I just sent you a test script. Can you send me that output? Maybe it's not doing what I think it's doing.
Yes, I'm talking about t/450_service.t The modifications in version 2.25 SHOULD have allowed nonblocking connects to give ECONNREFUSED immediately instead of the previously required EINPROGRESS. Apparently some optimizations on SMP systems may also come back with the real error instead of EINPROGRESS if the status can be determined early enough. But you say it is still fails? I think it must be a different situation. Everyone does not need to run a webserver (not even on localhost). But I do assume the sites tested do not block (firewall) tcp port 80. -- Rob On Wed, 20 Nov 2002, Jost Krieger wrote: > On Tue, Nov 19, 2002 at 04:13:20PM -0700, Rob Brown wrote: > > Ugh, you're killing me. You don't get the > > proper "Connection refused" error when trying > > to connect to localhost? Try this on the > > Solaris box: > > > > $ telnet 127.0.0.1 12345 > > Trying 127.0.0.1... > > telnet: Unable to connect to remote host: Connection refused > > $ telnet 127.0.0.1 12346 > > Trying 127.0.0.1... > > telnet: Unable to connect to remote host: Connection refused > > $ telnet 127.0.0.1 12347 > > Trying 127.0.0.1... > > telnet: Unable to connect to remote host: Connection refused > > $ > > > > Any three random ports should be fine. It > > needs to say something like that and under > > perl it must set $! = ECONNREFUSED, right? > > What did it say? > > All this works normally, it seems. > > You're talking about 450. right? > > I've instrumented it, and it looks like the (supposedly nonblocking) connect to > the closed port $port2 returns immediately with ECONNREFUSED. This seems to > mess up things. Is that unreasonable for 127.0.0.1? > > The other problems are that you are expecting everyone to run a webserver, it seems >... > > Jost
