Re: splitting ASCII output?

[Patrice Dumas]

On Mon, Apr 07, 2014 at 04:54:12PM -0600, Karl Berry wrote:
> 
>    texi2any.pl --plaintext --split=section -o dir/ -c OUTFILE=undef 
> manual.texi
>    which is rather ugly.  Any idea?
> 
> The only part that seems unexpected to me is the -c OUTFILE=undef.
> Can the -o be made sufficient?

It can, simply by setting OUTFILE to undef when using -o even if a
directory is detected.  


Here is what this would lead to:

1) ./texi2any.pl --plaintext --split=chapter manual.texi
 => output on stdout, not split.

2) ./texi2any.pl --plaintext --split=chapter manual.texi -o somewhere
 => output split in somewhere/.  Top element file is called somewhere.txt.

3) ./texi2any.pl --plaintext --split=chapter manual.texi -o somewhere/
 => output split in somewhere/.  Top element file is called manual.txt.

4) ./texi2any.pl --plaintext --split=chapter manual.texi -c OUTFILE=undef
 => output split in manual/.  Top element file is called manual.txt.

5) ./texi2any.pl --plaintext --split=chapter manual.texi -c TOP_NODE_FILE=jojo
 => output on stdout, not split.

6) ./texi2any.pl --plaintext --split=chapter manual.texi -o dir -c 
TOP_NODE_FILE=jojo
 => output split in dir/.  Top element file is called jojo.txt.




If 'TOP_NODE_FILE' is set in the default case for Plaintext (for example
to coco), in cases 2, 3, 4, the top element file will be called
coco.txt.


So 2 questions.  First is the above ok?  Second, should TOP_NODE_FILE be
set in the default case?

-- 
Pat