[Freeciv-Dev] Re: Map coordinate cleanups.

[Kevin Brown <kevin@xxxxxxxxxxxxxx>]

Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx> wrote:
> I haven't looked at the assembler output but is possible that the
> compiler removes the loop completely. The actual execution times
> however are too large for a body without the loop.

I'm quite sure the compiler didn't optimize the loop out, otherwise
execution would have been instantaneous.  In other words, what you
said.  :-)

> > > Besides, compilers can't optimize functions as well as inlined code.
> > 
> > Is this true of inlined functions?  I have suspicions that it is.  I'm
> > not sure how to devise a testcase for this...
> 
> AFAIK the current compilers can only inline methods if the compiler
> sees the method body at compile time. This leads to such ugly header
> files as seen in some C++ code.

Not much difference between that and macros, is there?  Especially
macros of the variety we're talking about.  There's no problem with
namespaces, either: just declare the functions static in the headers
(but we all knew this already, so why am I saying it?  :-).

I'd argue that, even if the function is in the header, it's *still*
cleaner looking than the equivalent macro (for multiline macros, at
least), if only because the macro has to have a continuation character
after each line.

This stuff only becomes an issue when you're dealing with a compiler
that can't do inline functions.  I doubt there are many of those that
matter anymore...



-- 
Kevin Brown                                           kevin@xxxxxxxxxxxxxx

    It's really hard to define what "unexpected behavior" means when you're
                       talking about Windows.