Shell scripts: determining filename length

[Dave Ihnat]

On Wed, Sep 07, 2005 at 11:18:02AM -0400, Magnus Andersen wrote:
> I'd use the wc command with the -m flag to ge the number of charactes used 
> in the file name. You could use an awk substr to mod the name if longer than 
> 64.

I'd actually use "cut -d'.'" to get the filename and suffix.
Use "basename" if there's a chance there'll be a path prepended.
Then use the "wc -m" as recommended to determine string length.
(Actually, I don't see how "wc -m" is different than the traditional
"wc -c", but then, I just _wrote_ cut and paste...what do I know...hmm...
they took my name out of the latest man page???)

As to stripping the filename, well, "cut -c1-64" is a lot lighter
than awk.  However, you have to worry about identical filenames in the
first 64 characters; I'd actually strip it down to something like 61
characters, and append a sequence number to guarantee uniqueness.

Something like:

  COUNT=0;
  MAXFNLEN=64;
  CUTFNLEN=61;

  cd $DESTDIR;

  for FILE in *
  do
    FNAME=`echo $FILE | cut -d'.' -f1`;
    FSUFF=`echo $FILE | cut -d'.' -f2`;

    # If there is no suffix, cut will return the filename again
    if [ "$FNAME" = "$FSUFF" ]
    then
    	FSUFF="";
    fi;

    if [ `echo -n $FNAME | wc -c` -gt $MAXFNLEN ]
    then
    	FNAME=`echo $FNAME | cut -c1-$CUTFNLEN``printf "%03d" $COUNT`;
	let COUNT=$COUNT+1;
    fi;

    if [ "$FSUFF" ]
    then
    	OUTFNAME=$FNAME.$FSUFF;
    else
        OUTFNAME=$FNAME;
    fi;

    <do whatever you're going to do with the filename here>

  done;

Cheers,
--
	Dave Ihnat
	ignatz at dminet.com