wrong line number in encseq encode error message about genometools

[satta]

What error message did you receive exactly? The -protein option (as well as the -dna option) forces the encoder to use a specific alphabet definition, while otherwise the alphabet is determined from the characters encountered in the input sequence. Check core/alphabet.c for the exact DNA and protein alphabet definitions, they are hardcoded there and used when you specify -protein or -dna. If there are no lower case characters in these definitions then they are counted as invalid. If that is not desired then either a different alphabet definition must be used or the hard-coded alphabet must be extended (maybe the better option).
Regarding the wrong line number: I assume that is an simple off-by-one error due to 0- vs. 1-based indexing. I am sure I can fix this easily if you tell me which error you got.

[kowsky]

The error message is:
"bin/gt encseq encode: error: illegal character 's': file "encseq_error_minex.fasta", line 3"

There are indeed only upper case letters hardcoded in alphabet.c.

[satta]

I see. It may be worth the effort of extending it by lowercase characters then (I assume that's the way it is already done for DNA).

[gordon]

These alphabet definitions are pretty old, I think @stefan-kurtz defined them way before GenomeTools was started. I vaguely remember that there was a reason why protein characters have to be upper case, but I can't recall it. Maybe Stefan knows more.