linux-2.2.1-ac3 and egcs-19990131

Marc Espie espie@quatramaran.ens.fr
Mon Feb 8 16:38:00 GMT 1999


In article < 19990208150407.A29197@cygnus.com > you write:
>Yes, -Winline should probably be implied by -Wall.

NO! Please don't do that. This is going to break (potentially) perfectly
good code that's using -Wall -Werror.

inline is a *hint* to the compiler, code whose semantics change when inline
stops getting inlined is bogus, and all the angry penguins of the world
won't change that.

Does egcs warns when it can't put variables in registers ?


More information about the Gcc-patches mailing list