Dov Feldstern wrote:
Hi, Abdel!
I appreciate it that you're trying to keep your eye out for RTL issues!
Unfortunately, in this case I think that the fix is wrong --- it does
more harm than good. Attached is a patch to reverse this changeset
OK, I now understand what was the problem and modified the code accordingly.
[...]
Basically, the problem is that an RTL paragraph inside an inset is
painted to the left of the inset, instead of inside it. The problem is
only if the paragraph itself is RTL --- i.e., RTL text within an LTR
paragraph inside an inset is painted correctly; and also, the problem is
only with a single-line paragraph: as soon as the paragraph (inside the
inset) becomes multi-line, the problem goes away. So I suspect this has
something to do with the "wide" stuff?
More or less yes. This is fixed now.
Note that there's also a problem with the cursor placement --- it's not
placed correctly either relative to the text, nor relative to the inset
--- but I think it will be easier to work this problem out after the
text itself is correct.
This is the next item on my agenda, but feel free to fix it before me ;-)
In the attached file there are also a few other problems --- various
crashes when moving the cursor around (e.g., try just pressing the down
key from the top of the document) --- but I assume these are all due to
the metrics issues, and that once those are fixed, then they will
disappear...
Should be fine now, please report any further problems.
Thanks,
Abdel.
