On Sun, Mar 18, 2001 at 04:04:13PM +0000, Alan Fry wrote:
> At 9:09 am -0500 17/03/01, Ronald J Kimball wrote:
> >I don't think Bart has misunderstood the problem at all.
>
> With the greatest respect I think you have both misunderstood what I
> was driving at.
I think we're all misunderstanding each other. :/
> >Take the first four characters, which indicate the type.
>
> That IS the problem. Those four merry little characters (which may be
> 'TEXT' or 'styl' or any number of things) are buried contiguously in
> a mush of all sorts of other unrelated stuff. The problem is how to
> pick them out of the mush.
The four characters I was referring to are not buried in a mush. They're
the *first* four characters, which are at the beginning of the string. :)
> >Take the next however many bytes, which indicate the length of the
> >section of that type. Take the contents of the section, using the
> >indicated length.
> >
> >Repeat as necessary.
>
> Absolutely. If you were to look at the small script in my original
> posting, you would see that is exactly what does happen.
Your script does something different from what I was intending to describe.
Here's a quick script that demonstrates what I think is the appropriate
approach (untested):
#!perl
my %size = ( TEXT => 'L',
styl => 'S',
);
{
local $/;
$string = <>;
}
while (length $string) {
s/^(.{4})// or die "string too short";
$type = $1;
$size = $size{$type} or die "unknown type '$type'";
$len = unpack $size, $string;
(undef, $content, $string) = unpack "$size A$len A*", $string;
print "Got '$type', $len bytes:\n$content\n";
push @{$content{$type}}, $content;
}
__END__
Ronald
