Created attachment 158515 [details]
Screen shot of editor displaying pointed Hebrew incorrectly.

Bidi text analysis (logical ordering, boundary analysis) does not work properly. Attached is a screen shot of a small XML file containing both English (left-to-right) and Hebrew (right-to-left) text.

The first <hebrew> tag text shows the Hebrew in correct display order. However, text selection does not work properly. If one tries to advance the text caret with control-right-arrow and control-left-arrow, the caret becomes stuck in the Hebrew. Regular right- and left-arrow movement proceeds visually. Shift-right-arrow and shift-left-arrow seem to select visually, but it is impossible to select the entirety of the Hebrew text in that manner. Dragging with the mouse allows one to select all the text, but the selection appears sometimes to work on the text in logical order and sometimes in visual order, depending on where one starts dragging. Shift-clicking with the mouse also appears to operate in visual order except for the first character, which ends up being excluded.

The second <hebrew> tag has severe display problems. The only difference between the two <hebrew> lines is the addition of vowel marks. The text is supposed to be (hopefully your browser will display this correctly):

זֶהוּ חֵלֵק עִברִי

What actually appears is the letters laid out left-to-right instead of right-to-left. The good news is that caret movement using arrow keys doesn't get stuck.

I should also mention that these rendering and caret movement problems are independent of the font. The attached image shows the editor using Times New Roman for XML text and the default Monospaced font for everything else. Finally (and this perhaps could be considered a separate but), the Monospaced font appears to be missing most of the Hebrew combining mark glyphs. The pointed text still appears (incorrectly) in left-to-right order.