Closed Bug 362527 Opened 18 years ago Closed 17 years ago

Malayam font support: fails on just Karthika

Categories

(Core :: Graphics, defect)

x86
Windows XP
defect
Not set
normal

Tracking

()

RESOLVED FIXED

People

(Reporter: BijuMailList, Unassigned)

References

Details

Attachments

(6 files)

this is a feature or bug!!! If firefox cant display some Malayalam word in WinXP default "Karthika" font. It switches to Anjali Old Lipi, when it is available. Test 1: * Ensure "Kartika" is the only Malayalam font on you computer. * goto http://snipurl.com/1370t#.E0.B4.A4.E0.B4.BF.E0.B4.B0.E0.B5.81.E0.B4.B5.E0.B4.A8.E0.B4.A8.E0.B5.8D.E0.B4.A4.E0.B4.AA.E0.B5.81.E0.B4.B0.E0.B4.82 http://varamozhi.wikia.com/wiki/Geographical_Locations_of_Malayalam_Bloggers#.E0.B4.A4.E0.B4.BF.E0.B4.B0.E0.B5.81.E0.B4.B5.E0.B4.A8.E0.B4.A8.E0.B5.8D.E0.B4.A4.E0.B4.AA.E0.B5.81.E0.B4.B0.E0.B4.82 * You shoud see ===Trivandrum=== * Chandrashekaran nair / Peyad * Ponnapan - the Alien * Chandu / Poojapura * "Chandrashekaran nair" is show as boxes see Karthika_without_Anjali.png Test 2: * install AnjaliOldLipi from http://prdownloads.sourceforge.net/varamozhi/AnjaliOldLipi-0.730.ttf?use_mirror=kent * Ensure "Kartika" is still the malayalam font at Tool > Options > fonts * goto http://snipurl.com/1370t#.E0.B4.A4.E0.B4.BF.E0.B4.B0.E0.B5.81.E0.B4.B5.E0.B4.A8.E0.B4.A8.E0.B5.8D.E0.B4.A4.E0.B4.AA.E0.B5.81.E0.B4.B0.E0.B4.82 http://varamozhi.wikia.com/wiki/Geographical_Locations_of_Malayalam_Bloggers#.E0.B4.A4.E0.B4.BF.E0.B4.B0.E0.B5.81.E0.B4.B5.E0.B4.A8.E0.B4.A8.E0.B5.8D.E0.B4.A4.E0.B4.AA.E0.B5.81.E0.B4.B0.E0.B4.82 * You shoud see ===Trivandrum=== * Chandrashekaran nair / Peyad * Ponnapan - the Alien * Chandu / Poojapura * "Chandrashekaran nair" is show in AnjaliOldLipi when others are in Kartika see Karthika_with_Anjali.png This wont be a big deal for people who have AnjaliOldLipi, but issue for those who dont have it.
Attached image Anjali.png
font set to Anjali works fine
Just FYI: Problem exist before Bug# 361778 fix I tested in Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20061129 Minefield/3.0a1 and after fix in Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20061201 Minefield/3.0a1
The problem here is that the entry that shows up as boxes is using codepoints for cillaksarams which do not exist in the Karthika font; in fact they are not even defined in the current version of Unicode, though they have been proposed for inclusion (I'm not sure what the status of the proposal is). The codepoints in question are: 0D7A MALAYALAM LETTER NN ൺ 0D7B MALAYALAM LETTER N ൻ 0D7C MALAYALAM LETTER RR ർ 0D7D MALAYALAM LETTER L ൽ 0D7E MALAYALAM LETTER LL ൾ 0D7F MALAYALAM LETTER K ൿ The current behaviour on Windows seems to be that if we can't find a font that provides glyphs for *all* the codepoints in a given Uniscribe item, we don't display *any* glyphs for that item. Is this really what we want to be doing?
Component: Layout: Fonts and Text → GFX: Thebes
QA Contact: layout.fonts-and-text → thebes
No, there is another bug filed (???) on going further than the uniscribeitem level -- I know how to do it in a way that will be fast but just haven't had the time. Once I fix that it'll fix this and a handfull of other bugs.
oops.. I add it this to bug 361783, sorry (In reply to comment #5) > The current behaviour on Windows seems to be that if we can't find a font that > provides glyphs for *all* the codepoints in a given Uniscribe item, we don't > display *any* glyphs for that item. Is this really what we want to be doing? In case of malayalam even if you display *other* glyphs supported in the font, still issue remains for those who view with Karthika font, as it will be difficult to read malayalam with out chill-aksaram, when the page is written with \u0D7A to \u0D7B. I dont know status of \u0D7A to \u0D7B. Some time unicode supported it. Then they found another alternative or something. So as an exception in case of malayalam chill-aksarams, if a font wont support chill-aksaram please convert it to alternate form using following formula for display. ൺ (u+0D7A) -> ണ്‍ (u+0D23 + u+0D4D + u+200D) ൻ (u+0D7B) -> ന്‍ (u+0D25 + u+0D4D + u+200D) ർ (u+0D7C) -> ര്‍ (u+0D30 + u+0D4D + u+200D) ൽ (u+0D7D) -> ല്‍ (u+0D32 + u+0D4D + u+200D) ൾ (u+0D7E) -> ള്‍ (u+0D33 + u+0D4D + u+200D) ൿ (u+0D7F) -> ക‍് (u+0D23 + u+0D4D + u+200D) __ (u+0D7?) -> യ്‍ (u+0D2F + u+0D4D + u+200D) see Chill_conversion_formula.html (attachment 247355 [details]) with Karthika font, to find how it works
(In reply to comment #5) > The problem here is that the entry that shows up as boxes is using codepoints > for cillaksarams which do not exist in the Karthika font; in fact they are not > even defined in the current version of Unicode, though they have been proposed > for inclusion (I'm not sure what the status of the proposal is). They are accepted for inclusion in some future version of Unicode, with a recent name change to MALAYALAM LETTER CHILLU xx. Contrary to comment #7, they have not ever been actually included in Unicode, because the Unicode stability policies make it impossible to remove a character once included. > The current behaviour on Windows seems to be that if we can't find a font that > provides glyphs for *all* the codepoints in a given Uniscribe item, we don't > display *any* glyphs for that item. Is this really what we want to be doing? That's bug 332649.
Depends on: 332649
This is not a mozilla bug. It is problem with the content of that page. There is no such code point in unicode and if you wanted to fix this you have to make that content use only unicode code points (use your formula on the content, it is a wiki, so you can do it yourself). This a serious issue to be decided by unicode and I suggest we close this bug. As of now these code points are only in the proposed 5.1 version of unicode. The actual problem here is with Anjali which included this code points in the font even though it is not in unicode. Actually it got included in unicode for just a week. The argument against those were so strong but developers of Anjali never bothered to remove those (presumably they supported these code points and they hoped it would again get included some time). But the real stability issue is now when they get included again. Because the combination which Biju mentioned in his formula has existed for more than 10 years and virtually every content on the internet follow that sequence. Now if you include these code points it would result in dual encoding for these characters making it simple for spoofing (IDN allows you to register domains in unicode) also it would make all computing operations complex to say the least. Sorting, searching, comparison ... will have to consider existence of two encodings for the same sequence.
(In reply to comment #9) > This is not a mozilla bug. Issue here is the characters before MALAYALAM LETTER CHILLU N ie ചന്ദ്രശേഖര not displaying attachment 247209 [details] The conversion formula on attachment 247356 [details] is only a work around I suggested. BTW: I have retest to whether the problem of ചന്ദ്രശേഖര not displaying is fixed or not
Did you check with latest firefox nightly build? Justin mentioned this issue is fixed. There no need for conversion as that particular code point is not unicode compliant and is a bug in the content+Anjali
Attached image karthika_only_fixed.png
tested and found it is PASS with out anjali - display other characters properly in Karthika, and chill "N" as box with anjali - display other characters properly in Karthika, and chill "N" in anjali, or an alternative font
Status: NEW → RESOLVED
Closed: 17 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: