develooper Front page | perl.perl5.porters | Postings from September 2000

Re: t/op/regmesg.t test 13 fails in bleadperl

Thread Previous | Thread Next
Dominic Dunlop
September 6, 2000 01:51
Re: t/op/regmesg.t test 13 fails in bleadperl
Message ID:
At 23:29 -0500 2000-08-17 (just as I went away), Jarkko Hietaniemi wrote:
>On Thu, Aug 17, 2000 at 10:03:32PM +0200, Dominic Dunlop wrote:
>  > [My stack check patch] didn't get in.  Main objection was that the manner
>  > of establishing available stack space varies widely between platforms
>  > (and according to whether you're threaded or not).  I only did the
>Also, when do you stop upping your stack?  When setrlimit() says no
>can do more?  When you get a signal telling you stop doing that?  When
>the system starts swapping?  When your system admin comes through your
>door (and I do mean *through*) demanding to know what the <bleep> you
>think you are doing?

That's not really a valid objection: firstly, setrlimit's going to 
stop giving you more stack when it hits whatever hard limit your 
sysadmin has imposed (so you say ingenuously as said person brushes 
splinters from their clothing, "But I assumed you would have set a 
sensible hard limit..."); secondly, Perl (and almost every other 
program of any complexity) has always felt at liberty to grab as much 
memory as it can get away with using malloc() (again subject to 
whatever hard limit may be in force): I don't see that letting it 
also get greedy via the stack segment makes it any less antisocial 
than it already is.
Dominic Dunlop

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