inotify patch

[Mike Frysinger]

On Wednesday 21 June 2006 04:47, Peter S. Mazinger wrote:
> On Wed, 21 Jun 2006, Joakim Tjernlund wrote:
> > This makes the configure step for cross compiling difficult. How should
> > one test for a feature without running a test program?
>
> As I understand the above
> #ifdef __NR_x should be OK even if cross-compiling (it is not
> bullet-proof, because syscall numbers were introduced partly earlier,
> before it was really implemented)

no, what he's worried about is cross-compiling applications against uclibc

so if a package does something like
AC_CHECK_FUNC(inotify_func)
AC_CHECK_FUNC(dnotify_func)
#ifdef HAS_INOTIFY
 < does inotify stuff >
#elif HAS_DNOTIFY
 < does dnoitfy stuff >
#else
 < does something else
#endif

and our uClibc has inotify stubbed to an ENOSYS, crappy applications will 
think inotify support is actually implemented (because the symbol exists) and 
disable code that would otherwise work
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 827 bytes
Desc: not available
Url : http://lists.busybox.net/pipermail/uclibc/attachments/20060621/a458434c/attachment-0002.pgp