mamarjan/gff3-pltools

gff3-ffetch segfaults

Closed this issue · 2 comments

pjotrp commented

wget ftp://ftp.wormbase.org/pub/wormbase/species/c_elegans/sequence/genomic/c_
elegans.WS232.genomic.fa.gz
wget ftp://ftp.wormbase.org/pub/wormbase/species/c_elegans/sequence/genomic/c_
elegans.WS232.genomic.fa.gz
gzip -d *.gz

gff3-ffetch mRNA c_elegans.WS232.annotations.gff3 c_elegans.WS232.genomic.fa
Segmentation fault

lomereiter commented

(one of Pjotr's links must be ftp://ftp.wormbase.org/pub/wormbase/species/c_elegans/gff/c_elegans.WS232.annotations.gff3.gz)

I think I've found where the issue comes from.
First, Pjotr did specify files in the wrong order, and gff3-ffetch didn't even notice whereas it clearly states in usage information that FASTA should come first. So what happens is the following:

  • gff3-ffetch tries to get FASTA from the end of first file
  • there's no FASTA at the end of the file, and get_fasta_range() returns null
  • however, if we look back at bio/gff3/conv/fasta.d, we see that the result is not checked for null, and property all is accessed. Here's where the segfault is likely to occur. (I haven't checked, though.)
mamarjan commented

Thanks Artem, I'll fix that right away.