Segmentation fault in ConvertTileset / calloc

#0 _int_malloc (av=0x7ffff7479e80, bytes=5177344) at malloc.c:4709
#1 0x00007ffff718d642 in __libc_calloc (n=<value optimized out>, elem_size=<value optimized out>) at malloc.c:4066

Install debug symbols for your libc. I need to know arguments passed to calloc (<value optimized out> is not enought).

> Install debug symbols for your libc.
It is installed indeed. Otherwise we wouldn't see the #0 _int_malloc staff.

> I need to know arguments passed to calloc (<value optimized out> is not enought).
Ok. I did this:
$ diff startool.cpp{.ORIG,}
3101c3101
< image = (unsigned char *)calloc(1, height * width);
---
> printf(">>>DEBUG: calloc(1,%d)<<< ",height * width); fflush(stdout); image = (unsigned char *)calloc(1, height * width);

And get this:
> extracted: tileset\Jungle.cv5 (1806, 82056 bytes)
> extracted: tileset\Jungle.vf4 (1808, 161472 bytes)
> >>>DEBUG: calloc(1,5177344)<<< Segmentation fault
$

Please note (again), I got a non-standard SC distribution. (the mpq is definitely modified externally).

> DEBUG: calloc(1, 5177344)<<< Segmentation fault

~5 Mb on the stack... it is a lot.
we should use a std::vector<unsigned char> image;
image.resize(height * width)

Nope, calloc allocate memory in heap (as malloc). So this is correct usage. Seems like some memory leaks or not enought free memory...

> Seems like some memory leaks or not enought free memory...
I don't think so. Glibc should _return_ a error code in such cases, not SEGFAULT.
This looks like a memory corruption.

Can you try to start binary valgrind? Maybe it show something more...

I confounded with alloca...

What is your type of modifications in mpq ?

[Expired for Stargus because there has been no activity for 60 days.]

Ok, this is written in linux malloc manpage:

Crashes in malloc(), calloc(), realloc(), or free() are almost always related to heap corruption, such as overflowing an allocated chunk or freeing the same pointer twice.

Ildar, what OS do you have? Also please run startool under valgrind. It will print all memory corruptions.

And can somebody other reproduce this bug?

[Expired for Stargus because there has been no activity for 60 days.]