On Feb 9, 13:30, "Yang Tse" wrote:
>Tor
>
>On the daily build log we can see that on 2-Feb there is a single AIX
>log which has completed quite nicely, and no more AIX logs. But since
>3-Feb _ALL_ AIX daily builds are failing due to an already in use 8990
>port.
>
>Are they failing due to the current test ?

No, it's the semi-normal problem of the server/sws not stopping.
It happens very often on Tru64 and occasionally on AIX, on the Tru64
build machine I run a kill-script to avoid it (it didn't always avoid
it so I improved the script only yesterday). On AIX it's been rarely
enough that I kill manually when I notice. In this case it had been
hanging since Feb 2 - I hadn't had time to log in and check.
Killed now.

>Or, has it been just a quirk which might remain in libcurl's test
>servers, and an old server happens to be running at that port ?

That's exactly it.

>In other words... Old testsuite quirk, or new issue introduced mot
>probably by me with some "compiler warning fix" ?

Old :-)

-Tor
Received on 2007-02-09