dig problem in 9.8.1b2?

[Bill Owens]

Working on a project related to TCP blocking on authoritative servers, I accidentally typed a command that caused an error in dig. It appears that combining '+nssearch' and '+tcp' flags triggers it, regardless of the server being used or the query:

kindofblue:~> dig @127.0.0.1 +nss +tcp .
socket.c:2499: REQUIRE(socketp != ((void *)0) && *socketp == ((void *)0)) failed, back trace
#0 0x81763fe in ??
#1 0x8176577 in ??
#2 0x819cbbd in ??
#3 0x8052f92 in ??
#4 0x8053333 in ??
#5 0x8191139 in ??
#6 0x8193cf2 in ??
#7 0x8193f9f in ??
#8 0x8194522 in ??
#9 0x804f0e3 in ??
#10 0xb7578e37 in ??
#11 0x804b291 in ??
Abort

That's on a Linux/Intel machine, here's the result on my PowerPC:

[turkle:~] owens% dig @127.0.0.1 +nss +tcp .
socket.c:2499: REQUIRE(socketp != ((void *)0) && *socketp == ((void *)0)) failed, back trace
#0 0x12bc64 in ??
#1 0x12bb68 in ??
#2 0x1548d0 in ??
#3 0x9874 in ??
#4 0xb00c in ??
#5 0x149298 in ??
#6 0x14bde4 in ??
#7 0x14c1d0 in ??
#8 0x552c in ??
#9 0x1d94 in ??
Abort

Odd thing, the command hangs for a few seconds on the PPC, between the #9 line and Abort. 

I don't think there's anything wrong with that combination of flags, though it's probably a bit uncommon. Anyway, that's the limit of my debugging skills, so I'll leave it to the people who know what they're doing to figure out what's broken ;)

Bill.