[FFmpeg-devel] [PATCH] Fix clang warning in libavfilter/vf_unsharp.c

[Måns Rullgård]

Eli Friedman <eli.friedman at gmail.com> writes:

> Patch attached; clang warns on " =- " because it looks like a typo for " -= ".
>
> -Eli
>
> Index: libavfilter/vf_unsharp.c
> ===================================================================
> --- libavfilter/vf_unsharp.c	(revision 23827)
> +++ libavfilter/vf_unsharp.c	(working copy)
> @@ -83,9 +83,9 @@
>      for (y = 0; y < 2 * fp->steps_y; y++)
>          memset(sc[y], 0, sizeof(sc[y][0]) * (width + 2 * fp->steps_x));
>
> -    for (y =- fp->steps_y; y < height + fp->steps_y; y++) {
> +    for (y = -fp->steps_y; y < height + fp->steps_y; y++) {
>          memset(sr, 0, sizeof(sr[0]) * (2 * fp->steps_x - 1));
> -        for (x =- fp->steps_x; x < width + fp->steps_x; x++) {
> +        for (x = -fp->steps_x; x < width + fp->steps_x; x++) {
>              tmp1 = x <= 0 ? src[0] : x >= width ? src[width-1] : src[x];
>              for (z = 0; z < fp->steps_x * 2; z += 2) {
>                  tmp2 = sr[z + 0] + tmp1; sr[z + 0] = tmp1;

OK, no matter what anyone else might say.  Writing it like that is
nothing short of crazy.

Aside, the first version of K&R C actually used =+, =- and so on for
increment, decrement, etc.  It was changed due to the confusion it
could cause.

-- 
M?ns Rullg?rd
mans at mansr.com