On Jul 28,  3:13pm, mycroft@mit.edu ("Charles M. Hannum") wrote:
-- Subject: Re: Problems with lint(1)

[stuff deleted]

| However, this is all moot, since I've fixed the bug in lint(1) (and
| another one, too).

Thanks for fixing this :-) My suggestion was just avoiding the real
cause of the problem.

| I think that fixes all the actual problems in lint(1) that I've run
| into so far -- although it would be nice if it knew about
| __attribute__((noreturn)).

Actually it would be better to use ___attribute__((__noreturn__)) so
that we do not run into namespace issues.

That might be a worth-while goal; we need to make a clear decision
about implementing __attribute__ directives in lint:

	- do we implement all/any supported gcc attributes?
	- how do we deal with ARGSUSED/NOTREACHED? It would be nice to
	  be consistant and use a unified syntax for all lint annotations
	  instead of having a mixture of lint comments and __attribute__
	  statements.

| There is still the varargs wart to deal with, and I noticed that
| exports symbols in crt0 and libm are not defined anywhere, but it's
| actually possible to build some lint-free code now.

These symbols need to be exported publicly, but then we need to deal
with namespace problems again.

christos