Rob Patro

I just want to say that I'd also strongly support this feature! There are several places where I'm currently using rust-htslib where I might consider / prefer noodles instead, but...

I think this is reasonable _if_ such a tradeoff is necessary. However, it's not immediately obvious to me that a generification of the reader/writer/record type implies a speed hit to...

Hi @zaeleus, The progress on this is *great* news! I'm excited to try it out. Regarding the thoughts on the performance impact of the current design, I'm afraid I'm going...

This is very well stated @claymcleod. I would submit, @zaeleus, that a large speed deficit compared to `htslib` (or, rather, `rust-htslib`) would unfortunately probably deter a non-trivial fraction of folks...

Hi @Martinsos, Thanks for breaking this out into a separate issue. I like your general assessment and proposed (potential) solutions. While I see the overhead for *very* short sequences, I'll...

Hi @Martinsos --- that's for all of the great & quick feedback! The usage scenario I envision is the following. I have many threads running independently and each one repeatedly...

Hi @Martinsos, I just wanted to update you on this. We did, indeed, end up using edlib in our current implementation of the idea to which I was alluding (pre-print...

Hi @Martinsos, Well, I'm interested in potentially adding some alignment functionality to our [transcriptome mapping tool](https://github.com/COMBINE-lab/RapMap), and potentially using that information downstream in our [transcript quantification tool](https://github.com/COMBINE-lab/salmon). We have some...

Let me think a little bit about the splitting claim. But to answer your question about the alignment type, I'm thinking both about global alignment and alignment allowing for soft...

I also tried calling the `done()` method of `hash_counter` before scope exit, but it seemed to have no effect; the function still hangs.