"Robinow, David" <[EMAIL PROTECTED]> wrote:
> > Thanks for the pointer, but did you see my original question?
> > How can I have
> > *only* thread local errno variable? Under Linux that
>
> #include <errno.h>
What then? As far as I understand, it will work *only* if I have global reent
struct for each thread, and then I must switch _impure_ptr pointer in each
thead *manually*.
> > functionality is provided
> > by overriding __errno_location() function. Under Cygwin
> > definition of my own
> > __errno() function doesn't help.
> Why do you want your own __errno function?
Probably I wasn't clear enough in my previous postings, sorry. It's not me,
it's an application that I want to port to Cygwin. Is it so unusual to have
*an application provided* errno variable (for every distinct thread)?
As I already mentioned, that approach works fine in Linux. As far as I know,
similar mechanisms are provided by FreeBSD (__error()), Solaris (___errno()),
UnixWare (__thr_errno()).
Is it true that Cygwin has no analogous functionality?
Anyway thanks for your help.
--
Dmitry.
--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple