More troubles with MIME parser

I have tried some more work with the MIME parser in the W3C Reference Library,
and I've run into trouble.

I have a multipart input file that has CR-LF pairs at the ends of the lines.
The MIME parser gets into HTBoundary_put_block, finds the first occurrence
of the boundary, and sets state to EOL_DOT.  It then falls into code that
looks like:

	    if (*b == '-') {
	    } else if (*b != LF) {
		me->dash = 0;
		me->state = EOL_BEGIN;

*b is the CR of the CR-LF pair that ends the line.  It is neither a hyphen
nor a linefeed, so the state reverts to EOL_BEGIN.

Eventually, we reach the end of the block of data, (l = 0), attempt to
PUTBLOCK to a target stream that has never been initialized, and dump core.

It also appears that the code can't work if the --BOUNDARY string isn't
all in the same block of the input.

Am I missing something, or is the parser seriously buggy at this point?

73 de ke9tv/2, Kevin KENNY   GE Corporate Research & Development
kennykb@crd.ge.com           P. O. Box 8, Room KWC273
                             Schenectady, New York 12301-0008 USA