This is the mail archive of the
mailing list for the GCC project.
Re: MIPS port and fixincludes
Well, I agree this is a minor issue, at least until we break
I think fixincludes should do as little as resonably possible
(for some definition of reasonable). Given that it's easy
enough for gcc to define whatever symbols the native standard
libraries expect, I don't think that what gcc predefines should
be an issue. Fixincludes should not be changing the header
files to suit gcc on this issue, though there may be other
reasons for fixincludes to make these kinds of changes (for
example to improve the ansi conformance of those headers). GCC
should define what ever those standard libraries expect it to
define (or what ever fixincludes changes the libraries to expect).
That said, the mips port defines __mips in all cases. It also
defines __mips__ in the case where host or target is openbsd.
This appears inconsistant, and may be an accedent, in which case
we should clean this up; there is no comment at the definition
to explain it's purpose. The specs for the mips port are
particularly twisty, makeing it difficult to be sure what's