A gapped alignment bug has been fixed that could, at low frequency, cause alignments to be significantly truncated at their beginning (amino terminus or 5' end).
Flat file parsing programs xdformat, gb2fasta, gt2fast and sp2fasta now recognize filename extensions that are suggestive of the input being compressed and automatically pipe the data through gunzip and unzip. (The gunzip and unzip programs must of course be in the user's PATH for this to succeed).
A new -wstrict option to guarantee the presence of at least one word hit in all reported alignments, when ambiguity codes in nucleotide database sequences might otherwise cause some alignments to be found that formally lack word hits.
