Discrepancy between geom_point2 and geom_point_borderless (and similar)

[strengejacke]

Fixes #395

[gemini-code-assist]

Code Review

This pull request introduces new 'halo' geoms, such as geom_point_halo and geom_jitter_halo, which provide a visual separation for overlapping points by adding a contour that matches the plot's background color. Additionally, it refactors *_borderless geoms as aliases and includes bug fixes for their rendering. Review feedback highlights an inaccuracy in the documentation regarding color aesthetics for the refactored geoms, a typo in the news file, and potential issues in the internal halo logic related to shape conflicts and background color detection.

[gemini-code-assist]

Typo in the function name alias: *_borless should be *_borderless.

Suggested change

	* Fixed the `*_borderless` aliases for `geom_point2()` or `geom_jitter2()`, which
	* Fixed the `*_borderless` aliases for `geom_point2()` or `geom_jitter2()`, which

[gemini-code-assist]

There are two improvements possible here:

1. Shape Conflict: If a user passes shape or pch in ..., it will conflict with the hardcoded pch = 21. It is safer to explicitly remove these from dots to ensure the halo effect (which requires a shape with a border and fill) is preserved.
2. Background Color Logic: The .get_theme_bg_color() function currently retrieves the grid line color (panel.grid.major$colour) rather than the actual background color (panel.background$fill). This might result in a halo that doesn't blend with the background if they differ (e.g., grey grid lines on a white background).

  dots$colour <- dots$shape <- dots$pch <- NULL
  dots$color <- .get_theme_bg_color()

  fun_args <- c(list(pch = 21), dots)