Massive Konqueror Regression

[David van Hoose]

Aaron J. Seigo wrote:
> On Saturday 13 August 2005 09:08, David van Hoose wrote:
> 
>>Aaron J. Seigo wrote:
>>
>>>there's something of a contradiction in this sentence ;) you can help
>>>development progress, though ...
>>
>>I'll provide information, but I cannot provide the source to the pages
>>in question.
> 
> 
> the source to the pages themselves usually isn't the best thing for the 
> developers anyways. what is best is a reduced, simplified bit of example 
> HTML/Javascript/CSS/whatever that shows the problem succinctly and directly. 
> this usually means save the page locally and then removing all the bits that 
> aren't related to the problem until you have a generic test case. this 
> sometimes takes a while because it requires hunting for what it triggering 
> the problem.
> 
> this is very valuable to the developers because it saves them from doing this 
> and allows them to quickly see what the problem is exactly and then go about 
> fixing it quickly.

But this is not always possible since some pages are several hundred 
lines long, and if you don't know what is broken, then you don't know 
what to put in the "simple test case." I've identified what it does and 
its effect, but I don't know where in Konqueror it is. I do know that it 
worked before and doesn't now.

>>To mention a really annoying KDE regression, every time I log out of
>>KDE, KDE Panel (kicker) segfaults.
> 
> 
> there is one more "crash on exit" bug left and hopefully i'll get it for 
> 3.4.3. it's a bug in kdelibs. the other 2 have been fixed.
> 
> this, btw, is where i find this list to be particularly humorous in name. so 
> many people on this list are attracted to it because it talks about "quality" 
> and everyone wants quality but then do the people actually install alphas and 
> betas? do they have a debug enabled install of KDE that they use for day to 
> day work? do they report backtraces and create test cases? do they triage bug 
> reports?
> 
> that's the most basic of things that one can do and it really helps when 
> people do it. the kicker crash on logout problems could have been identified 
> and fixed BEFORE 3.4.x if more end users had used the 3.4.x pre-releases for 
> day to day work with their normal configurations.

And why didn't the KDE developers find this bug from beta testing?

>>I've also experienced several 
>>deadlocks with the panel as well when configuring it from the Control
>>Center. It deadlocks and KDE must be forcibly restarted to correct it.
> 
> 
> this is almost certainly your X server .... there's not much kicker can do 
> itself to deadlock your entire session ...

Sorry, but this ONLY occurs with the new version of KDE so it is very 
much NOT an X problem.

Regards,
David van Hoose