REGRESSIVE vs default: different behaviour when printing to STDOUT

[lrauschning]

Hi Cedric & Jose,

while debugging the multiplesequencealign pipeline, I came across an inconsistency in the way tcoffee gives its output when told to print to stdout.
In particular, a file called 'STDOUT' is created when the regressive mode is used and STDOUT is given as the output, if the regressive mode is not used the output is printed to stdout (see also below) – I think the output gets handled quite differently in general, as the output format is also different by default.

I think it makes sense to handle the output consistently for these cases – also supporting '-' to specify printing output to stdout might make sense.
Best,
Leon

$ t_coffee -seq example.fa -outfile STDOUT
<snip>
# output in clustal-like format sent to stdout, debugging info in stderr
$ file STDOUT
STDOUT: cannot open `STDOUT' (No such file or directory)

$ t_coffee -seq example.fa -reg -outfile STDOUT
<snip>
# debugging info gets sent to stdout and stderr, 'STDOUT' is created as a file
$ file STDOUT
STDOUT: ASCII text