Re: [RFC] Alternate approach to keeping convenience variables

[Jim Blandy <jimb@red-bean.com>]

On 1/22/06, Daniel Jacobowitz <drow@false.org> wrote:
> This version of the patch addresses all the comments I've received,
> except for:
>
>   - TYPE_ZALLOC.  I think this is more appropriate as a macro,
>     personally, though I wouldn't argue too loudly if someone wanted to
>     functionify it at the same time as TYPE_ALLOC.

That's fine.  I'll have to put my typing where my mouth is.

>     GDB developers seem to have developed an allergy to C macros that
>     I just don't understand.  They're not _inherently_ obfuscating
>     or evil!  They can be both useful and elegant.

Nobody commented on my creative use of macros in m32c-tdep.c.  If I
were allergic, that'd definitely put me in anaphylactic shock.

I don't like TYPE_ZALLOC and TYPE_ALLOC because they don't get
anything special from being macros.  These aren't time-critical bits
of code.  They don't construct names.  They don't take a type as an
argument (err, a compile-time type, that is).  They don't expand to
statements.  They don't export an interface whose presence you'd like
to be able to test for easily at compile-time.  So why not get the
nicer syntax, type checking, and debuggability that functions provide?

De gustibus non est disputandum.