[racket] Non threadsafe Fortran library (ARPACK) vs Racket 5.3.3

Hi Michael,

You were right. My mistake was that I assumed that the size of a Fortran LOGICAL
was 8 bits, whereas it is 32 bits (at least in g77's case).

Many thanks for your help,

Alexander




"Alexander Shendi (Web.de)" <Alexander.Shendi at web.de> schrieb:

>Hi all,
>
>So here it is
>
>
>
>Michael Wilber <mwilber at uccs.edu> schrieb:
>
>>Without places, racket is single-threaded, so non-threadsafe code
>>shouldn't be a problem unless you're using places.
>>
>>When the presence or absence of JIT compilation is significant, that
>>tells me that maybe you're doing something undefined with memory.
>>
>>Would you mind showing us the source code?
>>
>>"Alexander Shendi (Web.de)" <Alexander.Shendi at web.de> writes:
>>> Hi,
>>>
>>> So I habe now installed DrRacket and everything seemed fine until I
>>tried to
>>> access libarpack.so via Racket's foreign interface. I seem to have
>>two
>>> problems:
>>> - the library is not threadsafe.
>>> - "racket" and "racket -j" give different results for the same
>>problem.
>>>
>>> This leads to the following question:
>>> How do I call non threadsafe foreign code from Racket? I have
>already
>>> tried the #:atomic? and #:in-original-place? keywords when using
>>> '_fun'.
>>>
>>> I can provide the offending Racket program and the source to ARPACK
>>> if required.
>>>
>>> Many thanks in advance for your help.
>>>
>>> Best Regards,
>>>
>>> Alexander
>>> --
>>> Diese Nachricht wurde von meinem Android-Mobiltelefon mit K-9 Mail
>>gesendet.
>>>
>>> ____________________
>>>   Racket Users list:
>>>   http://lists.racket-lang.org/users
>
>--
>Diese Nachricht wurde von meinem Android-Mobiltelefon mit K-9 Mail
>gesendet.
>
>------------------------------------------------------------------------
>
>____________________
>  Racket Users list:
>  http://lists.racket-lang.org/users

-- 
Diese Nachricht wurde von meinem Android-Mobiltelefon mit K-9 Mail gesendet.