Re: [darktable-devel] free(): invalid pointer

2012-10-29 Thread Ulrich Pegelow
Am 29.10.2012 21:40, schrieb johannes hanika: > sounds like broken installation/old plugins to me. valgrind ran > through clean (at least a couple of weeks ago when i last tried). > I get the same glibc messages here. And that's from a really fresh install with regular nuking the whole install dir

Re: [darktable-devel] free(): invalid pointer

2012-10-29 Thread johannes hanika
sounds like broken installation/old plugins to me. valgrind ran through clean (at least a couple of weeks ago when i last tried). j. On Tue, Oct 30, 2012 at 7:41 AM, Ulrich Pegelow wrote: > Feel free ;) > > Am 29.10.2012 19:29, schrieb Mikko Rasa: >> Uh, no, those messages have nothing to do wit

Re: [darktable-devel] free(): invalid pointer

2012-10-29 Thread Ulrich Pegelow
Feel free ;) Am 29.10.2012 19:29, schrieb Mikko Rasa: > Uh, no, those messages have nothing to do with memory leaks. "free(): > invalid pointer" means that free got a pointer that was already freed > earlier, or wasn't malloced at all. If detected, glibc will simply > ignore such pointers, bu

Re: [darktable-devel] free(): invalid pointer

2012-10-29 Thread Mikko Rasa
Uh, no, those messages have nothing to do with memory leaks. "free(): invalid pointer" means that free got a pointer that was already freed earlier, or wasn't malloced at all. If detected, glibc will simply ignore such pointers, but don't lull yourself into a false sense of security - the rea

Re: [darktable-devel] free(): invalid pointer

2012-10-29 Thread Jesper Pedersen
Roumano, On 10/29/2012 02:07 PM, Ulrich Pegelow wrote: > At some point we should have a look those warnings and debug them. > You can compile dt with Debug symbols, and run your scenarios with Memcheck: http://valgrind.org/info/tools.html#memcheck That should provide with the needed informati

Re: [darktable-devel] free(): invalid pointer

2012-10-29 Thread Ulrich Pegelow
Hi, not much to worry about. For some reason on your system environment variable MALLOC_CHECK_ is set to 1. This will cause glibc to spit out warning on all (potential) memory leaks in darktable or its libraries. Most user won't see this as they have MALLOC_CHECK_ not set. At some point we should

[darktable-devel] free(): invalid pointer

2012-10-29 Thread Roumano
Hi, Sorry Nop, still some "free() invalid pointer"when i went to darkroom mode (but can't Always reproduce) : ls -ald /opt/darktable/ ls: cannot access /opt/darktable/: No such file or directory sudo make install /opt/darktable/bin/darktable : *** glibc detected *** /opt/darktable/bin/darktabl