There are of course a near-infinite number of 404 "pages", and this is potentially a major issue.
The problem was actually fixed, and then accidentaly unfixed in the process of putting out the security release recently. Look for an update very soon with a fix (again

)
It's being worked on as quickly as possible. If it's that urgent, you're certainly welcome to grab the code and submit a patch if you can get to it faster than the coding team.
Some clue of how fast the coding team would otherwise get to it (and Victor, it's assigned to you) would be helpful in figuring out whether I can get to it faster!
This isn't as urgent as I first thought, as I've realised there's another solution to the Google issue (see below) - but still I expect fake 404 pages to pass a 404 code to the agent if not the server and would be happier if MODx did that.
If it's going to be more than a week or so, then I volunteer to pick up your previous fixed code and reapply it to 0.9.2.1. If you could just tell me where I can grab that fixed code from!
The bug report at
http://modxcms.com/bugs/task/255?tasks=last mentions a fix at :
http://modxcms.com/forums/index.php/topic,1052.0.html but I cannot access that forum page - if you could please PM me the code, it would make this do-able.
Alternatively, if you can't get the code to me, let me know and I'll investigate other avenues.
Fixing the Google issueThere are also 2 potential ways of solving the Google problem, by asking it not to index the fake 404 pages:
1) Custom robots.txt file with known (old) pages disallowed
2) Add
<META NAME="ROBOTS" CONTENT="NOINDEX, NOFOLLOW">
to a custom template for the 404 page. (I guess you only want NOINDEX there but I don't know if that would work so it's probably best to leave it as per the official example.)
For more on this angle, see
http://www.google.com/support/webmasters/bin/answer.py?answer=35303I'm going to put the meta tag in my 404 template, it seems the best way forward.