Abort in dig after zone transfer

[Chris Thompson]

In the mood for bug reporting now ... :-)

Sometimes dig gets into some sort of mess after doing a zone transfer 
from certain hosts, e.g.

$ dig +nocmd +nostats axfr dlv.isc.org @ns-int.isc.org >dlv.new

The complete zone is written to the output file, perfectly correct,
but dig then sits there doing nothing. If one uses ctrl-C (SIGINT)
it dies like this:

dighost.c:3353: REQUIRE(sockcount == 0) failed.
Abort (core dumped)

Here's a stack trace from the core file, for what it's worth:

$ pstack core
core 'core' of 13457:   dig +nocmd +nostats axfr dlv.isc.org @ns-int.isc.org
-----------------  lwp# 1 / thread# 1  --------------------
 febc54d7 _lwp_kill (1, 6) + 7
 feb713f3 raise    (6) + 1f
 feb51709 abort    (8047480, 6, 8249a88, febbe196, 8047360, 806db5a) + cd
 081830d7 default_callback (81afbcc, d19, 0, 81afbbc) + 5b
 0806db5a destroy_libs (8047364, feffb818, feffb818, 8047364, 8047480, 804739c) + c6
 08067d95 main     (6, 80473a8, 80473c4) + 105
 08064b3e _start   (6, 80474e8, 80474ec, 80474f3, 80474fc, 8047501) + 7a
-----------------  lwp# 3 / thread# 3  --------------------
 febc47fb __lwp_park (82521c8, 8252198, 0) + b
 febbf058 cond_wait_queue (82521c8, 8252198, 0) + 3c
 febbf50a _cond_wait (82521c8, 8252198) + 64
 febbf54c cond_wait (82521c8, 8252198) + 21
 febbf585 pthread_cond_wait (82521c8, 8252198) + 1b
 0819d78e run      (8252190) + 62
 febc44b7 _thr_setup (fea20a00) + 4e
 febc47a0 _lwp_start (fea20a00, 0, 0, fea0fff8, febc47a0, fea20a00)
-----------------  lwp# 4 / thread# 4  --------------------
 febc5037 ioctl    (8250260) + 7
 febc44b7 _thr_setup (fea21200) + 4e
 febc47a0 _lwp_start (fea21200, 0, 0, fe9efff8, febc47a0, fea21200)

This was with the 9.6.0-P1 version, but I've seen it with other versions.

Fetching the signed root zone from na.iana.org also tends to behave
like this.

-- 
Chris Thompson
Email: cet1 at cam.ac.uk