Home / Browse / Adding Unicode Sinhala to GNU/Linux / Mail / Archive

Adding Unicode Sinhala to GNU/Linux

Email Archive: sinhala-technical (read-only)

[sinhala-technical] Bug in BhashitaComplex font: Wrong ligature seen for dantaja sayanna touching itself

From: கா. சேது | කා. සේතු | K. Sethu <skhome@gm...> - 2011-08-30 11:13

While testing the sequences for touching consonants that I have
included in a Wijesekera keymap being made by me in KMFL
(http://kmfl.sourceforge.net/), I detected a bug in BhashithaComplex
and BhashithaComplexSans fonts for Linux released by ICTA.

The code points sequence for any two consonants touching is:  <CONS1>
200D 0DCA <CONS2>  wherein CONS1 and CONS2 are the two consonants
which are to touch with al-lakuna hidden. The bug I observe is when
both are ස (dantaja sayanna).

You can see it in the attached screen shot file:
touch-cons-LibreWriter.png taken in Libre Writer. With the correct
sequence  when either one of the Bhashita font is used the second
character changes to mayanna. Changing font to LKLUG shows it
correctly as  ස but of course LKLUG lacks the level 3 ligature to hide
al-lakuna and also show touching correctly.

The other combinations I tried - (ම,ම) (ක,ක), (න,න) and (ස.ම) all
yield correct ligatures.

I also looked at the ligature for ස touching ම under fontforge -
attached file touch_sa_ma_liga-bhashita-fontforge.png . The sub-tables
and the components fields for the Ligature do not have any additional
data to cover the case of both consonant being   ස. So why these
Bhashitha fonts yield this bug is beyond my level of know-how.

~Sethu

Thread View

Thread	Author	Date
[sinhala-technical] Bug in BhashitaComplex font: Wrong ligature seen for dantaja sayanna touching itself	கா. சேது \| කා. සේතු \| K. Sethu <skhome@gm...>

Re: [sinhala-technical] Bug in BhashitaComplex font: Wrong ligature seen for dantaja sayanna touching itself

From: Harshula <harshula@gm...> - 2011-08-30 13:27

Hi Sethu,

On Tue, 2011-08-30 at 16:43 +0530, கா. சேது | කා. සේතු | K. Sethu wrote:
> While testing the sequences for touching consonants that I have
> included in a Wijesekera keymap being made by me in KMFL
> (http://kmfl.sourceforge.net/), I detected a bug in BhashithaComplex
> and BhashithaComplexSans fonts for Linux released by ICTA.

Thanks. I've forwarded your email to Dineesha.

Why kmfl? Does it provide surrounding text support? Does it even provide
pre-edit buffer support?

cya,
#

Re: [sinhala-technical] Bug in BhashitaComplex font: Wrong ligature seen for dantaja sayanna touching itself

From: கா. சேது | කා. සේතු | K. Sethu <skhome@gm...> - 2011-09-01 05:16

2011/8/30 Harshula <harshula@...>:
> Hi Sethu,
>
> On Tue, 2011-08-30 at 16:43 +0530, கா. சேது | කා. සේතු | K. Sethu wrote:
>> While testing the sequences for touching consonants that I have
>> included in a Wijesekera keymap being made by me in KMFL
>> (http://kmfl.sourceforge.net/), I detected a bug in BhashithaComplex
>> and BhashithaComplexSans fonts for Linux released by ICTA.
>
> Thanks. I've forwarded your email to Dineesha.

Thanks for the prompt attention.

>
> Why kmfl? Does it provide surrounding text support? Does it even provide
> pre-edit buffer support?
>

For "Why kmfl" part I will give my opinions via another thread [
http://sourceforge.net/mailarchive/message.php?msg_id=20059140 ] I
started 3 years (!) back in this mailing list which was my seeking
definitions for Surrounding Text / Preedit Support.

On the rest two, very briefly here:

No it does not have pre-edit buffer support.

But for STS it can be considered as supporting one half of it. -
Referring to your definition in the above mentioned thread STS's
ability is to change what's on *left and right* of cursor without
having to store keys in buffer. KMFL's ability is only for whats on
left of cursor - already committed character sequence only left of
cursor position can be modified with a key event.

More later in the other thread.

~Sethu