210575 - The "virama" character is not being rendered correctly. It should form a conjunct consonant when used between two consonants.

Status: RESOLVED DUPLICATE of bug 166520

(Core :: Layout: Text and Fonts, defect)

Description

[J.M.N.]

User-Agent:       Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.2b) Gecko/20021016
Build Identifier: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.2b) Gecko/20021016

The "virama" character is not being rendered correctly.  It should form a
conjunct consonant when used between two consonants.  Please view this same page
using Internet Explorer to see how the virama should be rendered.  

The Unicode codepoint for the "virama" character is 094D.  This is the text from
the Unicode code chart for the Devanagari range:

[start of quote]
094D DEVANAGARI SIGN VIRAMA 
 - halant (the preferred Hindi name)
 - suppresses inherent vowel
[end of quote]

I am unclear as to how the TrueType font and the browser interact to substitute
the virama sign for a conjunct character (usually composed of 2 half characters).  


Reproducible: Always

Steps to Reproduce:
1. Visit any page that uses Devanagari Unicode
2. Check for virama signs that appear beneath consonants
3. View the same page with Internet Explorer to compare the results.


Actual Results:  
I was able to determine that the Devanagari font was not being rendered
correctly using the Mozilla browser.



The two characters which are joined by the virama should be replaced by a single
conjunct character which is usually generated dynamically by the TrueType font
from 2 half-characters.

Comment 1

[Andrea]

Moving to the Browser component.

Comment 2

[Andrea]

Ehm, Browser|Complex Text Layout.

Please reporter: could you try to confirm it with a recent build ?

Roy Yokoyama: please reassing as required.

Comment 3

[J.M.N.]

I have checked this bug on earlier versions of Mozilla as well as with Mozilla
1.4 Release Candidate 3 and Netscape 7.02.  The behaviour is the same on each of
them.  The virama character is placed beneath the letters that are meant to form
conjunct consonants.  

Comment 4

[J.M.N.]

 have checked this bug on Mozilla 1.4 Release Candidate 3 and Netscape 7.02  as
well as with earlier versions of Mozilla.  
 
The behaviour is the precisely the same on each of them.  The virama character
is placed beneath the letters that are meant to form conjunct consonants.  This
is incorrect.  Please compare this with the way in which the same characters are
represented using Internet Explorer.  I am not sure as to the mechanism used to
substitute 2 characters joined by a virama by a single conjunct character.  All
I am sure of is that this is not happening using the Mozilla font rendering engine.

Comment 5

[Roy Yokoyama]

reassign to proper owner to get better exposure

Comment 6

[Jungshik Shin]

jmn, is this only on Windows 98 or do you have the same problem on Win 2k/XP as
well? I don't think you do.
BTW, on Win 2k/XP, you have to activate/install the support package for one of
complex scripts (Devanagari, Tamil, Thai) before testing. If it's only on
Windows 98(Win 9x/ME), this is a dupe of 166520. If it's about Win 2k/XP, that's
very likely because you didn't activate/install CTL support package (available
on Win2k/XP installation CD.). 
 
It's filed for Win98, I'm marking this as a dupe of bug 166520.

See also bug 206550. I18N release notes for 1.4 is going to have a section
devoted to  this issue because there have been so much confusion and
misunderstanding on CTL support in Mozilla. 

*** This bug has been marked as a duplicate of 166520 ***

Comment 7

[J.M.N.]

Hello Jungshik Shin,

I have just read the notes on bug 166520 and now understand that the problem is
due to the fact that MSIE bypasses the poor Unicode support on pre-Win2K
versions by using the Uniscribe DLL.  I hope that Mozilla will use a platform
independent solution to resolve this issue.

I agree that this bug is a duplicate of 166520, however, apart from the virama
issue I haven't noticed any other problems with the display of Devanagari for
Hindi or Sanskrit with the current versions of Netscape 7.02 and the 1.4 release
of Mozilla.

Comment 8

[Jungshik Shin]

Hello J.M.N, 
I guess misrendering of consonant conjuncts is most conspicuous, but there 
must be other defetcs. You can try, for instance, my test case attached to bug 
204286. 
 

Comment 9

[J.M.N.]

Hello Jungshik Shin

I will certainly do as you request and attempt to work through your
test-case as soon as I have some free-time.  The number of
conjunct-consonants in Devanagari is almost open-ended so it might take a
while to test this fully.

Thank you for explaining the technical details of Complex Text Layout to me.
It has opened up for me an entirely new avenue of research.

In the meantime please check the latest comment I have added to the Complex Text
Layout issue.  I could have also mentioned the latest Flash Player and the
various SMIL players (RealOne, QuickTime, etc.) which also claim to support
Unicode encodings in this context but I think that the general point was clear
without having to produce a comprehensive list.

It does seem to be an issue that most organisations outside of Unicode.org,
ISO, W3C, Mozilla and the Linux communities seem to be avoiding.

Good Luck !

J.M. Nicholls
London

Comment 10

[J.M.N.]

Hello again Jungshik Shin,

I have just taken a quick look at your test-case page using Mozilla 1.4 on
Windows 98:

http://bugzilla.mozilla.org/attachment.cgi?id=122383&action=view

I have opened 2 windows onscreen side-by-side :

1)    With Mozilla
2)    With Internet Explorer

I can confirm that most of the Mozilla renderings are still incorrect.  None
of the conjunct consonants are working and the placement of the other
diacritical characters is also dysfunctional.

These combinations seem to work OK :

RRAn + VIRAMAn
TAd + RAd
KAn + AAvs + CANDRABINDUn

I hope this helps.

Good Luck !

J.M. Nicholls
London