John Marshall

From the commit message: > as well as if the start (or mate start) is off the end. Um… not sure anyone has made a case for allowing this?

HTSlib is a C library so it doesn't really matter what C++ has :smile: For some reason I thought enumeration constants were in a separate name space (so did not...

I am updating PR #721 at present and am planning to add new entries as `bzip2_compression`, `fastq_format`, etc rather than plain `bzip2` and `fastq`. Does this seem like a sufficient...

> Would it be better to use an hts prefix, resulting in hts_fmt_sam, hts_compress_gzip etc.? As usual, it's a tradeoff. It depends how people feel about the hideous ugliness, etc…

Since 2017, both SAM and VCF have gained clarity around what constitutes a legal contig name (and VCF's prohibition of colons has been removed). The SAM spec also has some...

PR #1441 fixed the minor typo, but the filename inconsistency remains.

For `conda create` and `conda install` see also #10491 — the user-facing documentation does not really explain what can appear in a `package_spec` on these command lines.

To activate HTSlib's workaround for macOS not having _lzma.h_, you need to run the _configure_ script. Does this vt build process run HTSlib's _configure_?

Here is a better _hts_utils.h_ patch that was applied two years ago: fb0288be9346cbf0f822dc9104c3a063ed0c521e.

When you look at how that function was reaching into htslib internals, that's unsurprising. Fortunately sanity prevailed two months after that release and can also form the basis of the...