just wondering if anyone else in Galaxy land was having this problem, I emailed tophat.cufflinks@ to let them know
We have been experiencing segmentation faults with cufflinks 1.0.3. Our platform is CentOS 5.5, and we compiled from source using gcc 4.1.2 (Red Hat 4.1.2-48).
The segmentation fault occured when we passed in a GTF annotation file with the -G option. Using gdb, I tracked it down to line 218 in bundles.cpp:
string rs = (rna_seq) ? rna_seq:"";
It appeared that in our case rna_seq was uninitialized and did not point to a valid C string.
char* rna_seq; if (loadSeqs && faseq) { rna_seq = rna.getSpliced(faseq, false, &seqlen); }
...
if (loadSeqs) { string rs = (rna_seq) ? rna_seq:""; ... }
As you can see, it would be possible for the first if-block to not execute, but the second one to still execute (loadSeqs = true, faseq = false). I tried making the following change to the declaration of rna_seq to initialize it to NULL:
char* rna_seq = 0;
this stopped the segmentation fault we were seeing
-- Glen L. Beane Senior Software Engineer The Jackson Laboratory (207) 288-6153
galaxy-dev@lists.galaxyproject.org