On Sun, 2017-08-06 at 11:51 +0200, Yury V. Zaytsev wrote:

> We have a patch that you could try, it would be great if you could tell 
> whether this fixes the problem for you:
> 
> http://midnight-commander.org/attachment/ticket/3846/3846_mcview_hook.patch

Thanks, that fixes the problem.

> The initial report lacked a critical piece of information, and 
> specifically that you have pressed "Ctrl+x q" during panel initialization 
> and this caused the crash, and not merely observed a "random crash on 
> startup".

Yeah, initially I thought it was a random crash. It was only after some
more investigation that I discovered it was actually reproducible.

> This is the reason why I originally assumed that either the pointer wasn't 
> set to NULL after getting freed, or something else overwrote it to point 
> to unitialized memory.

I see.

> Yes, this log has been really useful and now it is more or less clear 
> what's going on.

Good to hear, thanks for the info.

> I think that you are misinterpreting your valgrind report.

Yeah, that is clear from your description of the problem.

> If you run mc under valgrind ... which we routinely do
...
> we do use static analysis tools like cppcheck
...
> we have been using Coverity for the last 5 years or so

Good to hear :)

> as fuzzing is concerned ...

I see, thanks for the info.

-- 
bye,
pabs

https://wiki.debian.org/PaulWise

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to