fix: textfield masking and element transform positioning in session replay

[ioannisj]

💡 Motivation and Context

Closes #239

Fixes mask positioning for TextFields and elements under transforms (e.g. ScreenUtil scaling, rotations).

• Mask positioning: Switched from localToGlobal to getTransformTo(ancestor) using the screenshot container as the coordinate reference. This correctly handles ScreenUtil scaling, Transform.scale, Transform.rotate, and nested transforms.
• Added RenderEditableParser: TextField text is rendered by RenderEditable, whose layout bounds can be much smaller than the actual text (e.g. 1.3px vs 39px with isDense: true + ScreenUtil). This parser uses preferredLineHeight to calculate the correct mask height.
• TextField masking delegation: When maskAllTexts is enabled, TextField masking is now handled at the RenderEditable level (better bounds) rather than widget level. Widget-level masking only applies for obscureText fields when maskAllTexts is disabled.
• ImageMaskPainter transform support: Mask drawing now applies the element's Matrix4 transform, correctly positioning masks in the screenshot's coordinate space regardless of scaling or rotation.

💚 How did you test it?

• Added MaskingTestsScreen with test cases for text, images, TextFields, transforms, and PostHogMaskWidget.
• Verified with customer's MRE

📝 Checklist

• I reviewed the submitted code.
• I added tests to verify the changes.
• I updated the docs if needed.
• No breaking change or entry added to the changelog.

If releasing new changes

• Ran pnpm changeset to generate a changeset file
• Added the release label to the PR

[marandaneto]

use a typedef or a data class to the return type so you dont have to repeat this and its self documented

[marandaneto]

why don we check this anymore?