Re: [PATCH] mask ADT: new mask.h file [2/22]

[Matthew Dobson <colpatch@us.ibm.com>]

On Mon, 2004-03-29 at 16:27, Paul Jackson wrote:
> Matthew wrote (of my recommendation to not use the mask type directly):
> > Is this necessary, or just convenient?
> 
> Technically as you suspect, just convenient, except in the case of the
> mask_of_bit() macro, as you observe.
> 
> I was adhering to the K.I.S.S. school here - just tell the user one
> recommended way of using things, suppressing my engineering urge to
> explain alternatives that had no real advantages.

That's what I figured.  Just looking for a clarification.


> > I think that it wouldn't be terribly ugly to split out the 1 unsigned
> > long special cases (bitmap_and, bitmap_or, etc) with #ifdefs.
> 
> Do you have in mind an ifdef per function, or putting
> several functions inside an ifdef?  If you think it
> looks better - show us the code ;).

I was thinking of having a large ifdef'd section for the one word case. 
I'll run that up and see if it does in fact look any cleaner.


> Here's my cumulative patch of changes that I have gained so far
> from your excellent feedback, and a couple I've noticed:

<diff snipped>

Looks good.  I'll keep chugging along through the patches! :)

-Matt