Fix generating skeleton (#20224)
Fixes #20105
Summary of the issue:
markdownTranslate.py previously removed inline markdownlint comments during preprocessing, so those comments were not preserved in generated XLIFF skeleton content. This could break round-tripping and comment fidelity for markdown files that embed inline lint directives.
Description of user facing changes:
No direct end-user NVDA feature change. This affects markdown/XLIFF tooling behavior used in docs workflows: inline markdownlint comments are now retained in skeleton output.
Description of developer facing changes:
Removed preprocessMarkdownLines usage across markdown/XLIFF processing paths.
Updated inline markdownlint regex handling to capture and preserve inline markdownlint comments in skeletonizeLine.
Added getSkeletonContentFromXliffText to avoid over-aggressive .strip() on skeleton text.
Added a new fixture file:
markdownlint_inlineComments.md
Added unit coverage:
test_generateXliff_preservesInlineMarkdownLintCommentsInSkeleton
Uses module loading + patching of getRawGithubURLForPath to support file URI in test context.
Description of development approach:
This PR was heavily assisted through CoPilot.
The change removes line preprocessing that was stripping semantic inline comments and instead handles markdownlint comments at skeletonization time so comments remain in output. Skeleton text extraction was adjusted to preserve meaningful whitespace/newline boundaries. Test strategy was expanded with a dedicated fixture and targeted assertion of skeleton content.
