3.0.0: job table full causes fg to fail?

[Janos Farkas <chexum@xxxxxxxxxxxxxxx>]

Howdy people!

I've just run into a problem, and can't imagine how it could be fixed.
It's probably present in earlier versions too.

I've been trying to stress my home-net with some concurrent pings, and
suddenly..

[iceq] /usr/src 514> ping ancient >/dev/null &
[47] 24995
[iceq] /usr/src 514> ping ancient >/dev/null &
[48] 24996
[iceq] /usr/src 514> ping ancient >/dev/null &
zsh: job table full
[iceq] /usr/src 514> ping ancient >/dev/null &
zsh: job table full

   Heh, cool, limits?  Hmm.. it's right, I shouldn't have been doing
   this...  So let's get rid of them...  I used to fg, and ctrl-c
   out of it, but:

[iceq] /usr/src 514> fg
fg: no current job

   What?  Now, how am I supposed to kill 'em?

[iceq] /usr/src 515> killall ping
zsh: job table full

   WHAT?  :)  Hmm... Okay, I don't have a free slot to run killall,
   but fg shouldn't have been tricking me?

[iceq] /usr/src 516> _