On Mon, 2010-05-03 at 10:25 -0700, Marek Olšák wrote:
> José,
>
> the first patch removes the PIPE_FORMAT_ prefix in a string returned
> by util_format_name, it makes debug logs shorter. The second patch
> adds util_format_is_plain.
>
> diff --git a/src/gallium/auxiliary/util/u_format.h
> b/src/gallium/auxiliary/util/u_format.h
> index fb6ade5..d851c31 100644
> --- a/src/gallium/auxiliary/util/u_format.h
> +++ b/src/gallium/auxiliary/util/u_format.h
> @@ -332,10 +332,10 @@ util_format_name(enum pipe_format format)
>
> assert(desc);
> if (!desc) {
> - return "PIPE_FORMAT_???";
> + return "???";
> }
>
> - return desc->name;
> + return desc->name+12;
> }
Just return desc->short_name. It's lower case so it's easier to read.
Also please modify src/gallium/drivers/trace/tr_dump_state.c to use the
full caps name, as retrace relies on it.
> static INLINE boolean
>
> diff --git a/src/gallium/auxiliary/util/u_format.h
> b/src/gallium/auxiliary/util/u_format.h
> index d851c31..350b817 100644
> --- a/src/gallium/auxiliary/util/u_format.h
> +++ b/src/gallium/auxiliary/util/u_format.h
> @@ -338,6 +338,18 @@ util_format_name(enum pipe_format format)
> return desc->name+12;
> }
>
> +static INLINE boolean
> +util_format_is_plain(enum pipe_format format)
> +{
> + const struct util_format_description *desc =
> util_format_description(format);
> +
> + if (!format) {
> + return FALSE;
> + }
> +
> + return desc->layout == UTIL_FORMAT_LAYOUT_PLAIN ? TRUE : FALSE;
> +}
> +
This is fine.
I have no problem with adding more such functions here, but for the
record, a better practice would be to call util_format_description once
and then do necessary operations with it. And typically there are
several. That is I'd rather see all inline functions in u_format.h
taking as argument the struct util_format_description *desc, expect
util_format_description().
> static INLINE boolean
> util_format_is_s3tc(enum pipe_format format)
> {
>
> May I push them?
Jose
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
