[Scummvm-cvs-logs] CVS: scummvm/scumm/smush codec37.cpp,1.10,1.11
Max Horn
max at quendi.de
Sun Mar 2 15:23:40 CET 2003
- Previous message: [Scummvm-cvs-logs] CVS: scummvm/scumm/smush codec37.cpp,1.10,1.11
- Next message: [Scummvm-cvs-logs] CVS: scummvm/backends/dc dc.h,1.7,1.8 dcmain.cpp,1.5,1.6 input.cpp,1.5,1.6 time.cpp,1.1,1.2
- Messages sorted by:
[ date ]
[ thread ]
[ subject ]
[ author ]
At 23:35 Uhr +0100 02.03.2003, Marcus Comstedt wrote:
>Max Horn <max at quendi.de> writes:
>
>> Uh, that's pretty ugly and hard to maintain...
>
>Considering how the _rest_ of the code looked, that didn't seem to be
>a major concern in this file... :-)
>
>
>> I already see the typos
>> sneaking in whenever we have to fix something here. Why not use clean
>> simple nested look for the SCUMM_NEED_ALIGNMENT variants? Any
>> reasonable compiler should automatically unroll them, shouldn't it?
>> I.e something lie
>>
>>
>> t = *src++;
>> #if defined(SCUMM_NEED_ALIGNMENT)
>> for (int i = 0; i < 4; i++)
>> for (int j = 0; j < 4; j++)
>> (dst + pitch * i)[j] = t;
>> #else
>> ...
>> #endif
>
>I think the whole file could use a cleanup. There's the same code
>cut'n'pasted four times all over the place. Maybe a macro that's
>instansiated four times to get the different variants instead?
All be fine by me :-)
Max
- Previous message: [Scummvm-cvs-logs] CVS: scummvm/scumm/smush codec37.cpp,1.10,1.11
- Next message: [Scummvm-cvs-logs] CVS: scummvm/backends/dc dc.h,1.7,1.8 dcmain.cpp,1.5,1.6 input.cpp,1.5,1.6 time.cpp,1.1,1.2
- Messages sorted by:
[ date ]
[ thread ]
[ subject ]
[ author ]
More information about the Scummvm-git-logs
mailing list