On Tue, Nov 27, 2001 at 01:33:25PM -0500, Christopher C. Chimelis wrote: > > On Tue, 27 Nov 2001, Daniel Jacobowitz wrote: > > > As with that bug, no, GCC should complain about dollars starting > > identifiers. Try using b$c instead of $b. > > Oddly enough, our powerpc gcc packages have --no-dollars-in-identifiers > enabled by default, despite gas having no problems handling things like > $b (for the record, alpha, mipseb, and sparc gas also can handle the > sample code in the bug report without a problem and none of the gcc > packages on those machines required an explicit > --dollars-in-identifiers). x86 seems to be the oddball in this regard in > that gas can't handle cases like $b currently and yet our gcc debs allow > those cases by default. > > So, my question is, what are the criteria for determining whether or not > gcc should reject identifiers with dollars? I'm of the opinion that, at > least for Debian, gas support for that should make that > determination. After all, if gcc just warns about cases like $b, but gas > rejects the code, then the warnings are redundant unless they say > something like "dollar identifier may not be supported by the assembler on > this target". I still vote for disabling dollared identifiers by default > on x86 and, now, enabling it on ppc (after proper testing, of course).
You won't be able to build X86 kernels if you do that :) Well, not with things like NTFS support, at least. I don't know of any other offenders offhand. I'm most strongly of the opinion that this isn't Debian's problem. Our GCC should support whatever GCC has decided to support. -- Daniel Jacobowitz Carnegie Mellon University MontaVista Software Debian GNU/Linux Developer
