James Bonfield

Maybe :-) It feels more like an improved version of Unix `comm`. Anyway in the short term you should probably use bedtools merge: https://bedtools.readthedocs.io/en/latest/content/tools/merge.html

Are we making a rod for our own back by having curl as a plugin? Agreed we should finalise the json tokeniser, but not being able to use it in...

@anderskaplan I fully agree that abort() is a bad thing for a library. There are a few places it may be deemed acceptable (coding errors rather than data/system driven errors),...

We can't easily remove or rename fields without breaking existing code, so this is an htslib 2.0 (no roadmap as yet) feature. However would accessor functions help clarify this? We...

Release 1.10 is a major step forward in terms of symbol tables. All internal functions are no longer exported from the library. However this is mainly a FYI as the...

Possibly not. Most of the timeouts are just because it's possible to generate bit streams of very highly compressed data, such that decoding takes a lot of memory and a...

There are newer versions of that file. The lzmaspec.cpp contained in the git repo I linked above is dated 2015, and a quick google search finds some newer still. Eg...

Thanks for the response. I thought it was interesting though to see differences in htslib's BAI usage and htsjdk's BAI usage. I'm sure BAM users would like faster random access...

It's also incorrect to say "unexpected newline" because this error will be triggered on a header that has zero newlines (and also doesn't start with an \@). Not that we...

I don't understand why a coordinate sorted file that misses the coordinate sorted tag would make a difference, at least not from an htslib perspective. There is some code to...