Unicode Escapes handled incorrectly

[Yorwba]

I do not have TextMate to reproduce this locally, but the grammar is also used by GitHub's syntax highlighting, right?

In this comment thread on Hacker News I learned that Java expands Unicode escapes like \u002F\u002A (for /*) everywhere.

This means that code like

someService.addTrustedKey(SecuritySettings.getTrustedKey("SomeService"));
/* For local debugging:
  // Copy this into your client config: 
  DBG_KEY=\u0022\u0029\u003B\u002F\u002A\u002F
  someService.addTrustedKey("\u0022\u0029\u003B\u002F\u002A\u002F");
*/

is equivalent to

someService.addTrustedKey(SecuritySettings.getTrustedKey("SomeService"));
/* For local debugging:
  // Copy this into your client config: 
  DBG_KEY=");/*/
  someService.addTrustedKey("");/*/");
*/

but is highlighted differently (thanks to HN user mistercow for the example).

I hope it is clear that this difference in highlighting is pretty deceiving when doing code review, for example. A complete fix would require replacing all literal characters in regular expressions (e.g. \*) by something like (\*|\\u002[aA]) and it gets worse for \s etc. This could probably be done by some kind of preprocessor for the syntax, but someone would have to write it.

A simpler temporary fix would be to mark Unicode escapes of ASCII characters as invalid everywhere, to highlight to the unsuspecting that something fishy might be going on.