[kaffe] HAVE_BROKEN_SO_RCVTIMEO

[Guilhem Lavaux]

Ito Kazumitsu wrote:

>Hi,
>
>  
>
Hi Ito,

Could you send me the config.log ? It should have said that SO_RCVTIMEO 
is 'broken' on freebsd 5.1.1 ... We have already noticed that the test 
doesn't work either on solaris.

Cheers,
Guilhem.

>Now on my FreeBSD 5.1-RELEASE machine, regression test for
>SoTimeout.java fails.  I think the following may be the
>cause of this failure.
>
>Now that kaffe/kaffe/kaffevm/systems/unix-pthreads/syscalls.c has been
>changed as follows:
>
>===================================================================
>RCS file: /cvs/kaffe/kaffe/kaffe/kaffevm/systems/unix-pthreads/syscalls.c,v
>retrieving revision 1.18
>retrieving revision 1.19
>diff -u -r1.18 -r1.19
>--- kaffe/kaffe/kaffevm/systems/unix-pthreads/syscalls.c        2004/07/29 14:24
>+++ kaffe/kaffe/kaffevm/systems/unix-pthreads/syscalls.c        2004/08/19 09:51
>@@ -469,7 +469,7 @@
>                int timeout, int* out)
> {
>        /* absolute time at which time out is reached */
>-#if defined(SO_RCVTIMEO) && !defined(__FreeBSD__)
>+#if defined(SO_RCVTIMEO) && !defined(HAVE_BROKEN_SO_RCVTIMEO)
>        int ret;
>        struct timeval old_tv;
>        struct timeval new_tv;
>
>then configure on FreeBSD should define HAVE_BROKEN_SO_RCVTIMEO.
>
>But actually it says:
>
>checking whether setsockopt(SO_RCVTIMEO) is broken... no
>
>And HAVE_BROKEN_SO_RCVTIMEO is not defined.
>
>_______________________________________________
>kaffe mailing list
>kaffe at kaffe.org
>http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
>
>  
>