core dump in 0.5.5 and 0.5.8

[RichardCorbett]

Hi.
When I try and merge some bisulfite novoalign bams I get a core dump. I've tried this on Centos5 and Centos 6 with the 2 versions above and I consistently get this error soon after the run begins.

xhost11 /projects/rcorbettprj2/novoAlignBioQC/A27712 $

/gsc/software/linux-x86_64/sambamba-0.5.5/sambamba_v0.5.5 merge -p A27712.bam ./*sorted.bam;
Segmentation fault

any ideas?

[RichardCorbett]

Hmmm. I was able to run some other samples successfully. There is quite likely a problem with one of the bams I am trying to merge.

[lomereiter]

Does any of your files contain long reads (~10K bases)? They weren't handled properly until recently (#182).

[RichardCorbett]

Hi Artem,
Nope, just 125bp illumina reads. I suspect one of the bams were truncated (we've had some network errors lately). I'll try re-running the alignment and see if that fixes things. It might take a few days (novoalign after all).

[RichardCorbett]

Hi Artem,
As expected, having re-run the alignments allows me to successfully to merge/dupmark the bams. There must have been something wrong with my original bams.