Shaun Jackman

I believe the current behaviour is to use `$type_$serial`, where `$serial` is a serial number. I'd suggest using (in order of preference): - `$Name` - `$ID` - `$type_$serial` `$type:$seqid:$start..$end` is...

Regarding coordinates, see the related issue https://github.com/genometools/genometools/issues/597

I'm definitely :+1: on #571, and a sed script to pick the appropriate FASTA ID wouldn't be so bad. I could suggest adding a `-retainnames` option. It might be nice...

The above hacky shell script does not work with `-join` easily. I'd find it really useful to have this issue or #571 implemented.

https://stackoverflow.com/questions/65351813/whats-the-right-way-to-populate-a-rust-ndarray-from-an-iterator-over-structs > The vector that is collected into is directly used as the backing storage for the array https://docs.rs/ndarray/latest/ndarray/type.Array2.html#method.from ```rust let arr: Array2 = iter .map(|row| [row.a, row.b, row.c]) .collect::()...

@pmelsted

:+1: (is my vote implied?)

@ekg @jts @lh3 @noporpoise @pb-jchin Sound good?

Unanimous agreement is certainly preferable. Let's see what happens. I wanted to jot down some criteria so that I'm not merging changes to the spec based on my own whim.

This error is reproducible on my machine at work, but if I don't see it at home or on CircleCI, I wouldn't worry about it. No worries at all! I'm...