develooper Front page | perl.perl5.porters | Postings from October 1999

Re: perlthread.pod -- First draft of thread docs

Thread Previous | Thread Next
From:
Rob Cunningham
Date:
October 26, 1999 07:22
Subject:
Re: perlthread.pod -- First draft of thread docs
Message ID:
3815B928.84F19E7E@ll.mit.edu
Dan,
I have bet the software barn on two things occuring:
	(1) Perl threading and regexp working properly. (Else--why use threaded perl
at all?)
	(2) Perl threading getting more efficient.

One is must-have, two is would-be-nice.

I'll add a comment before I start a war. For what I need to do, the best thing
about perl is its regexp parser, which is more full-featured than anything
else I can find. It's speed is good too, which is critical. When I needed to
structure my code to simultaneously parse multiple streams, I could have
either (a) rewritten my perl objects in C++ (starting over) or (b) made use of
perl's threading mechanism, and kept most of the code base. I did the latter,
even though threading was listed as experimental. I plan to work to help make
the next version of perl threading work, and we have committed some resources
to help make that happen.

Next Question: Will somebody knowledgeable about the debugger fix the perl
debugger as I described in bug report [ID 19990820.008]? I even included a
simple test case...  I desparately need a working debugger in the threaded
environment; adding print statements is getting wearisome.

Rob

Dan Sugalski wrote:
> 
> At 08:43 AM 10/26/99 -0400, Rob Cunningham wrote:
> >Yes. 62 came out just as we were poking at the ugly bits of perl, so we
> >started work there. I hope that people really stop adding features and
> >concentrate on getting the included features to work properly.
> 
> Then here's a question--would you be willing to trade off some of the speed
> that came with _62 to get a thread-safe (well, safer, at least) core? No
> promises of atomicity or anything, but making unsynchronized access to
> shared variables not potentially core?
> 
> >And as for bravery--we're not! We need this fix, or we wouldn't do it.
> 
> Hey, you're in there and haven't gone gibbering off into the night waving
> to Aldeberaan. You're not giving yourself enough credit here... :)
> 
>                                         Dan
> 
> ----------------------------------------"it's like this"-------------------
> Dan Sugalski                            even samurai
> dan@sidhe.org                           have teddy bears and even
>                                         teddy bears get drunk

-- 
Dr. Robert K. Cunningham       Information System Technology Group
  rkc@ll.mit.edu                     MIT Lincoln Laboratory           
*** My comments, my opinions: my responsibility.
PGP key available from http://pgpkeys.mit.edu:11371

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About