New text fields should compute maxlength considering composed character sequences.
See comments in http://bugzilla.opendarwin.org/show_bug.cgi?id=6987
Bumping to P1 since this causes a text field regression
These are all text field regressions so they should all be P1.
The API for this is in ICU, in <unicode/ubrk.h>. If we use a UBRK_CHARACTER iterator, then we can find characer boundaries. I think I'll tackle this.
I've got the fix for this written; just need to write some layout tests. I also fixed a regression where changing maxlength would not immediately truncate the value that's already in the text field.
Created attachment 7587 [details]
patch, including change log and a layout test