> In fact, it is not the length or the '\\\\\\\\' strings needed and I
> know that there are some work-arounds.
> Still I think that it is a PITFALL: naturally when trying to replace \
> by \\, knowing that '\' are escaped characters, you would write:
> gsub(/\\/,'\\\\').
> 
> The need to either write gsub(/\\/,'\\\\\\\\') or use a work-around
> gsub(/\\/){'\\\\'}, indicates (for me) than something is broken..

That seems harsh. =)

The only problem is that you're dealing with two parsers that happen to use
the same character to do the same thing. It's a side-effect of the
pure-OO'ness, and the truth is that if the ruby parser were to compensate for
this special case, you'd be bastardizing the OO, since the parser shouldn't
care where the string is headed.

It's an example of how sometimes specialized tools (awk, perl, etc.) can be
more convenient than more generic, all-purpose stuff, albeit in a minor way.

If anything, it's indicative of one of the things I LIKE about Ruby: it's
largely free of any of Perl's famed "magic".

But anyways, this is a really silly thing to be writing such long e-mails
about.