Conflicting use of command line options

[sciurius]

Try

mma - < mysong.mma

This produces the error message "Input from STDIN specified. Use -f to set an output filename." Understandable.

But what if we do not generate output?

mma -n - < mysong.mma

I'd say this is legal, even useful. But mma throws a Usage error.

[bvdp]

This works:

  mma -f /dev/null - <test.mma

And, actually, I think it's a bug. Try changing the following:

1. main.py, line 125:

 ######################################
# Create the output filename

if not MMA.debug.noOutput:
    MMA.paths.createOutfileName(".mid")

2. options.py, line 265:

    # if a single '-' is left on the cmd line user want stdin. We set the
    # the input filename to numeric 1 which can't be entered.

    if gbl.infile == '-':
        gbl.infile = 1

        if not gbl.outfile and not(MMA.debug.noOutput):
            error("Input from STDIN specified. Use -f to set an output filename.")

Did a quick test here and it seems to work. Honestly, I've never had a need for stdin or for no output Just read and create things

 

[sciurius]

Thanks. I use this feature quite often during experiments for things like

echo printchord m7 | mma -n -

echo printactive | cat t.mma - | mma -

But I know, I'm an old fashioned command line scripter .

[bvdp]

Thanks. I use this feature quite often during experiments for things like

echo printchord m7 | mma -n -

echo printactive | cat t.mma - | mma -

But I know, I'm an old fashioned command line scripter .

Oh yeah ... the good old days. My little brain is already aching just looking at this

[sciurius]

And, actually, I think it's a bug. Try changing the following:

Still get errors. Shouldn't there be a case in opts() to actually handle the -n?

E.g. line 79 in parse.py:

        elif o in ('-d', '-o', '-p', '-s', '-r', '-e', '-c' , '-n'):
  

[bvdp]

No. The -n is handled by the debug code.

Line 80 of options.py:
 
    elif o in ('-d', '-o', '-p', '-s', '-r', '-w', '-n', '-e', '-c'):
              MMA.debug.cmdLineDebug(o[-1])
       
Check later today for the latest version.

[sciurius]

My mistake... I did intend to refer to options.py but accidentally mentioned parse.py.

In any case, it is clear that the handling of -n is not in the 16.06b sources and apparently you fixed it already in yours.

[bvdp]

Ahh, yet an another argument for github either that or shame on me for not doing more interm releases. See my new posting for great information!