James Bonfield

It dawned on me we can do a poor mans threading here using calmd | mpileup. Eg: Orig ---- ``` $ time bcftools mpileup --threads 2 --fasta-ref $HREF -Ou ~/scratch/data/9827_2#49.1m.bam...

Thanks to an office discussion (with @dkj) this brings ideas for the multi-sample bcftools too. Compare (fake 10x 1 sample, but go with it...): ``` time bcftools mpileup --threads 2...

That's probably my fault giving incorrect advice to Thomas - I didn't realise there was even an mmap hFILE already, although is there a reason why it's hidden away in...

It may be best to use "samtools view -c" to count records. Otherwise you're largely just benchmarking how long it takes to print up SAM records (which isn't something htslib...

I'm starting to look at this again now. I prefer @jmarshall's implementation using the 'm' field in the mode string as it's cleaner to use and doesn't require juggling of...

Starting from @jmarshall's branch, I extended this to replace mFILE.c with hfiles instead. It needs more testing and is a work in progress, but this is the code so far:...

Most commands don't support globbing. It's typically your shell that does it for you. Eg try `echo bgzip *.vcf` and you'll see that _.vcf gets expanded up and passed as...

Just to be clear, are you calling this "cat" as file1.bed.gz, file2.bed.gz, file3.bed.gz etc are internally sorted with the first item in file2 coming after the last in file1, and...

It's an interesting one. For SAM and VCF of course there are already dedicated merge tools. We could do another for BED, but tabix is totally generic. It has configurations...

If we require it to be indexed first, then it becomes an easier proposal and no additional collation order information is needed. (See my dawning realisation in the "Edit:" comment...