Snippet for handle recreation support in GC via operation pattern

[HeikoKlare]

This ticket has been reduced to adding a snippet (see comments). Previously the intent was as follows:

Description
After having implemented the multi-zoom support in GC via storing GC executions as operations that can be applied to handles for other zooms afterwards, it should not be possible to draw something with a GC (onto a control/image) without an ImageGcDrawer) but still have a sharp rendering on whatever zoom is required for the context later on.

We should evaluate if this properly works. We know at least two cases in which this can be tested:

• The BufferedGraphicsSource in GEF.
• The double buffering used in the WindowBuilder project (see the discussion around Double-buffering with recent HighDPI changes eclipse-platform/eclipse.platform.swt#2159 (reply in thread)). We have implemented a fix in SWT to make it working again, which could be reverted to test if it not works because of the operation pattern in GC.

This is probably something @akoch-yatta or @HeikoKlare have to test.

[akoch-yatta]

I had a short test run with WindowBuilder yesterday and did not find any (new) issues

[HeikoKlare]

I just intensively tested again with WindowBuilder again and did not find any new issues as well. I even tested against older states of WindowBuilder, in particular to cover the changes made during the 4.36 Eclipse release because of the changes to image, around this discussion: eclipse-platform/eclipse.platform.swt#2159 (comment)

In my understanding, only remaining thing is that BufferedGraphicsSource does obviously not work properly yet (used by GEF when DOUBLE_BUFFERING is not enabled for the Canvas of a LightWeightSystem). Do we want to look for a solution for that, @akoch-yatta? At least, we might document the underlying issue (at and example) to make it comprehensible for more developers (and even ourselves) and to be able to determine where the responsibility for fixing the behavior would be.

[akoch-yatta]

I don't think the effort necessary is worth it - at least unless there are more places affected by this. In GEF the BufferedGraphicsSource is not used that much anymore anyway. I think adding a Snippet plus explanation would do it for me for now.