Good morning.

For the last years I have been experimenting from time to time a very difficult to reproduce crash. I've seen it at eBay and at Amazon, for example, but I have never reported it because I was unable to reproduce it exactly.

Now I have the way to reproduce it, so here are the instructions to follow (please note that following this instruction may crash K-Meleon, so save your work before trying):

1) Open Amazon's web and type on the Amazon's search bar "tubin balalaika", or directly go to this address:
http://www.amazon.com/Eduard-Tubin-Symphony-Balalaika-Orchestra/dp/B0000016D6/ref=sr_1_1/002-3389737-3244036?ie=UTF8&s=music&qid=1180088514&sr=8-1

2) Click at the first search result "Edward Tubin: Symphony No. 1 in C minor ..." but asking to open it as a background layer (middle clicking on my PC does opens layers on the background).

3) While the foreground is still the Amazon's search results, click on the background newly opened layer. It will completely crash K-Meleon.

If you right click on the search link and save the url address of the crashing layer (without opening it), you restart K-Meleon and open directly the crashing link it will sometimes show properly without crashing.

I'm using official spanish 1.1 over WXP SP2.

As it involves an Amazon page (which changes often) it's IMHO important trying to reproduce it now that we can and before it's too late. Can somebody please try it?

I hope someone can reproduce it.


PS. Well, curiously, if you open the previously reported link in this same layer it will open perfectly, but if you open it as a background layer and then click over it, it will also crash. The link goes to the Amazon's search results, the crashing link was suppossed to be the Tubin's CD page, not the search results page which is previous.



Edited 1 time(s). Last edit at 06/07/2007 07:46PM by enaitzjga.

I guess, this would be the correct link to Amazon's search results?

Well, WFM, with both de-DE and es-ES user agent.

KKO, your link opens perfectly, but, again, if I open the first result shown in your link as a background layer, when I click on the new background layer it crashes.

Doesn't it happen to you too?

As I've said: Works For Me

And please, let somebody confirm your 'bug' next time before posting it to the BTS.

We've lots of bug reports like this in the BTS that nobody could ever reproduce.

What is the problematic link in your search page linked; details, listen to samples, more about this product, ...?

K-Meleon in Spanish

The main crashing link is the first result of any of Amazon's search pages posted above, but only if you open them as a background layer and then click over it.

Any of those pages wich makes my K-Meleon crash when opened in background layer and then focused on it can perfectly be seen if you just left-click over them and see them in the normal way.

Well, I have been making a few calls this morning to find up what was wrong.

I have 2 friends with modern systems (P4 and latest AMD) that uses 95% the same software than me, so the only difference between their systems and mine is the hardware. They've been unable to reproduce the error, so I deduce it's not related to a third party software which I'm using.

I also know a guy who has the exact same computer as myself, including sound and video cards and motherboard BIOS revision (Dell Pentium III 933 Mhz, 512 Mb RAM, ATI Radeon 9000 Pro and early Sound Blaster Live), but he uses different software. He had the same problem as myself, he was the very first person I see able to reproduce the error.

So looks like a case closed. Must be some kind of weird incompatibility between the hardware and something in the crashing link, maybe some driver issue, who knows?

What's important for the Project is that this is most probably not a bug, so IMHO we can close this bug.

Quote
kko
And please, let somebody confirm your 'bug' next time before posting it to the BTS.

We've lots of bug reports like this in the BTS that nobody could ever reproduce.

Dunno, in my opinion, getting a 100% reproducible bug is reason enough to file a bug report, so people can confirm the bug or mark it "can't reproduce".

The way I see it, the difference between Enaitz's report and a lot of the other unconfirmed bug reports is that his report has a few easy steps to follow, and it only takes a few seconds to try without having to do anything fancy. Plus, he replies to questions. The result is a bug that almost gets closed in one day (I'm just waiting if Dorian has any comments).

Thanks rmn, I'm only trying to help creating the best browser ever.

As it was a long time searched bug (by me), when I found the way to get a 100% success while reprocing it (again, by me), I got excited and suppossed everyone would be able to reproduce it. Now I see that was not right.

The reason for posting it at the BTS was to get the more possible people helping on solving that, if you see it, at the forums not always you can find the most valuable help while trying to test bugs. This thread has been viewed by 70 people, and the only people who has tested the easy to follow steps and reported the results has been KKO and Desga2, people who usually checks the BTS. Even with a so urgent looking name to the thread as the one I choosed.

Don't know if people at the forums are to shy or what, but looks like if you want serious check for a possible bug you have to go directly to the BTS.

In my earlier post I forgot to mention that I've also tested this, and it didn't crash either.

Could very well be a driver problem... I've heard that bad hardware/driver can cause very odd bugs. In my old computer, IE used to consistently freeze whenever I scroll a page--didn't happen in other apps--and every hardware person I asked seemed to blame it on the graphic card driver.

Yes, that's what I now supposed it should be the problem.

I'm using latest drivers of everything, I may should try changing them, or trying replacing some parts of the hardware. It's a pitty 'cause my old Dell Pentium III works fine for my needs, and I was not thinking in replacing it.

Anyway I found an AMD Sempron 2500 based full computer (minus the monitor) at eBay for 130 EUR and I bought it for my a friend who asked me to find a cheap computer for him. I may have to start considering something similar for myself ...

Good evening.

I've had the oportunity to test a beta version of the future 1.5 K-Meleon, and it seems that the Bug 912 is fixed.

I'm going to report everything I've done, step by step, please excuse me if this goes too long, but I want to make an useful report.

First I've reproduced again the 912 bug on the 1.1 es-ES, just to be sure it was still there and happening. It was happening as always.

Then I've simply extracted the 1.5 beta to the folder of my current spanish and slightly modified 1.1 K-Meleon, overwritting everything and not deleting the profiles folder (so the files at the defaults folder of the 1.5 beta should have not been used). When I started it for the first time (and later everytime) it appeared a pop up tittled "Macro parser error" and telling some lines of detected errors. The bug 912 kept unfixed and happening everytime.

Then I completely uninstalled my spanish K-Meleon, downloaded and installed official en-US 1.1, and before starting it for the first time, extracted the 1.5 beta over it (overwritting when needed). I did not install Flash Player. When running it there was no "Macro Parser Error". I have been completely unable to reproduce the 912 bug, there was no way at all to reproduce it.

Lastly I've installed Flash Player and tested again the 912 bug. It was still impossible to reproduce.

So, looks like I was not crazy and Bug 912 is history.

Thanks Dorian!

enaitz,

FWIW, I tried to reproduce your crash on my W98se laptop using KM 1.0.2, updated to Seamonkey 1.1.2 and did not get the crash.

W98se
PIII-600 Intel Coppermine
320 Megs RAM
Compaq BIOS
ESS Technology ES1970 Canyon3D / ES1978 Maestro-2E Sound Card PCI


(Sorry, don't read the forums as much as I used to and almost never go to the bug reporting system.)

N

Thanks for trying. This bug loves me so much that won't go with anybody else but me

BTW after some more experiments I'm close to find the key for it: If I overwrite 1.1 with all the files from the 1.5 b14 the bug is fixed, if I overwrite 1.1 with all the files from 1.5 b14 except the new menus.cfg the bug is still there. But if I overwrite the 1.1 only with the menus.cfg from the 1.5 b14 the bug is not fixed.

So it must be the menus.cfg in relation with other file.

enaitz,

That is good detective work. I played with 1.1 for awhile, but pending the time to figure it out, I'm back to 1.0.2 for now.

One of these days I'll find the time.

<wry grin>

N

I detected this "Macro Parser Error" in 1.1 when i was make this patch:
http://kmeleonbrowser.org/forum/read.php?3,72300

When i tried run KM 1.1 with my 'groups.kmm' patch file in 'Macro' subfolder in profile folder, "Macro Parser Error" is showed in lines 18 and 19.
I don't know why it is, can be for a coincidence with the same file in Macros folder (out of profile).

Greetings.