Jason,

I think you missed the words "GCC dialect of C".  Yes, GCC *is* within
the c99 rules by making this change, but it *is* also a change over
the accepted, relied-upon[*] GCC behaviour in gcc 1.x and gcc 2.x.  If
I'd gotten a report about it for gc 1., I'd have accepted it as a bug,
and I beleive RMS would have too. Thats all I'm saying.

[*] possibly even explicitly implementation-defined?  its been too
long since I read the source/texinfo.