Re: Paralizing configure

[Peter Rosin]

Den 2011-02-09 05:42 skrev Miles Bader:
> On Wed, Feb 9, 2011 at 4:39 AM, Ralf Corsepius <address@hidden> wrote:
>>> At the least, constructs like
>>> AC_CHECK_HEADERS([stdint.h unistd.h fcntl.h sys/mman.h sys/stat.h])
>>> could check all the entries in parallel.
>>
>> Only simple scenarios, in which headers and header search paths are static -
>> This often doesn't apply to complex configurations.
> 
> Wait, I'm confused ... how can the results of checking for "stdint.h"
> affect the subsequent check for "unistd.h"?
> 
> [I meant, all the entries _within a single invocation_ (of
> AC_CHECK_HEADERS, etc) could be checked in parallel...]

Yes, in a perfect world.  If you e.g. do this:

AC_CHECK_HEADERS([ \
        windows.h \
        winsock2.h \
        winsock.h \
        ], [], [], [
AC_INCLUDES_DEFAULT
#define WIN32_LEAN_AND_MEAN
#ifdef HAVE_WINDOWS_H
#include <windows.h>
#endif
#ifdef HAVE_WINSOCK2_H
#include <winsock2.h>
#endif
#ifdef HAVE_WINSOCK_H
#include <winsock.h>
#endif
])

The outcome of the winsock.h test depends on if the winsock2.h test
succeeded or not.  At least for some versions of the headers, as
provided by MS.

Cheers,
Peter