DNA letter check logic

[ftegenfe]

Hello, when I run diamond on AA fasta files, it sometimes fails claiming there are only DNA letters even if that is not true. Looking in the code I see this conforms with the logic. This concerns the function SequenceFile::load_onepass() . I wonder isn't it more logical to fail the check only if ALL sequences in the test range (CHECK_FOR_DNA_COUNT) are DNA? I am aware of the --ignore-warning option but I encountered this in a package which uses diamond internally. It's a suggestion - I fixed this by rearranging the sequences in the fasta file :)

[bbuchfink]

Yes that would probably make sense changing that, I'll look into it.