[PATCH] Fix cpp ICE (regression from 2.95.x or 3.0.4)

[Neil Booth]

Jakub Jelinek wrote:-

> The following patch fixes it by doing what gcc <= 2.95.x did, ie.
> silently disable -dM, -dI and -dN when -M or -MM is given, plus
> only uses the last of -dM, -dD and -dN options given on the command line

Would you also add something about this behaviour to the manual?  In the
absence of testcases (I know it's hard) having it all fully documented
is the best way to avoid regressions.

Thanks,

Neil.