable virtual memory in Windows.
| 7:06 pm on Jan 30, 2010 (gmt 0)|
|I believe Mozilla have accepted this as a leak and offered two pieces of advice... |
1) Restart Firefox periodically.
2) Don't disable virtual memory in Windows.
This should be standard practice with all desktop applications.