PGS: RLE Pixel Count vs Dimensions

[wswartzendruber]

Does anyone know of any professionally-produced PGS streams with objects containing line counts that don't agree with the object's declared dimensions?

Everything in my collection has [width] number of pixels per line and [height] number of lines, but I'm not sure if that's always the case.

[cubicibo]

In an epoch, once you define an object ID, the provided dimensions are used to do memory allocation and that memory layout shall remain until the end of the epoch.
While you could provide RLE data that decode to an object with smaller width and height than the target memory slot, you end up with undefined behaviour. The outcome depends solely on the vendor implementation of the decoder in the said case.
E.g, the composition controller could copy the object according to the defined slot dimensions, displaying residual object data not erased by the new, smaller, object.

Beside, how would you compute the decode duration in this case? To the current object shape or the buffer?

[wswartzendruber]

I may as well call this work-in-progress Cubicibo.PGS4NET.dll.

I won't do that, of course. But if this project goes anywhere it will have been because of you.