fix: prevent strip_codes from matching across OSC 8 sequences

Changes the OSC sequence regex from .*? to [^\x1b]* to prevent
matching across escape sequence boundaries. This fixes a bug where
text between OSC 8 hyperlinks was incorrectly stripped in terminals
with proper OSC 8 support (e.g., Ghostty).

The new pattern stops at the first escape sequence terminator rather
than potentially continuing to a later one, preventing over-matching.

Fixes two failing tests:
- test_strip_codes_preserves_text_between_osc8_hyperlinks
- test_strip_codes_with_ui_link

Author: bostonaholic

lib/cli/ui/ansi.rb

@@ -42,7 +42,9 @@ def printing_width(str)
         #
         #: (String str) -> String
         def strip_codes(str)
-          str.gsub(/\x1b\[[\d;]+[A-Za-z]|\x1b\][\d;]+.*?\x1b\\|\r/, '')
+          # Strip SGR (color) codes, OSC sequences (including hyperlinks), and carriage returns
+          # Use [^\x1b]* instead of .*? to prevent matching across escape sequences
+          str.gsub(/\x1b\[[\d;]+[A-Za-z]|\x1b\][\d;]+[^\x1b]*\x1b\\|\r/, '')
         end
 
         # Returns an ANSI control sequence