James Bonfield

> For this PR to proceed, it would be useful for the maintainers to comment on the direction: in particular, I am now leaning towards dropping the punctuation characters entirely...

What's the status of this, other that "draft"? I had a play again this morning and think there are still some open questions. Eg: - Why does the external struct...

As an attempt to fix the READ1/READ2 bug, my first stab was this: ``` @@ -5014,8 +5018,16 @@ static unsigned long sam_parse_flag1(const char *s, char **endptr) case 'R': if (strcmp(name,...

I'm trying to think of how practically it would be implemented to track how many soft-clipped reads overlap. For the right hand soft-clip, we could record when we're discarding a...

Thank you for raising this issue. I am not a VCF maintainer and admit that I found this confusing too when I looked into it. However I see this very...

Also, I note that "0 / 1 | 2" comment was there in the initial checkin of the initial VCF spec, so we have no more history to know where...

It does appear to be implicitly defined in BCF though as the spec indicates the | vs / affects the next value, not the previous one.

https://github.com/samtools/htslib/pull/976 was an attempt to fix this for CRAM, but I forget why we rejected it as an idea. (I think trying to represent whole genome alignments in BAM is...

That's likely a bug in bgzip as it ought to be using the magic number instead of filename. However I see similar login in tabix, which only works on (for...

Now the mem-io branch has been merged, is it possible to untangle this PR by removing those merges and rebasing on develop? I gave it a try, but it's a...