devel/nspr failing prtests

[Joe Marcus Clarke]

On Thu, 2006-05-11 at 14:44 +1000, Geoffrey Giesemann wrote:
> Should nspr be passing all its prtests on a 6.1-RELEASE x86 box?

Apparently not.

> 
> The following tests are causing grief:
> 
>  - forktest fails
>  - socket never finishes - ktrace shows it looping endless around:
>         50838 socket   CALL  poll(0xbeef3e20,0x1,0x1388)
>         50838 socket   RET   fork 0
>         50838 socket   CALL  kse_release(0x8076f40)
>         50838 socket   RET   kse_release 0
>         50838 socket   CALL  poll(0xbeef3e20,0x1,0x1388)
>         50838 socket   RET   fork 0
>         50838 socket   CALL  kse_release(0x8076f40)
>         50838 socket   RET   kse_release 0
>         (etc etc)
> 
> (or should this be directed into the appropriate Bugzilla?)

Actually, the socket failure might point to a bug in our IPv6 stack.
The test that locks up is the UDP client/server IPv6/IPv4 test (and all
the TCP tests pass).  The fork test looks like a potential problem in
libpthread.  Running the same test using libthr instead of libpthread
does not trigger the abort message, but the test appears to lock up in
[lthr].

That said, we do not appear to be having any practical problems with
nspr.  That is, Gecko browsers and Evolution (some of the biggest nspr
consumers) operate fine.

Joe

-- 
PGP Key : http://www.marcuscom.com/pgp.asc
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: This is a digitally signed message part
Url : http://lists.freebsd.org/pipermail/freebsd-gnome/attachments/20060513/557ef7d1/attachment.pgp