On Jan 19, 2006, at 6:23 PM, David Lee wrote:
On Thu, 19 Jan 2006, Andrew Beekhof wrote:
BSC passes fine - I have no objections.
BSC (presumably BasicSanityCheck): what platform? (Your earlier
email was
about OSX).
Yeah, OSX. I'll be running it on sles9 (x86) shortly.
Heartbeat 2.x (at least with IPaddr) has worked for some time on
Solaris.
But BSC on Solaris has never passed. Even with this patch (and its
two
minor requisite precursors in CVS earlier this week) BSC still
seems not
quite right. But at least it's doing far better than before.
Whilst I have a selfish interest in Solaris, nevertheless the most
important thing to check is that my patch doesn't break the other
platforms.
Aside: On _any_ platform that has streams, the streams-specific
part of
"ipcsocket.c" should be compile-time selectable by defining
HB_IPC_METHOD
as value HB_IPC_STREAM. Then anyone can get involved in debugging it!
[Just to re-iterate: the "#if ..." logic near the top of
"ipcsocket.c" has
the default intention of using "sockets" wherever possible,
wherever it
can find an associated credentials mechanism. It only falls back
to the
streams mechanism (which seems to have its own inherent
credentials) as a
last resort (or second-last if we count the "dummy" code).]
We should probably make sure this is in before 2.0.3
Yes. I would certainly very much like to see this. That's why I'd
like
non-Solaris folk to make sure I haven't broken anything for them.
Thanks, Andrew. Your work is appreciated!
No probs, I'm actually hoping this may have helped some intra-program
IPC weirdness I've been seeing on OSX...
So many itches to be scratched :-)
--
Andrew Beekhof
"If it weren't for my horse, I wouldn't have spent that year in
college" - Unknown, courtesy of Lewis Black
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
