Closed Bug 248690 Opened 20 years ago Closed 20 years ago

enable Punjabi/Panjabi (pa) and Bengali/Bangla (bn) in language.properties

Categories

(Core :: Internationalization, defect)

defect
Not set
normal

Tracking

()

RESOLVED FIXED

People

(Reporter: sukhuk, Assigned: smontagu)

References

Details

(Keywords: intl)

Attachments

(3 files)

User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7) Gecko/20040614 Firefox/0.8 Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7) Gecko/20040614 Firefox/0.8 Under the languages section of the general preferences, the languages Punjabi [pa] seems to be entirely missing! I'm pretty certain it's in there in other Mozilla based products! Also, in fonts and colours it would be preferable to have an option for Gurmukhi (script used to write Punjabi). Reproducible: Always Steps to Reproduce: 1. Click Tools > Options 2. Go to the general section and click "Languages..." 3. Go to the drop down list labelled "Select a language to add" Actual Results: No Punjabi language is listed. Expected Results: Listed an option for Punjabi [pa].
steffen, care to look at this and comment/update?
Punjabi (also called Panjabi), code pa, is currently disabled here: http://lxr.mozilla.org/mozilla/source/intl/locale/src/language.properties#196 This file is common to Mozilla and Firefox. -> Browser, I18N Can't find a bug there -> confirming.
Assignee: firefox → smontagu
Status: UNCONFIRMED → NEW
Component: Preferences → Internationalization
Ever confirmed: true
OS: Windows XP → All
Product: Firefox → Browser
QA Contact: mconnor → amyy
Hardware: PC → All
Summary: Missing Punjabi in list of languages → enable Punjabi/Panjabi (pa) in language.properties
Version: unspecified → Trunk
It doesn't show up in Mozilla's add language dialog as well. The difference is that you can enter "pa" manually in Mozilla. You can't in Firefox because I left out that functionality when porting the dialog (bug 181541).
Would it be possible to add it permanently to the main list? It is a language spoken by over 60 million people!
I don't know how well we support Gurmukhi on different platforms. There are also possible problems in adding a Gurmukhi langGroup for Punjabi (which we would need to do to have it appear in fonts preferences), since I understand that Punjabi can also be written in Devanagari or a modified Arabic script.
(In reply to comment #5) > I don't know how well we support Gurmukhi on different platforms. Well, I can say it's fully supported on Windows using Uniscribe. I'm not sure about Linux support with CTL but I'm pretty certain Pango supports it also. In addition, Gurmukhi is one of the most simple Indic scripts to implement. > There are also > possible problems in adding a Gurmukhi langGroup for Punjabi (which we would > need to do to have it appear in fonts preferences), since I understand that > Punjabi can also be written in Devanagari or a modified Arabic script. Punjabi can be written in Devanagari but it is comparitively rare (read: hardly ever). Most Punjabi pages using [pa] online are written in Gurmukhi. I'm not sure how much Shahmukhi (the Arabic script) is used online. The solution could be pa-IN = Gurmukhi, pa-PK = Shahmukhi. However, as it's implemented on Windows, pa seems to generally (actually, almost exclusively) refer to Gurmukhi.
(In reply to comment #6) > The solution could be pa-IN = Gurmukhi, pa-PK = Shahmukhi. However, as it's > implemented on Windows, pa seems to generally (actually, almost exclusively) > refer to Gurmukhi. That was my impression too from some unsystematic surfing. (I wish I knew a search engine where you could search for pages with particular attributes!) In fact, the only Shahmukhi pages I came across use image files instead of text.
The fact is that Shahmukhi seems to have been largely ignored. Even though Punjabi is meant to be the most widely spoken language of Pakistan, its use is generally not encouraged. It isn't even an official language of Pakistan. As the Indian side doesn't use Shahmukhi, there has been very little effort to support it. I can't really see the situation changing anytime soon. So, will it be possible to add [pa] to the list?
Status: NEW → ASSIGNED
We can just enable pa in language.properties file, but that doesn't mean much without other changes. Nonetheless, I'm not against enabling it in the file. JFYI : Pango doesn't seem to support Gurmukhi, yet in its Xft backend although it used to in X11 backend. (even if it does, Mozilla doesn't use Pango although I'm gonna upload my 1.7build with a patch to make Mozilla make use of it : bug 215219). On Windows, we don't use Uniscribe directly (see bug 218887) but still complex script rendering works to a limited extent on Win 2k/XP. On Mac OS X, we need to use ATSUI which we don't use at the moment (in addition to this, text rendering/font handling on OS X need a lot of work). Due to this platform parity issue, I have been hesitant to add a whole bunch of new 'langGroups' (script groups) to the font preference. Perhaps, we just have to go ahead....
Keywords: intl
Is there any sort of time frame that we can look at to get all these (very significant) issues with complex scripts sorted? For example, is there any chance they might make Firefox 1.0 build?
Also, for testing (and possible distribution with Mozilla), a GPL Gurmukhi font with Unicode 4.0 support is available at http://guca.sourceforge.net/typography/fonts/saab/.
Thanks for the reference to the font. On Windows 2k/XP, it should more or less works (unless you use 'text-justify') as long as uniscribe supports it (Mozilla uses Uniscribe indirectly, which is why 'text-justify' doesn't work). With bug 218887 fixed, it'll work with justified text as well. On Linux, Mozilla alone can't do anything unless Pango supports it. (does Owen Taylor, the principal Pango developer, know about the font? If he knows that, it will help him support Gurmukhi sooner) Anyway, I'm planning to contribute my patched version of Mozilla 1.7 and firefox 1.0 (gtk2+xft+pango build) to ftp.mozilla.org.
That's great news. The font is the first of its kind and relatively new so I doubt Owen Taylor knows about it. If you could pass on details to him that will be very helpful. Gnome version 2.6 supports Unicode Gurmukhi correctly (there are some inconsistences present that are incompatible with Windows) and as far as I know that uses Pango also.
Any update on this?
Is there any chance of getting this enabled so it can be fixed for Firefox 1.1? A simple switch to enable pa and pa-IN should be all that's required.
Attached patch Preliminary patch (deleted) — Splinter Review
There are some things I'm not sure about in here, also testing is blocked by bug 283461.
I'm not exactly sure what I'm reading here in this patch, but I presume that pa and pa-IN will default to Gurmukhi and pa-PK to Shahmukhi (Arabic + some extra bits)? Also, I'm not sure if this is the right place to comment about another issue, but I'll do so anyway: At the moment, Gecko seems to default to using a Devanagari font when showing Dandas (0964, 0965) in Gurmukhi text - even if the Gurmukhi font supports the Danda character. Will this patch enable a specific font to be used for the Dandas in Gurmukhi text? By that I mean will it treat the Dandas in the Devanagari block as if they were Gurmukhi text?
Sorry for replying to myself: > I'm not exactly sure what I'm reading here in this patch, but I presume that pa > and pa-IN will default to Gurmukhi and pa-PK to Shahmukhi (Arabic + some extra > bits)? I noticed the line: + { "x-gurmukhi", 0x0A05, (const FcChar8 *)"pa-IN" }, Should that be: + { "x-gurmukhi", 0x0A05, (const FcChar8 *)"pa" }, Also you've got a lot of 0x0A05 dotted around. I'm sure there is a reason that they start at 0x0A05 and not 0x0A01 (because Tamil and Devanagari start at 5 too) but I just thought I'd bring it to your attention.
(In reply to comment #17) > I'm not exactly sure what I'm reading here in this patch, but I presume that pa > and pa-IN will default to Gurmukhi and pa-PK to Shahmukhi (Arabic + some extra > bits)? Yes, that's correct. (Ideally we should use be able to select fonts for Shahmuki and Urdu which are differennt from the fonts for Arabic, but that's another issue). > Also, I'm not sure if this is the right place to comment about another issue, > but I'll do so anyway: At the moment, Gecko seems to default to using a > Devanagari font when showing Dandas (0964, 0965) in Gurmukhi text - even if the > Gurmukhi font supports the Danda character. Will this patch enable a specific > font to be used for the Dandas in Gurmukhi text? By that I mean will it treat > the Dandas in the Devanagari block as if they were Gurmukhi text? Can you point me to a Gurmukhi page with Dandas so I can test this? If the page is marked up with lang="pa", it should use the Gurmukhi font. (In reply to comment #18) > I noticed the line: > > + { "x-gurmukhi", 0x0A05, (const FcChar8 *)"pa-IN" }, > > Should that be: > > + { "x-gurmukhi", 0x0A05, (const FcChar8 *)"pa" }, > > Also you've got a lot of 0x0A05 dotted around. I'm sure there is a reason that > they start at 0x0A05 and not 0x0A01 (because Tamil and Devanagari start at 5 > too) but I just thought I'd bring it to your attention. > This is one of the parts of the patch I was unsure about. I'm hoping Jungshik will help me out.
Attached image Internet Explorer Dandas (deleted) —
Attached image Firefox Dandas (deleted) —
(In reply to comment #19) > Can you point me to a Gurmukhi page with Dandas so I can test this? If the page > is marked up with lang="pa", it should use the Gurmukhi font. http://guca.sourceforge.net/pa/faq/ Please see attached the screenshots - it shows how it appears that another font is being used for Dandas. Internet Explorer doesn't do this.
(In reply to comment #19) > (In reply to comment #18) > > Also you've got a lot of 0x0A05 dotted around. I'm sure there is a reason > This is one of the parts of the patch I was unsure about. I'm hoping Jungshik > will help me out. The field occupied by U+0A05 used to play the role of 'A' in Latin letter ('the' representative character of a script whose glyph must be present in a font that claims to support the script) in nsFontMetricsXft.cpp. However, the field is not used any more so that we have to get rid of it. As for X11 corefonts, I think we don't have an option other than just mapping most of new (non-Latin/Cyrillic/CJK/Greek) langGroups (see bug 237434) to iso10646-1 Anyway, if we decide to open a flood gate for 1.8b2, let's add many others as well (see bug 237434). Btw, what are default fonts to specify for Punjabi/Panjabi?
Default fonts are: Raavi (Windows) Saab (GNU GPL'ed font - most popular free font) AnmolUni (New GNU GPL'ed font) Lohit Punjabi also Code2000 and Arial Unicode MS
Summary: enable Punjabi/Panjabi (pa) in language.properties → enable Punjabi/Panjabi (pa) and Bengali/Bangla (bn) in language.properties
Depends on: 288638
Simon, what do you think of combining this bug and bug 237434? If you don't mind, I'll make a combined patch.
Btw, if we're gonna add Gurmukhi and Bengali, we should add other South/South Asian scripts (Malayalam, Telugu, Khmer, Kanada, and so forth) supported by Uniscribe (on Windows XP) and/or Pango support. (On Mac OS X, we have to fix Gfx to work better with AAT and ATSUI to take advantage of OS support of complex script)
(In reply to comment #26) > Simon, what do you think of combining this bug and bug 237434? If you don't > mind, I'll make a combined patch. Sounds like an excellent idea :)
it's fixed by my check in for bug 237434
Status: ASSIGNED → RESOLVED
Closed: 20 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: