[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: PuTTY Testing
- To: Development Discussions <developer at arabeyes dot org>
- Subject: Re: PuTTY Testing
- From: Nadim Shaikli <shaikli at yahoo dot com>
- Date: Tue, 23 Mar 2004 21:26:04 -0800 (PST)
--- ahmad khalifa <ahmadkhalifa at hotmail dot com> wrote:
> at last, i can say that PuTTY can does its own shaping, and Bidi.
> Bidi is done by my Implementation of the Bidi algorithm, i have
> updated it to do Mirroring, and dont seem to have any bugs left.
> (although some might arise from mirroring).
Here are a few things I've encountered that I think should be fixed
before we push for the patch's adoption.
1. Do some repetitive 'ls -lF' and 'clear', putty crashed on me a
number of times (there might be a memory leak somewhere).
2. If you view that favorite test file 'ar_utf8' there is a string
in there that notes, "Here are some Random words:" for whatever
reason the ":" dances around and changes location. 'cat' the
file (or less it) and then scroll back to that line (towards
the top to see what I mean).
BTW: that line is not shown correctly (ie. I don't think Bidi
is doing the right thing there.
3. The LAM+ALEF which gets combined into a LAA still exhibits a
space after it. That space should really be visually gone.
4. Do please add 2 buttons in Terminal->Features that notes,
a. Disable Shaping
b. Disable Bidi
> anyway, i tested it and seem to get 100% Bidi Compliance,
> anyone care to correct me ?
The visual look of raw data looks semi-clean (note above).
I think once things are settled looking into Control-Characters
will be warranted in order to properly place cursors and/or string
search boxes, etc (that's for much later on).
> there are 2 patches in cvs, 'arabic_putty.patch' which does
> shaping and flipping(no BIDI). and 'putty.arabic' which does
> shaping and Bidi.
There is no reason to upkeep the one that doesn't do Bidi
(ie. we can get rid of it). Beyond the files that are in CVS
now (which is fine), we really ought to have ONE patch file
which we submit which would encompasses all the new files
within (nothing major, but thought I'd mention it again).
> in cvs you'll also find my implementation of the Bidi Algorithm
> mentioned above, its akbidi.h and akbidi.c (bad name huh?)...
Not a fan of the name, but do please run the tests on your code
that were mentioned before to see how it compares to the other
algorithms. Once that is done, I think we should maintain this
new Bidi code (and the shaping) as part of their own mini-project.
> nadim: how about another ping for Simon...
Sure, I have no problem pinging (do also CC your posts to get them
moving as well). I do think that we need to solve the above 3-4
points before proceeding to them (or at least have an idea on what's
involved) - as the next thing we should do is to get the patch adopted
(bug fixes and other issues (major and minor) to be ironed out post
patch acceptance). What PuTTY now does is what we initially set out
to do which is to get it simple display (ie. 'cat') Arabic files without
Ahmad, really really really great work.
Do you Yahoo!?
Yahoo! Finance Tax Center - File online. File on time.