-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Darren New wrote:
> Christopher Smith wrote:
>> The "inline" keyword is strictly a hint, much like the "register"
>> keyword.
> 
> Yes. But if I give you a .h and a .o and no .c, you're going to have a
> heck of a time writing an optimizer that figures out what's in the .o
> file and knows it can skip some of the code, yes?

Again, that's not how a bounds checking iterator or null checking smart
pointer template would be shared though, because you can't instantiate
types from a .o.

That said, you can indeed have a linker that does inlining and dead code
elimination. Indeed a lot of the commercial linkers do that right now
(not sure about GNU ld.. I'm guessing not). See the Herb Sutter article
for more on just how often one has an opportunity to inline.

- --Chris
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGcitROagjPOywMBARAgPPAJ0cJTkxdd4R13z2o0hIXEx9+j3qOQCgvD0V
AhloDmjB6xVvqr4O61ItYHc=
=uBUm
-----END PGP SIGNATURE-----

-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Reply via email to