Re: [Alephmodular-devel] Autoconf committed

[Br'fin <br...@ma...>]

On Monday, January 6, 2003, at 08:46  PM, Alexander Strange wrote:

>
> On Monday, January 6, 2003, at 08:34 PM, Br'fin wrote:
>> I am curious as to what you have in mind for cleanup within 
>> render.cpp though.
>>
>> -Jeremy Parsons
>
> There's large amounts of old #ifdefed out code, commented-out code (in 
> scottish_textures too), ugly and badly-formatted code, misspelled 
> comments, etc. And a few inefficiences where it calls memset() when an 
> inlined loop would do the same thing faster  (cross-library calls are 
> extremely slow in OS X), although this really counts as an 
> optimization.

On Render.cpp

Most of the #ifdef'd code that I see is under the auspices of 
QUICKDRAW_DEBUG. As such, I am inclined to leave it in under the 
assumption that debugging code whether on or off is potentially useful, 
Especially if someone knows what they're doing in there and needs to 
know.

I have not done a similar review of commented out code. But I'd be 
applying the same kind of decisions. Truly obsolete code? It's out of 
there. Disabled but useful in some fashion? for instance Debugging code 
or code that appears to relate directly to M1 behaviour, may stay.

On memset: leave it alone for now. Right now we have most of the 
original memset's still in place. If we're going to do anything about 
them, then I would rather it be done uniformly and across the board. 
Loren's obj_clear and other template based calls at the end of 
csmacros,h hold a certain appeal to me in this regard. ( Even if they 
themselves are implemented in terms of memset and memcpy. :) )

scottish_textures.cpp:

Yeah it does look like it has some nicely labeled obsolete code at the 
end. Mmm, I wonder what's so nicely unique about the one that's labeled 
obsolete and 'unnecessary, but do not delete' :)

-Jeremy Parsons