bug#69992: Minor improvement to image map transformation logic

[David Ponce]

Many thanks for this feature, which is particularly useful to
automatically recalculate the map of computed images like SVG.

To make the code faster, by avoiding multiple scans of the map for
copy and parsing, I propose the following patch which factors most of
the code into the functions `image--compute-map' and `image--compute
-original-map'. I have done some tests on my side which are
conclusive.

Furthermore, I wonder if the term :base-map would not be more
descriptive than :original-map?

Thanks again for this valuable feature!

image.el.patch
Description: Text Data