[win32] Introduce ZoomContext in Image

[akoch-yatta]

This PR extends the internal zoom used in Image in the windows implementation to a complex object. This is preparatory work to provide a consistent behavior independent of the auto scale mode, e.g. when an Image is drawn with a GC.

The introduced ZoomContext is only used in one place with two different values for targetZoom and nativeZoom, in Image#applyUsingAnyHandle. This is the correct way to fill ZoomContext here, but the zoom does not matter for the usages of this method anyway.

The complex ZoomComplex object introduced in Image is a first step in unifying zoom semantics. We noticed that depending on the configured autoscale mode, different zoom calculations must be used. This led to multiple regressions in the previous releases. One simple example is the font zoom. In most cases it is equal to the native zoom of the monitor or primary monitor. With a fixed zoom, it should use the fixed zoom instead.
The goal is to move this logic into the (probably windows only) ZoomContext object, that will provide an abstract layer for this and hopefully improves testability as well.

[github-actions]

Test Results

   546 files  ±0     546 suites  ±0   31m 9s ⏱️ + 3m 28s
 4 425 tests ±0   4 408 ✅ ±0   17 💤 ±0  0 ❌ ±0 
16 746 runs  ±0  16 619 ✅ ±0  127 💤 ±0  0 ❌ ±0 

Results for commit 16da749. ± Comparison against base commit 51d6029.

♻️ This comment has been updated with latest results.

[HeikoKlare]

Thank you for the latest updates of the PR!
The change is sound and adds the necessary capabilities for properly distinguishing native and target zoom to the Image class. This change alone (without the follow-up changes) does not have any user-perceivable effect. I tested with different auto-scale settings (monitor-specific on/off, fixed auto-scale) and did not find any changes in behavior.