Someone please tell me what form of insanity I'm suffering from so I can get it cured.
I'm running Matthew's native data store on a Linux 2.4.18 system with Sun/Blackdown Java 1.3.0. My freesite has been inserted, etc. If I'm logged into the Linux box and run this command: dwarf$ echo -e '\000\000\000\002ClientGet\nURI=SSK at wlUk5niRuNDUL4lzvdHu6DamhW8PAgM/Greg_Wooledge//\nHopsToLive=5\nEndMessage\n' | nc localhost 8481 Then I get a normal-looking FCP response: DataFound MetadataLength=76 DataLength=76 EndMessage DataChunk Length=76 Data Version Revision=1 EndPart Document DateRedirect.Target=freenet:SSK at wlUk5niRuNDUL4lzvdHu6DamhW8PAgM/Greg_Wooledge End But if I try this command from a different host on the LAN: jekyll$ echo -e '\000\000\000\002ClientGet\nURI=SSK at wlUk5niRuNDUL4lzvdHu6DamhW8PAgM/Greg_Wooledge//\nHopsToLive=5\nEndMessage\n' | nc dwarf 8481 Then I get just the first line -- or zero lines sometimes -- of the FCP response. I never get anything after "DataFound". This is 100% reproducible. I've tried "(nc dwarf 8481; sleep 3)" from jekyll with no change (still fails). I've tried "nc 192.168.2.4 8481" (dwarf's IP address) from dwarf itself with no change (still works), so it doesn't seem to be a loopback-versus-eth0 issue. If I run the same command from jekyll, but point it at my monolithic data store node (hostname: pegaus, OpenBSD + Kaffe) then I get the correct FCP response. Is this a bug in the native data store code? Or in the Linux JVM? Or in Linux? Or is it time for me to play in the rubber room? -- Greg Wooledge | "Truth belongs to everybody." greg at wooledge.org | - The Red Hot Chili Peppers http://wooledge.org/~greg/ | -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 187 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20020923/dcce8fc4/attachment.pgp>
