Patch to restrict extended identifiers to -fextended-identifiers option
Mark Mitchell
mark@codesourcery.com
Tue Sep 20 16:39:00 GMT 2005
Joseph S. Myers wrote:
> As per <http://gcc.gnu.org/ml/gcc-patches/2005-03/msg01665.html>, Zack
> no longer being in GCC development to do this patch himself, there
> having been no sign on the GCC lists of an audit of all uses of
> identifiers in the compiler for assumptions that break when
> identifiers have characters outside A-Za-z0-9_$ (and there being at
> least the case of diagnostic output as a known unfixed case of such an
> assumption), this patch restricts UCNs in identifiers only to be
> accepted when a special option -fextended-identifiers is passed. In
> so doing, it updates documentation which still indicated this feature
> was not implemented at all to indicate that there is a limited (UCNs
> but not the characters they denote accepted) experimental
> implementation conditional on the new option. It doesn't deal with
> the general question of documenting the phase 1 handling properly
> (under "Implementation-defined behavior" in cpp.texi), but some
> possible wording for that to improve on the present situation is
> suggested in bug 9449.
>
> Bootstrapped with no regressions on x86_64-unknown-linux-gnu. OK to
> commit?
OK.
It's certainly an improvement to clarify the documentation to indicate
tat there is in fact some support. The new option is perhaps more
controversial, but I think it's appropriate. It can become the default
if and when there's consensus that the implementation is sufficiently
complete; until that point, it's still available to users who need this
functionality by using the option.
--
Mark Mitchell
CodeSourcery, LLC
mark@codesourcery.com
(916) 791-8304
More information about the Gcc-patches
mailing list