Fixes#12338, related to #37616
This change improves URL detection in the terminal by removing trailing
periods that appear to be sentence punctuation rather than part of the
URL structure. It builds upon the parentheses sanitization work from
#37076 by consolidating both approaches into a unified
`sanitize_url_punctuation` function.
## Changes
- Combines parentheses and period sanitization into a single
`sanitize_url_punctuation` function
- Uses optimized single traversal with `fold()` for parentheses counting
(addressing code review feedback)
- Removes trailing periods using heuristics to distinguish sentence
punctuation from legitimate URL components
- Removes multiple trailing periods (always considered punctuation)
- Removes single trailing periods when they appear after alphanumeric
characters or slashes
- Preserves periods that are part of legitimate URL structure (e.g.,
version numbers, IP addresses, subdomains)
- Maintains existing parentheses balancing logic from #37076
## Implementation Details
- **Parentheses handling**: Counts opening and closing parentheses,
removes trailing `)` when unbalanced
- **Period handling**: Uses `take_while()` iterator for efficient period
counting
- **Performance**: Single pass counting with optimized loop to avoid
redundant work
- **Code clarity**: Uses let-else pattern for readable conditional logic
## Testing
- Added comprehensive test coverage for both parentheses and period
sanitization
- Tests cover balanced vs unbalanced parentheses cases
- Tests cover various period scenarios including legitimate URL periods
vs sentence punctuation
- All existing tests continue to pass
## Release Notes
- Improved terminal URL detection by further trimming trailing
punctuation. URLs ending with periods (like
`https://example.com.`) and unbalanced parentheses (like
`https://example.com/path)`) are now properly detected without including
the trailing punctuation.
Joseph Mearman
0ac1752668