John Marshall

(See jmarshall/samtools@aefbf50cf18fef678beb2b1bf364cda68c2e174c for a draft of the samtools `--keep-tag`/`--remove-tag` functionality re-implemented via this API.)

I'll take complete silence as signalling approval of the basic approach :smile: The new iterator-oriented field deletion functions are now `bam_aux_remove()` and `bam_aux_remove2()`, as `…remove` is fairly widespread in HTSlib...

~~Don't refresh your memory of the code as it stands in this PR. Back in April, I got most of the way through refactoring it as discussed. I may have...

Oops, actually it is #1361 that has been substantially refactored from the current state of its PR. I have an in-progress commit adding `bam_aux_remove_if()` to the existing PR. I was...

(That force-push was just to bring this up to current **develop** unchanged. Further pushes will make actual changes.)

I've addressed the review comments to `bam_aux_remove()` and think that function should be ready for merging. I've added the proposed API for `bam_aux_remove_if()`, which facilitates removing several aux fields at...

Gah, I'll remove that last commit and bring `bam_aux_remove_if()` back as a separate PR later — this one has been hanging around quite long enough! So you can review this...

I think I know which other samtools branch that is :smile: Feel free to (force-)push to this PR's branch. Or use the “Squash and merge” UI button in due course,...

`hopen`/`hread`/`hwrite`/`hclose` are pretty obviously based on the `open`/`read`/`write`/`close` system calls, and you test whether you're at the end of the file in the same way: You try to read, and...

I know you know; I was pointing out that _it has a statement that the structure should be considered as opaque_ mixes up cause and effect. A philosophical point perhaps....