Re: [putty] Broken Bidi in PuTTY

[Jacob Nevins <jacobn+puttybugs at chiark dot greenend dot org dot uk>]

Simon Tatham writes:
> Nadim Shaikli <shaikli at yahoo dot com> wrote:
> 
> > Simon, I just downloaded the "latest development snapshot" PuTTY binary [1]
> > and it looks like its Bidi support is either broken or disabled for some
> > reason (my guess is that it was mistakenly disabled on purpose).
> 
> I haven't done anything that _should_ have broken it, so if this is
> true then it's certainly a bug that needs fixing.
> 
> >  + 2004-12-02:r2941  Works great without issues
> >  + 2004-12-19:r5011  Seems to be broken
> 
> (Presumably you mean 2004-12-02:r4941 :-)
> 
> > I simply looked at a mixed Arabic/English file and it looks like the Arabic
> > is being rendered properly its just not being flipped so instead of reading
> > HELLO, I'm seeing OLLEH (given caps are Arabic).
> 
> When I try displaying my small UTF-8 test file (currently checked in
> to the PuTTY source repository as putty/testdata/utf8.txt) I get
> nothing but rectangular blobs, but the blobs appear to be arranged
> in the right order as far as I can see (cursor position mapping and
> selection all appear to do the right thing). Do you see the problem
> with that file? If not, can you send me a file that _does_ exhibit
> the problem, and I'll add it to my test collection.

The behaviour when cat'ing the attached file (derived from
testdata/utf8.txt) has definitely changed between 2004-12-02 and the
current code; more precisely, the behaviour changes at r4964 (rewrite of
getType()). I wouldn't like to say for sure which is right (since I
can't even see the glyphs), but I'm guessing the previous behaviour.

This file has a line which is predominantly LTR but contains RTL
sequences (unlike the utf8.txt lines which are all predominantly RTL).
We should probably add something like it as a test case.

Mixed LTR and RTL text: جرير رضي back to LTR.