bug#5226: Malayalam conjunct formation breaks down at the end of long words

bug#5226: Malayalam conjunct formation breaks down at the end of long words

[Praveen A]

Hi,

Many Malayalam conjuncts at the end of long words are broken down to
its basic forms.

Example 1: നക്ഷത്രാവശിഷ്ടങ്ങളും or
'\u0d28\u0d15\u0d4d\u0d37\u0d24\u0d4d\u0d30\u0d3e\u0d35\u0d36\u0d3f\u0d37\u0d4d\u0d1f\u0d19\u0d4d\u0d19\u0d33\u0d41\u0d02'
(ങ്ങ is broken into ങ ് ങ)
Example2: എനിയ്ക്കിഷ്ടായില്ല or
'\u0d0e\u0d28\u0d3f\u0d2f\u0d4d\u0d15\u0d4d\u0d15\u0d3f\u0d37\u0d4d\u0d1f\u0d3e\u0d2f\u0d3f\u0d32\u0d4d\u0d32'
(ല്ല is broken into ല ് ല)

To reproduce this bug: Copy the above example to a buffer and observe
the deference in rendering when a space is added in the middle of the
word.

This is tested with CVS version of emacs with most recent Indic rendering fixes.

- Praveen
-- 
പ്രവീണ്‍ അരിമ്പ്രത്തൊടിയില്‍
<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)
Join The DRM Elimination Crew Now!
http://fci.wikia.com/wiki/Anti-DRM-Campaign

bug#5227: Malayalam conjunct formation breaks down at the end of long words

[Praveen A]

2009/12/16 Praveen A <pravi.a@gmail.com>:
> Hi,
>
> Many Malayalam conjuncts at the end of long words are broken down to
> its basic forms.

Screen capture video of the bug
http://blip.tv/dashboard/episode/2999711

- Praveen
-- 
പ്രവീണ്‍ അരിമ്പ്രത്തൊടിയില്‍
<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)
Join The DRM Elimination Crew Now!
http://fci.wikia.com/wiki/Anti-DRM-Campaign

Re: bug#5226: Malayalam conjunct formation breaks down at the end of long words

[Kenichi Handa]

In article <3f2beab60912160328y1084addbw7555543aaf2fcf0c@mail.gmail.com>, Praveen A <pravi.a@gmail.com> writes:

> Many Malayalam conjuncts at the end of long words are broken down to
> its basic forms.

> Example 1: നക്ഷത്രാവശിഷ്ടങ്ങളും or
> '\u0d28\u0d15\u0d4d\u0d37\u0d24\u0d4d\u0d30\u0d3e\u0d35\u0d36\u0d3f\u0d37\u0d4d\u0d1f\u0d19\u0d4d\u0d19\u0d33\u0d41\u0d02'
> (ങ്ങ is broken into ങ ് ങ)
> Example2: എനിയ്ക്കിഷ്ടായില്ല or
> '\u0d0e\u0d28\u0d3f\u0d2f\u0d4d\u0d15\u0d4d\u0d15\u0d3f\u0d37\u0d4d\u0d1f\u0d3e\u0d2f\u0d3f\u0d32\u0d4d\u0d32'
> (ല്ല is broken into ല ് ല)

> To reproduce this bug: Copy the above example to a buffer and observe
> the deference in rendering when a space is added in the middle of the
> word.

> This is tested with CVS version of emacs with most recent
> Indic rendering fixes.

Again, I cant' reproduce it.  Sigh...

Which font is it?

Please try this and show me the result of "log1" and "log2" files.

% MDEBUG_FLT=3 emacs NO_MIDDLE_SPACE.txt >& log1
% MDEBUG_FLT=3 emacs WITH_MIDDLE_SPACE.txt >& log2

Here NO_MIDDLE_SPACE.txt is a file containing only this line:
എനിയ്ക്കിഷ്ടായില്ല
and WITH_MIDDLE_SPACE.txt is a file containing only this line:
എനിയ്ക്കിഷ്ടായി ല്ല

> Screen capture video of the bug
> http://blip.tv/dashboard/episode/2999711

When I visit that page, login is requested.  What username
and password to use?

> ---
Kenichi Handa
handa@m17n.org

Re: bug#5226: Malayalam conjunct formation breaks down at the end of long words

[Praveen A]

[-- Attachment #1: Type: text/plain, Size: 1118 bytes --]

2009/12/17 Kenichi Handa <handa@m17n.org>:
> Again, I cant' reproduce it.  Sigh...

:-(

> Which font is it?

This bug is present with all fonts (Meera, Rachana, Lohit Malayalam...)

> Please try this and show me the result of "log1" and "log2" files.
>
> % MDEBUG_FLT=3 emacs NO_MIDDLE_SPACE.txt >& log1
> % MDEBUG_FLT=3 emacs WITH_MIDDLE_SPACE.txt >& log2

log files are attached.

>> Screen capture video of the bug
>> http://blip.tv/dashboard/episode/2999711
>
> When I visit that page, login is requested.  What username
> and password to use?

Sorry about that, the link I had given was not public. The public link
is http://blip.tv/file/2980014

I have also uploaded another screencast which shows the rendering
change when text is selected.

http://blip.tv/file/2984156

-Praveen
-- 
പ്രവീണ്‍ അരിമ്പ്രത്തൊടിയില്‍
<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)
Join The DRM Elimination Crew Now!
http://fci.wikia.com/wiki/Anti-DRM-Campaign

[-- Attachment #2: log1 --]
[-- Type: application/octet-stream, Size: 10417 bytes --]

[-- Attachment #3: log2 --]
[-- Type: application/octet-stream, Size: 9111 bytes --]

Re: bug#5226: Malayalam conjunct formation breaks down at the end of long words

[Kenichi Handa]

In article <3f2beab60912162328lb9a4bb1t15f5cd828b67b257@mail.gmail.com>, Praveen A <pravi.a@gmail.com> writes:

> > Please try this and show me the result of "log1" and "log2" files.
> >
> > % MDEBUG_FLT=3D3 emacs NO_MIDDLE_SPACE.txt >& log1
> > % MDEBUG_FLT=3D3 emacs WITH_MIDDLE_SPACE.txt >& log2

> log files are attached.

Thank you for them.  But the log shows that the Malayalam
rendering is not using the latest Emacs, at least, it seems
that the latest change for malayalam-composable-pattern is
not working.

Please show me the result of this:

(aref composition-function-table #xD15)

Mine is this:

(["\\([ക-ഹ]‌?്‌?\\)*[ക-ഹ]\\(്[‍‌]?\\|[െ-ൈൊ-ൌ]?[ൢ-ൣ]?[ാ-ൄൗ]?[ം-ഃ]?\\)\\|[അ-ഔൠ-ൡ]\\(‌?്[ക-ഹ]\\)?[െ-ൈൊ-ൌ]?[ൢ-ൣ]?[ാ-ൄൗ]?[ം-ഃ]?\\|‌്[ക-ഹ]\\|[ഀ-ൿ]" 0 font-shape-gstring])

---
Kenichi Handa
handa@m17n.org

Re: bug#5226: Malayalam conjunct formation breaks down at the end of long words

[Praveen A]

2009/12/17 Kenichi Handa <handa@m17n.org>:
> Thank you for them.  But the log shows that the Malayalam
> rendering is not using the latest Emacs, at least, it seems
> that the latest change for malayalam-composable-pattern is
> not working.
>
> Please show me the result of this:
>
> (aref composition-function-table #xD15)
>
> Mine is this:
>
> (["\\([ക-ഹ]‌?്‌?\\)*[ക-ഹ]\\(്[‍‌]?\\|[െ-ൈൊ-ൌ]?[ൢ-ൣ]?[ാ-ൄൗ]?[ം-ഃ]?\\)\\|[അ-ഔൠ-ൡ]\\(‌?്[ക-ഹ]\\)?[െ-ൈൊ-ൌ]?[ൢ-ൣ]?[ാ-ൄൗ]?[ം-ഃ]?\\|‌്[ക-ഹ]\\|[ഀ-ൿ]" 0 font-shape-gstring])

I copied (aref composition-function-table #xD15) to the Lisp
Interaction buffer and typed c-x c-e. Should I do it any differently?

This is what I get
(["[ഀ-ൿ‌‍]+" 0 font-shape-gstring])

- Praveen
-- 
പ്രവീണ്‍ അരിമ്പ്രത്തൊടിയില്‍
<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)
Join The DRM Elimination Crew Now!
http://fci.wikia.com/wiki/Anti-DRM-Campaign

Re: bug#5226: Malayalam conjunct formation breaks down at the end of long words

[Kenichi Handa]

In article <3f2beab60912170651l6e87b7ddxeffb4ff4b08d58a4@mail.gmail.com>, Praveen A <pravi.a@gmail.com> writes:

> I copied (aref composition-function-table #xD15) to the Lisp
> Interaction buffer and typed c-x c-e. Should I do it any differently?

No, that is the right way.

> This is what I get
> (["[ഀ-ൿ‌‍]+" 0 font-shape-gstring])

Then, you are not using the latest CVS code.  At least,
lisp/language/indian.el is not up-to-date or not
byte-compiled when you build your emacs.  Please check it.

---
Kenichi Handa
handa@m17n.org

Re: bug#5226: Malayalam conjunct formation breaks down at the end of long words

[Praveen A]

2009/12/18 Kenichi Handa <handa@m17n.org>:
> Then, you are not using the latest CVS code.  At least,
> lisp/language/indian.el is not up-to-date or not
> byte-compiled when you build your emacs.  Please check it.

I did a uninstall, distclean, cvs update and rebuilt emacs.
Also I removed the code needed for handling ZWJ from my .emacs file.

Now I'm getting this.
(["\\([ക-ഹ]‌?്‌?\\)*[ക-ഹ]\\(്[‍‌]?\\|[െ-ൈൊ-ൌ]?[ൢ-ൣ]?[ാ-ൄൗ]?[ം-ഃ]?\\)\\|[അ-ഔൠ-ൡ]\\(‌?്[ക-ഹ]\\)?[െ-ൈൊ-ൌ]?[ൢ-ൣ]?[ാ-ൄൗ]?[ം-ഃ]?\\|‌്[ക-ഹ]\\|[ഀ-ൿ]"
0 font-shape-gstring])

Now it is working beautifully. Thanks a lot.
Praveen
-- 
പ്രവീണ്‍ അരിമ്പ്രത്തൊടിയില്‍
<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)
Join The DRM Elimination Crew Now!
http://fci.wikia.com/wiki/Anti-DRM-Campaign