Historically, this function could *happen* to create bitmaps of
0x0 or even -1x-1 sizes and return a valid DynamicSprite
object. Try to ensure the image library is passed valid
parameters in such case, and keep going, for backwards compatibility.
From upstream 82194b261b4649281742c45498a4c34f452fd350

…pes"

From upstream 8d1751508b7f7d9925931144de1e3cb8ed1d2f51

* Fixed _lastAliveTs was not reset right after calling sys_evt_process_pending();
this caused events polling called continuously after timeout happened once.
* Set timeout to 60 fps (16.777 ms) instead of random 10ms.
* Do the timeout test only each 1000th loop iteration.
* Few other tiny adjustments in code.

Partially from upstream 60aaef003021fec2f75d4dc6db1b24f0b5d2be65

From upstream 57e866411b77af9e3a942bccf4a8c59bfa3d12cd

From upstream dd50a9976f77d504952171209c1597483318e1c2

…exture

This mistake was introduced early on in 7ed5186 and not fixed still in the later fix 38d0dd3
From upstream 1f27837ef14585d8050a4b49b46faa002f87a881

…e data

This complements the original idea, but also as a side-effect fixes a mistake introduced by another hotfix 38d0dd3.
Because the ID remained intact in the texture, the engine would not know to stop using it and replace after a dynamic sprite was deleted.
From upstream 87cc8e099169b82c2393282921e4d8295af4d0a9

From upstream c694126d77503b9d5528bd95427a8ef6ee48aa6c

Partially from upstream ad45540abe6c776574e02e6e57fba7758410ec40