On Sat, Mar 24, 2012 at 2:54 PM, Mark H Weaver wrote:
> Nala Ginrut writes:
> > I know it's not the best implementation, I didn't consider the
> > multi-threads situation.
>
> Yes, unfortunately there is no way to set the locale on a per-thread
> basis. This is a limitation of the C library.
>
Nala Ginrut writes:
> Will it be a proper solution if I use "monitor"?
No, that wouldn't work. For one thing, what if the body of
'with-locale' takes a long time? Also, consider this:
Thread 1: (with-locale "foo" (strftime ...))
Thread 2: (format ...)
If the 'format' and the 'strftime' ha
According to our talking on IRC. I dropped locale specific procedure
setlocale in my code. I choose (web http) write-date instead.
And my conclusion is to avoid locale specific things in multi-thread
context. Writing a new procedure for the same
issue is a better solution.
Regards.
On Sat, Mar 2
Nala Ginrut writes:
> According to our talking on IRC. I dropped locale specific procedure
> setlocale in my code. I choose (web http) write-date instead.
> And my conclusion is to avoid locale specific things in multi-thread
> context.
It's okay to use locale-dependent operations in a multi-thr
BT Templeton writes:
> GOOPS defines an metaclass for this (similar
> to AMOP's funcallable-standard-class):
>
> scheme@(guile-user)> (define-class ())
> scheme@(guile-user)> ((make #:procedure car) '(1 2))
> $1 = 1
>
> Would it work to simply have all Python metaclasses inherit from
> ?
On s
Mark H Weaver writes:
> BT Templeton writes:
>
>> GOOPS defines an metaclass for this (similar
>> to AMOP's funcallable-standard-class):
>>
>> scheme@(guile-user)> (define-class ())
>> scheme@(guile-user)> ((make #:procedure car) '(1 2))
>> $1 = 1
>>
>> Would it work to simply have all Python
