796174 - Don't use high-quality scaler for non-downscales

Status: RESOLVED FIXED

(Core :: Graphics: ImageLib, defect)

Description

[Joe Drew (not getting mail)]

Attachment: ensure only downscales are scaled

Right now we do a high-quality scale on unity scales. This is obviously wrong.

Comment 1

[Justin Lebar (not reading bugmail)]

Comment on attachment 666747 [details] [diff] [review]
ensure only downscales are scaled

Heh.

I'm not sure what the extra function is buying us, but either way is fine.  If you keep the function, can you mark it static or put it in an anonymous namespace?

Comment 2

[Joe Drew (not getting mail)]

I couldn't think of a clearer way to express "Only downscales" than having three conditions, and IME having a separate function is helpful when you start adding tons of conditions.

Comment 3

[Justin Lebar (not reading bugmail)]

(In reply to Joe Drew (:JOEDREW! \o/) from comment #2)
> I couldn't think of a clearer way to express "Only downscales" than having
> three conditions, and IME having a separate function is helpful when you
> start adding tons of conditions.

I mean we already have a function saying "are we going to downscale this?", so if you wrote it as

 bool
 RasterImage::CanScale(gfxPattern::GraphicsFilter aFilter,
                       gfxSize aScale)
 {
 // The high-quality scaler requires Skia.
 #ifdef MOZ_ENABLE_SKIA
   if (!gHQDownscaling || mAnim || !mDecoded ||
       aFilter != gfxPattern::FILTER_GOOD ) {
     return false;
   }

   if (scale.width > 1.0 || scale.height > 1.0) {
     return false;
   }

   if (scale.width == 1.0 && scale.height == 1.0) {
     return false;
   }

   gfxFloat factor = gHQDownscalingMinFactor / 1000.0;
   return (aScale.width < factor || aScale.height < factor);
 #endif

That would be plenty clear to me.

Comment 4

[Joe Drew (not getting mail)]

http://hg.mozilla.org/integration/mozilla-inbound/rev/629c399c6f8c

Comment 5

[Joe Drew (not getting mail)]

Backed out: https://hg.mozilla.org/integration/mozilla-inbound/rev/90332928696a
and relanded: https://hg.mozilla.org/integration/mozilla-inbound/rev/6925601f4299

Comment 6

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/6925601f4299