On Sat, Dec 19, 2020 at 11:01:52AM -0500, Stefan Monnier wrote:
> > It's well known that perl's regexp is very powerful for its capability
> > of text processing. So, which is more powerful, emacs or perl, in this
> > scenario?
> 
> Probably Snobol or Icon.

Tcl ;-P

While somewhat tongue-in-cheek, I have some (by now older) reference to
back it [1]. The gist is that Tcl actually implements Thompson's clever
strategy (NFA -> power set DFA, yadda, yadda), while the P* languages
(Perl, Python, PHP, Ruby... to me, they are quite similar, all of them
[2]) all seem to use backtracking implementations. They do have a more
convenient (and strictly more powerful!) RE language, but they do pay
the price in terms of the occassional exponential runaway [3].

Now to the tongue-in-cheek part: Tcl has evolved since [1] (which was
around 2007) -- as have the other P*s. Among other things, I hear that 
they have overhauled their regexpery (which is something awe-inspiring:
most projects, once they have a regexp engine in them, DON'T TOUCH IT.
Too much black magic and forgotten lore in there :-) But Tcl folks are
like that.

Perhaps Tcl cas chosen the "convenience" path, and their regexp engine
resembles more that of the P*s these days.

Cheers

[1] "Regular Expression Matching Can Be Simple And Fast
     (but is slow in Java, Perl, PHP, Python, Ruby, ...)"
    <https://swtch.com/~rsc/regexp/regexp1.html>

[2] But I do enjoy (a limited amount) of those bragging contests
    "My Python is better  than your Perl, nyah, nyah". They tickle
    the hobby anthropoligist in me ;-)

[3] This afforded me a job asignment a long, long time ago

 - t