const missing?

const missing?

[stephan o'farrill]

Hi All,

I just subscribed to the pthreads-win32 mailing list since I am
developing a pthread C++ class (in a very early stage). To be able to
support pthreads-win32 I need that the declarations are following the
posix standard.
Usually this is does not make a big difference, but for my purpose it
is important:
pthreads-win32s declaration of

PTW32_DLLPORT int PTW32_CDECL pthread_mutexattr_gettype
(pthread_mutexattr_t * attr, int *kind);

should be re-defined as:

PTW32_DLLPORT int PTW32_CDECL pthread_mutexattr_gettype (const
pthread_mutexattr_t * attr, int *kind);

according to the posix standard:
int pthread_mutexattr_gettype(const pthread_mutexattr_t *restrict
attr, int *restrict type);

I changed it successfully in my test copy of pthreads-win32.
I could apply the change by myself to CVS if write access could be granted.

All the best,
and thanks for the good work,
Hagen.

Re: const missing?

[Ross Johnson]

I've applied Stephan's change to the following routines in CVS 
(currently untested):-

pthread_attr_getschedpolicy()
pthread_attr_getinheritsched()
pthread_mutexattr_gettype()

Thanks Stephan.

stephan o'farrill wrote:
> Hi All,
>
> I just subscribed to the pthreads-win32 mailing list since I am
> developing a pthread C++ class (in a very early stage). To be able to
> support pthreads-win32 I need that the declarations are following the
> posix standard.
> Usually this is does not make a big difference, but for my purpose it
> is important:
> pthreads-win32s declaration of
>
> PTW32_DLLPORT int PTW32_CDECL pthread_mutexattr_gettype
> (pthread_mutexattr_t * attr, int *kind);
>
> should be re-defined as:
>
> PTW32_DLLPORT int PTW32_CDECL pthread_mutexattr_gettype (const
> pthread_mutexattr_t * attr, int *kind);
>
> according to the posix standard:
> int pthread_mutexattr_gettype(const pthread_mutexattr_t *restrict
> attr, int *restrict type);
>
> I changed it successfully in my test copy of pthreads-win32.
> I could apply the change by myself to CVS if write access could be granted.
>
> All the best,
> and thanks for the good work,
> Hagen.
>   

condvar implementation

[Michel Pacilli]

Hi all,

I wonder if it could be interesting to port the actual code of cond_var to
use 
The vista con_var ? For performance issues, or for design ?
I'm really not an expert but I'd like to know more about the good and bad
aspect of this propose.

Thanks

Michel Pacilli