Opened 13 years ago
Closed 11 years ago
#968 closed defect (fixed)
v.net.salesman problem with memory allocation
|Reported by:||marisn||Owned by:|
v.net.salesman does something wrong during memory allocation. Normaly it results into assertion in malloc call.
v.net.salesman: malloc.c:4633: _int_malloc: Assertion `victim->fd_nextsize->bk_nextsize == victim' failed.
Valgrind reports invalid write (will attach full valgrind output as an attachment).
Number of cities:  ==6589== Invalid write of size 8 ==6589== at 0x401FAF: main (main.c:209) ==6589== Address 0xd18cbc8 is 0 bytes after a block of size 40 alloc'd ==6589== at 0x4C25EE5: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==6589== by 0x529118F: G__malloc (alloc.c:41) ==6589== by 0x401F64: main (main.c:207)
GRASS 6.4 r41215 on ~AMD64 Gentoo. 6.5 is also affected.
I'm unable to reproduce with Spearfish. Will attach sample dataset.
Change History (5)
by , 13 years ago
comment:1 by , 13 years ago
Can't attach sample location. Download it from here: http://www.gisnet.lv/ggmlinux/salesman_crash.tar.gz
Command to reproduce issue:
v.net.salesman input=roads_and_cities output=my_route type=line,boundary alayer=1 nlayer=2 ccats=17,15,8,14,23,5
comment:2 by , 13 years ago
|CPU:||Unspecified → x86-64|
|Version:||unspecified → svn-releasebranch64|
comment:3 by , 12 years ago
|Milestone:||6.4.0 → 6.4.1|
Still an issue for 6.4 and 6.5 SVNs.
comment:4 by , 11 years ago
|Milestone:||6.4.1 → 6.4.3|
|Status:||new → closed|
The recent v.net.+ improvements have solved the problem. The sample command is now processed properly. Closing.
Valgrind output of v.net.salesman run