[335] in Zephyr_Bugs
Re: neskaya zephyrd losing
raeburn@ATHENA.MIT.EDU (raeburn@ATHENA.MIT.EDU)
Fri Jan 17 17:26:49 1992
The bogus address was because inet_ntoa was being passed sockaddr_in instead
of sockaddr_in.sin_addr in the call to syslog; I'm not sure of the cause of
Sounds like a glitch in the translation to C. I had defined an
inet_ntoa taking a sockaddr_in to cut down the repetitious ".sin_addr"
all over the place, so it'd be easier to read.
>Zephyr locations have been unreliable and unpredictable too.
This is due to a known bug in the old zephyr server code (which is not present
the new server). Unfortunately, the location database corruption will
continue to occur until b11zephyr is running the new server as well.
What bug was this?
The only severe ones I knew of in the ancient server b11zephyr is
running is that brain-dumps would not transfer the locations, and
sometimes locations wouldn't go away. The first got fixed while I
worked on it.
The problem yesterday was that two successive zlocate requests would
get different results; I've never seen that before.
