Tor Arntsen wrote:
> On Mar 9, 17:03, Joe Halpin wrote:
>
>>Yes, it's 0 on Solaris 9 as well. Do you mind if I post the code you
>>sent to comp.unix.solaris and see what they say about it?
>
>
> Not at all, feel free to post it.
> I ran that test on linux and irix boxes, and some more I can't recall,
> and socktype was always != 0. Not on aix though and that's why it failed
> on those same tests that are now failing on your solaris(intel) build.
>
> On the other hand, if the port problem was exactly the same as the AIX
> one it should have been taken care of by the fix we installed for AIX.
> The code now looks like this (lib/ftp.c):
>
> rc = getaddrinfo(hbuf, NULL, &hints, &res);
> ..
> for (ai = res; ai; ai = ai->ai_next) {
> ..
> if (ai->ai_socktype == 0)
> ai->ai_socktype = hints.ai_socktype;
>
>
> so if the only problem was ai_socktype==0 then it should have worked
> also on solaris. You may have to dig down into gdb single-stepping,
> or at least run a test where you can look at the logs (the autobuild
> removes all residue, of course..)

Actually I think the problem may be more obvious. I'm not sure IPv6 is
enabled on my machine (I thought it was):

solaris$ ifconfig -a
lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index 1
         inet 127.0.0.1 netmask ff000000
elx0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2
         inet 192.168.1.2 netmask ffffff00 broadcast 192.168.1.255

I'm not sure why the autobuild is trying to use it though. The only
configure flag I'm using is --enable-debug. I'll look into that.

thanks again

Joe
Received on 2004-03-09