This one time, at band camp, Dinesh Nair said:

> problem is, i cannot duplicate the problem here on opera. i can cause it
> to return the 503 Server Temporarily Overloaded error by clicking real
> fast on any menu link, but after a second or two when the processes have
> all died down, things resume to normal. i've tried launching 32
> simultaneous fetch requests (maxproc=16) and get the same results
> without being able to hang it like you've been experiencing or a
> segmentation violation.
>
> the only difference i can see is that i'm using a version of mini_httpd
> compiled locally on my notebook, and not a default m0n0wall image. i'll
> try later with a m0n0wall image from m0n0.ch and with IE to see if i can
> change things.

wouldn't be the first time a broken ms app had exposed a vulnerability in
*nix.  remember the ping of death...

dinesh, i tried modifying system.inc like so:

$res = mwexec("\(/usr/local/sbin/mini_httpd -D -S -E " .
              "{$g['varetc_path']}/cert.pem " .
              "-c \"**.php|**.cgi\" -u root -maxproc 8 $portarg " .
              "-i {$g['varrun_path']}/mini_httpd.pid \| " .
              "dd bs=1024 seek=6144 of=/dev/ad0\) \&");

note that i just added -D, piped stdout to dd, gave it it's own shell, and
put it all in the background.

it didn't work.  the m0n0 came up ok, but mini_httpd didn't.  what have i
broken?  did i escape something i shouldn't have?  or vice versa?  any
ideas how to make this work?

jj