OK, committed then. I have yet to study yet the other part of the proposal regarding diatritical characters. Patch 3 has a conflict for the regression tests, so a rebase would be needed. That's not a big deal though to resolve the conflict. I am also a bit confused by the newly-generated unaccent.rules. Why nothing shows up for the second column (around line 414 for example)? Shouldn't we have mapping characters?
That concerned me, as well. I have confirmed the lines are not empty. If you open the file in a text editor (I'm using KDE's Kate), and insert a standard character at the beginning of one of those lines, the diacritic then appears, combined with the character you just entered. The only program I've found that wants to display them on their own is vi (and I only just thought of trying that).
From what I can tell, this is likely a font issue:
