Increase robustness of functions that read fasta format files

[andrewdavidsmith]

In smithlab_os.*pp there is a function that reads a fasta format file to extract a sequence with name specified as an argument to the function. The way this function reads the file allows for at most 1M chars per line, corresponding to a fixed size input buffer. This is somehow broken and should be fixed. The suggestion is to fix it in the same way as for the other function for reading fasta, which reads all sequences in the file.