255566 - All Urdu Characters are not joining as suppose to join.

Status: RESOLVED FIXED

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

Description

[Nazir]

User-Agent:       Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.7) Gecko/20040707 Firefox/0.9.2
Build Identifier: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.7) Gecko/20040707 Firefox/0.9.2

Urdu like Arabic change shape in writing with reference to the Characters
comming befer or after. If you see the page in IE Characters are properly
joining whereas in Mozilla (firebox or any) few Characters don't join properly.
Please note every thing in the page is in unicode (charset=utf-8). This bug
force Urdu Viewer / Web Designer to avoid Mozilla.


Reproducible: Always
Steps to Reproduce:
1.
2.
3.

Comment 1

[Olivier Cahagne]

Can you post a screenshot comparing Gecko/IE6 rendering, at least highliting
what's wrongly rendered ?
Can you produce a minized testcase ?

Comment 2

[Nazir]

Attachment: Gif sketch for the Problem Top one is from Firebox and bottom one is from IE. See the third word from right the bottom one is correct.

In The firebox Sample the last alphabet of the Third word is not joining the
previous alphabet, whereas in Ie it is joined properly.

Comment 3

[Nazir]

Attachment: a Urdu page with three fonts

Comment 4

[Nazir]

http://www.unicode.org/charts/collation/chart_Arabic.html
Please click the link and search for 6cc,
1)	Every time you enter alphabet x6cc it changes its shape @  earlier /
following alphabet. It changes to fbfe,fbff,fbfd or fbfc.
2)	What happening in Mozilla everytime when the alphabet x6cc comes before the
alphabet x6ba it is not changing. It show the shape of 6cc whereas it is suppose
to change to fbfe. It changes to fbfe in IE.  In all the other occurrence
alphabet x6cc changes to respective shape in Mozilla but not with  x6ba.

Comment 5

[Olivier Cahagne]

Marking NEW since it has a testcase and good analysis (eventhough I cannot judge
being on urdu specialist).

Comment 6

[Simon Montagu :smontagu]

Taking bug.

Comment 7

[Simon Montagu :smontagu]

Attachment: Patch

This patch adds the missing mapping from U+06BA ARABIC LETTER NOON GHUNNA to
the matching presentation forms U+FB9E and U+FB9F, and while I was in the files
I checked gJoiningClass against the current version of
http://www.unicode.org/Public/UNIDATA/ArabicShaping.txt and made the following
corrections:

U+0649 ALEF MAKSURA	    from eRJ (right-joining) to eDJ (double-joining)
U+0656 SUBSCRIPT ALEF	    from eNJ (non-joining)   to eTr (transparent)
U+0657 INVERTED DAMMA	    from eNJ (non-joining)   to eTr (transparent)
U+0658 MARK NOON GHUNNA     from eNJ (non-joining)   to eTr (transparent)
U+06BA NOON GHUNNA	    from eNJ (non-joining)   to eDJ (double-joining)
U+06DD END OF AYA	    from eTR (transparent)   to eNJ (non-joining)
U+06DE START OF RUB EL HIZB from eTR (transparent)   to eNJ (non-joining)

Comment 8

[Jungshik Shin]

Comment on attachment 156499 [details] [diff] [review]
Patch

r=jshin

Comment 9

[Nazir]

(In reply to comment #7)
> Created an attachment (id=156499)
> Patch
> This patch adds the missing mapping from U+06BA ARABIC LETTER NOON GHUNNA to
> the matching presentation forms U+FB9E and U+FB9F, and while I was in the 
files
> I checked gJoiningClass against the current version of
> http://www.unicode.org/Public/UNIDATA/ArabicShaping.txt and made the following
> corrections:
> U+0649 ALEF MAKSURA	    from eRJ (right-joining) to eDJ (double-joining)
> U+0656 SUBSCRIPT ALEF	    from eNJ (non-joining)   to eTr (transparent)
> U+0657 INVERTED DAMMA	    from eNJ (non-joining)   to eTr (transparent)
> U+0658 MARK NOON GHUNNA     from eNJ (non-joining)   to eTr (transparent)
> U+06BA NOON GHUNNA	    from eNJ (non-joining)   to eDJ (double-joining)
> U+06DD END OF AYA	    from eTR (transparent)   to eNJ (non-joining)
> U+06DE START OF RUB EL HIZB from eTR (transparent)   to eNJ (non-joining)

Thanks for Patch, How can I check in my computer, or I have to download new,
Nazir

Comment 10

[Simon Montagu :smontagu]

Patch checked in. If you download a nightly build tomorrow, you should see the
fix. The rendering will probably still not be perfect in all fonts until bug
218887 (or equivalents in other platforms) is fixed.

Comment 11

[Simon Montagu :smontagu]

Comment on attachment 156499 [details] [diff] [review]
Patch

Requesting approval for branch and aviary. This is a minimal risk patch since
it only changes data tables, and improves rendering of Urdu, a language with
over 100 million speakers world-wide.

Comment 12

[Mike Kaply [:mkaply]]

Comment on attachment 156499 [details] [diff] [review]
Patch

a=mkaply for aviary and 1.7.3

Comment 13

[Nazir]

(In reply to comment #11)
> (From update of attachment 156499 [details] [diff] [review])
> Requesting approval for branch and aviary. This is a minimal risk patch since
> it only changes data tables, and improves rendering of Urdu, a language with
> over 100 million speakers world-wide.
> 

Hi Simon,
How come it is fixed in Firefox bot not in Mozilla 1.7.3

Comment 14

[Simon Montagu :smontagu]

1.7.3 wasn't released from the main 1.7 branch. Apart from some specific
security fixes it's the same as 1.7.2