Hi,
Luke Palmer wrote:
> On 7/14/05, Nicholas Clark <[EMAIL PROTECTED]> wrote:
>> This is more a "note to collective 'self'" question than one I expect
>> the answer to right now. (The answer I expect right now is a glib "it
>> will")
>>
>> How will the perl6 compiler cope with people creating threads inside
>> BEGIN blocks?
>
> A glib "it won't".
>
> Unless someone can come up with a reason you would want to do that.
class SomeCryptMod {
my %tables_which_are_expensive_to_create =
precalc_tables(); # called at BEGIN time
sub precalc_tables () {
# Use threads to speed up calculation:
for 0..10000 -> $i {
# Or whatever the thread API is.
Thread.create:{ precalc_table_for $i };
}
Thread.wait_for_all_threads;
}
...;
}
If you compile that module then, the compiler will only have to
calculate %tables_which_are_expensive_to_create once (as the results
are serialized in the .pbc), and the code could use threads to speed up
calculation.
(But I'm fine with outlawing it, or making it a new feature in Perl
6.x.y, or whatever :))
--Ingo
--
Linux, the choice of a GNU | self-reference, n. - See self-reference
generation on a dual AMD |
Athlon! |
