This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: define_constraints patch, re-revised
>>>>> Zack Weinberg writes:
Zack> This is a good point, though. How about a brand-new header just
Zack> for those inline functions? We'd have to touch all the insn-*.c
Zack> files (people are going to want to call satisfies_constraint_* from C
Zack> conditions in MD patterns) but it will prevent this. Perhaps we should
Zack> move the constraint-related macro definitions there too (we would then
Zack> have to tweak a small number of MI files that use those macros).
Another approach is to leave the inline functions where they are
but protect them with a new macro, e.g., #ifdef GCC_CONSTRAINTS. Then one
defines the macro before including tm_p.h for the appropriate files that
may reference satisfies_constraint_*.
Either method is less easy than I was expecting for this
conversion. This may explain why many port maintainers were reluctant to
become early adopters.
David