|FROM ||Neil Gunton
|SUBJECT ||Re: [Hangout - NYLXS] Embperl and threaded mpm?
|On 1/2/22 06:10, Ruben Safir wrote:
> Is it a good idea to adopt this package and update it?
> It wil likely die otherwise
I think it's pretty much "dead" anyway, at least in terms of new people
and websites using it. But then so is Perl, and even mod_perl. Hell,
even Apache is seen as passe by some people now. I still use all those,
and Sendmail, so call me a dinosaur. They work, and I have all my
configs dialed in for those tools. And I have a huge codebase (well, 20
man year's worth of work on my part for my website). So I'm going to use
Embperl for as long as I can. But I couldn't imagine using it to develop
a website for clients at this point, in good conscience. So who's using
it? I have no idea. I don't think Gerald Richter has been all that
interested in developing it for a decade or more.
If someone wanted to take it on and keep it relevant, even develop it
and expand and document all the cool features properly (I have never
used "recipes" or whatever Gerald called them) then maybe it could have
a rennaissance. I don't know. I don't have time, I have enough on my
plate just developing my own stuff.
People like to work on new stuff. Even on established open source
projects, you see people doing total rewrites every so often, just (I
guess) because it's easier to work on new code than maintain old. It's
an old story. I'm going through this right now with Thunderbird - 91.x
seems to be a lot of new code, because stuff that used to just work, now
doesn't. That's the downside to rewrites - all those years of
accumulated coding wisdom, little tweaks, down the drain.
So I'd hope that if anyone does take on Embperl, they won't rewrite from
scratch. Perl 6 is another example. It basically killed Perl, because
everybody was eternally waiting for Perl 6, which made Perl 5 look
passe, and eventually everybody just stopped waiting and used PHP or
Ruby or Python or whatever. The Perl devs really screwed the pooch on
Perl 6, in my opinion as a peripheral observer and developer. I don't
really care how cool the language is. If it's not a smooth and seamless
transition from Perl 5 to Perl 6, then it's useless to me. I don't care
about cool language features, I just want to get stuff done, and not
break all the code I have accumulated over the last 20 years.
Embperl is still great. But increasingly I find I use real Perl subs
rather than the [$ sub $], due to the better handling of local
variables. Most of the big bugs I've had have been due to leaking
variables across different Embperl subs. So now I only try to use those
for very basic stuff. If it's at all complex, then I convert that code
to a Perl sub, inside [- -], not [$ sub $]. The feature that I really
like about Embperl is the object oriented directory structure, which
lets you redefine files and subs in sub folders. I may be wrong, but it
really works for me and I like it. That's the big thing I wouldn't want
to give up if I had to move away from Embperl, really - the folder-based
hierarchy of subroutine overriding. Maybe there's clever ways to get the
same effect by other means, I'm sure there are, but Embperl still "just
works", it's stable so I hope it continues to work for a long while yet.
I don't know if Gerald is still out there reading these emails, but if
he is - THANK YOU. You gave us an awesome tool, and I hope if I ever get
my business successful one way then I'll be able to repay you somehow.
To unsubscribe, e-mail: embperl-unsubscribe-at-perl.apache.org
For additional commands, e-mail: embperl-help-at-perl.apache.org
Hangout mailing list