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

Re: [PATCH 5.005_62]Allow XS code to lock arbitrary scalars

Thread Previous | Thread Next
Gurusamy Sarathy
November 16, 1999 10:06
Re: [PATCH 5.005_62]Allow XS code to lock arbitrary scalars
Message ID:
On Tue, 16 Nov 1999 12:33:12 EST, Dan Sugalski wrote:
>>All that said, every time I see people patching USE_THREADS code I
>>wonder if it's all going to be for nothing.  I don't see much hope
>>for salvaging the existing model of USE_THREADS where prolific
>>locking is needed.
>Can we get a ruling on whether threads as they stand are in or out? And a
>statement of direction if they're in?

I can't make that judgement yet at this point, I want to wait to see
how the other way fares first in comparison.  (My *guess* at this
point would be that existing USE_THREADS will be out in the cold soon.)

>I really don't mind writing code that ultimately gets tossed out because it
>sucks or there's a better way to do things--it's a learning experience if
>nothing else, and the next time around the code's usually better. What I
>don't want to bother doing is writing code that is going to be tossed
>because I'm going the wrong way.

IMHO, you're going the wrong way, but I've said that before.

>I mean, at this point I've got code for Thread::Stop roughed out,
>preliminary mutex tracking set, and a scheme to make unsynchronized scalar
>access safe with about an order of magnitude less overhead than my last
>try. Getting them all going is a good chunk of work, and I don't mind doing
>it. I think it needs to be done, and that's not the point. I have other
>things to occupy my time, though, and I'd just as soon not tilt at these
>windmills without cause.

Well, I hope you're not saying nobody ever warned you.  (Watch out for
those neon "experimental" signs all over, buddy. :-)

>If the current threading model is dead, can we just shoot the poor thing
>and be done with it? Otherwise let's commit to it and figure out what needs
>to be done. This limbo state isn't doing anyone any good.

It's ain't dead until it's dead.  A deep coma has set in.  Whether
it is worth reviving is dependent on the quality of life afterward.


Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About