... Could you explain what this means a little bit? I just tried it and it seems that it works not only in the BOOT section but everywhere else too. I ...
Jan Dubois
jan.dubois@...
May 1, 1998 8:34 am
23824
... open(FILE,"+<outfile"); ... -- Nick Ing-Simmons <nik@...> Via, but not speaking for: Texas Instruments Ltd....
Nick Ing-Simmons
nik@...
May 1, 1998 8:41 am
23825
... I'm puzzled by this thread. Pragmas are lexically scoped. To see tell what pragmas have been invoked I'd just look further back in the code. There's no...
Tim Bunce
Tim.Bunce@...
May 1, 1998 11:33 am
23826
... Be careful with your interpretation here. There may well be more problems (perhaps many more) that have not been found simply because two threads didn't...
Tim Bunce
Tim.Bunce@...
May 1, 1998 11:33 am
23827
This is a bug report for perl from jan.echternach@..., generated with the help of perlbug 1.13 running under perl 5.003. [Please enter...
Jan Echternach
echter@...
May 1, 1998 11:33 am
23828
... Here are my results - this on a single CPU UltraSparc (so the noise in results is not a multi-CPU issue as it is still there). I normalized to non-threaded...
Nick Ing-Simmons
nik@...
May 1, 1998 11:33 am
23829
Hi, I mailed the following to comp.lang.perl.misc and someone suggested I should send this via perlbug... -- I think I've stumbled across a PERL bug while...
Patrick Verdon
patrick@...
May 1, 1998 11:33 am
23830
... Not at all. Safe can be used to resolve the global issue and substituting print with an imported my_print should work well enough. Tim....
Tim Bunce
Tim.Bunce@...
May 1, 1998 11:33 am
23831
... Sorry for being obtuse, but are big numbers faster or slower? -Ben -- Ben Sugars <bsugars@...> Senior Webmaster, CANOE Canadian Online Explorer, ...
Benjamin Sugars
bsugars@...
May 1, 1998 2:00 pm
23832
... DS> The big problem (at the moment) with locked data is that locking doesn't DS> block data access, only other lock attempts. That's a bug. I can't see how...
Chaim Frenkel
chaimf@...
May 1, 1998 2:01 pm
23833
... I do hope that in regards to determining whether a module is thread safe or not, that the means is put into test suites that either are run by default in...
Larry W. Virden
lvirden@...
May 1, 1998 2:05 pm
23834
... BS> Sorry for being obtuse, but are big numbers faster or slower? Could we have this fixed. Could perlbench be fixed to print out what the numbers mean?...
Chaim Frenkel
chaimf@...
May 1, 1998 2:10 pm
23835
... DS> What, then, would you suggest it be changed to? I thought it was a block, DS> but I've been wrong before, and undoubetdly will be wrong again. CODEREF....
Chaim Frenkel
chaimf@...
May 1, 1998 2:11 pm
23836
... calling newCONSTSUB will work at any time, however the reason it needs to be called during BOOT is so that the subs are defined before the compiler sees...
Graham Barr
gbarr@...
May 1, 1998 2:11 pm
23837
... [No flame, just curiosity - really!] [>>> a number of comments edited for brevity here...] ... I see (and agree with, in some ways) your points. But Perl's...
Moore, Paul
Paul.Moore@...
May 1, 1998 2:51 pm
23838
... That would make it a bit difficult for someone to learn more about perl threads, so that they could decide if they wanted to try them, wouldn't it? -- ...
Larry W. Virden
lvirden@...
May 1, 1998 2:52 pm
23839
... That sounds very useful, both generally, as well as in the Perl context. If you can make them available, I for one would be interested in seeing them... ...
Moore, Paul
Paul.Moore@...
May 1, 1998 2:52 pm
23840
... Ouch! Tim....
Tim BunceTim Bunce
Tim.Bunce@...
May 1, 1998 4:35 pm
23841
... Nice work. Borland version number should be 5.02, not 4.02. ... Not correct. Detach says "I'm not interested in this thread's return value." On win32,...
Gurusamy Sarathy
gsar@...
May 1, 1998 4:36 pm
23842
... The number is speed. Big numbers are better. So a threaded perl is only 40% of speed on an unthreaded perl on the call cases. ... -- Nick Ing-Simmons...
Nick Ing-Simmons
nik@...
May 1, 1998 4:37 pm
23843
... Solaris threads are faster, but not dramatically: A) perl-5.00464 path = /home/road36/free-sw/perl5/bench/opt/bin/perl cc = gcc optimize...
Nick Ing-Simmons
nik@...
May 1, 1998 4:37 pm
23844
... I'll rework things so block's not misused this way. Thanks. Dan ... Dan Sugalski (541) 737-3346 even samurai SysAdmin...
Dan Sugalski
sugalskd@...
May 1, 1998 4:37 pm
23845
Jan Echternach <echter@...> wrote ... Mended in perl5.004 and later. Mike Guy...
M.J.T. Guy
mjtg@...
May 1, 1998 4:38 pm
23846
... results ... I'm not sure that this is a safe thing to do, as the entered sub could spawn off another thread. Without the locks in place, the other thread ...
Dan Sugalski
sugalskd@...
May 1, 1998 4:38 pm
23847
... Yes. It has got no better recently. My belief is that is mainly caused by all the MUTEX LOCK/UNLOCK stuff in pp_entersub(). Copy-pasted below. This gets...
Nick Ing-Simmons
nik@...
May 1, 1998 4:39 pm
23848
... NI> The number is speed. Big numbers are better. So a threaded perl is only NI> 40% of speed on an unthreaded perl on the call cases. Why are subs getting...
Chaim Frenkel
chaimf@...
May 1, 1998 4:39 pm
23849
... Wouldn't it make the most sense to plaster "If a module doesn't have an explicit disclaimer in its documentation that it is thread safe then one should...
... GS> Note that perl *never* kills a thread currently. Killing a thread GS> is considered *very* bad form (at least on win32, and I presume on GS> POSIX)....
Chaim Frenkel
chaimf@...
May 1, 1998 4:40 pm
23852
... results ... Hey, at least all his numbers were above 30. I had one test that bottomed out at 22 in one of my runs... Dan ... Dan Sugalski (541) 737-3346...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
