so you've made it unsigned long? what happens in places where the
value is set to or checked for -1 (avision.c) ?
allan
On Tue, May 5, 2009 at 10:09 PM, Michael Cronenworth wrote:
> m. allan noah wrote:
>>
>> yes- it would be nice if SANE_Pid was an opaque type, perhaps just an
>> array index or
m. allan noah wrote:
> so you've made it unsigned long? what happens in places where the
> value is set to or checked for -1 (avision.c) ?
>
I had to change those returns and surrounding functions. More work may
need to be done.
m. allan noah wrote:
> yes- it would be nice if SANE_Pid was an opaque type, perhaps just an
> array index or pointer to a struct. Right now the ranges of valid
> values are not the same pthread vs fork.
>
Actually I changed the typedef in sanei_thread.h from long to type
pthread_t and made ot
yes- it would be nice if SANE_Pid was an opaque type, perhaps just an
array index or pointer to a struct. Right now the ranges of valid
values are not the same pthread vs fork.
allan
On Tue, May 5, 2009 at 4:50 PM, Michael Cronenworth wrote:
> Hi,
>
> In sanei/sanei_thread.c there's this little
Hi,
In sanei/sanei_thread.c there's this little tid-bit of code:
DBG( 2, "pthread_create() created thread %ld\n", (SANE_Pid)thread );
"thread" is of type pthread_t, which is an extremely opaque type. You
won't get a reliable answer by casting it as long and printing it for
debugging purposes.
