P4
Eva points out that:
1. The spec says that stops can by specified either with percentage or with values (in px, em etc.). In case of absolute values, the javafx Stops object are directly created with those numbers, but in reality only values between 0 and 1 are supported by the Stop object. This doesn't fail, but the created gradient has incorrect stops. I believe the values for Stops should be recalculated based on gradient dimensions (radius in case of radial gradient, from-to points in case of non-proportional linear gradient).
2. radius in radial gradient is said to be mandatory in our CSS spec. However the parser sometimes doesn't handle is at such. This results in NPE when you don't specify radius in radial gradient style. I believe this should end with some meaningful error (or the radius shouldn't be mandatory, I don't know if this is better).
1. The spec says that stops can by specified either with percentage or with values (in px, em etc.). In case of absolute values, the javafx Stops object are directly created with those numbers, but in reality only values between 0 and 1 are supported by the Stop object. This doesn't fail, but the created gradient has incorrect stops. I believe the values for Stops should be recalculated based on gradient dimensions (radius in case of radial gradient, from-to points in case of non-proportional linear gradient).
2. radius in radial gradient is said to be mandatory in our CSS spec. However the parser sometimes doesn't handle is at such. This results in NPE when you don't specify radius in radial gradient style. I believe this should end with some meaningful error (or the radius shouldn't be mandatory, I don't know if this is better).