[ Home | Freetime works | FreeBSD | KFUG ]

Mozilla: Unix Non-Latin1 Printing Improvement Plan

Editor: CHOI Junho < cjh@kr.FreeBSD.org >

Developers:

Feature Owner: CHOI Junho: <cjh@kr.FreeBSD.org>
XFE: CHOI Junho: <cjh@kr.FreeBSD.org>
XP: N/A
Document: CHOI Junho: <cjh@kr.FreeBSD.org>
People May Willing to Help: Liu Hai: <liuhai@irdu.nus.edu.sg>

This page is for improving HTML to PostScript conversion library of Mozilla(an NPL'ed web browser released from Netscape Communications)(lib/xlate in the source). Its purpose is mainly improving Non-latin1 language printing support and other generic PostScript converion issues.

After some discussion in mozilla-i18n mailing list(or netscape.public.mozilla.i18n), based on the Liu Hai's unofficial work done for Netscape, this plan is started.

In Unix, PostScript is widely accepted printing solution. Although it supports Non-latin1 fonts in several ways(Type 1, 0, 3, CID, OCF), we always suffer from the fundamental problem - "This font is not my printer!". Widely used latin-1 font(Times, Courier, Helvetica, etc) is on every PostScript printer, but other encoding fonts aren't(e.g. I have no chance to see the printers have CJK PS font yet). If the printer have some additional memory or disks, you can download the fonts to the printer and use it, but the printer can do this is a few. Ghostscript is a good solution for this(it is user-level program so you can add/remove fonts, and you can print out PS document to non-PS printers), but Mozilla cannot assume the existence of ghostscript(Mozilla should work without relying on other applications, I think).

Specification:

Liu's work originally includes:

Multiple faces(bold, italic, bold-italic) support for Non-latin1 characters
Page decoration(page number/url/title output)
External metric file support

The works should be done until Mozilla source code freeze(please refer the schedule):

Non-latin1 title print(currently just one face(latin1)).
Give more hints to external filters(charcode, URL, etc...)
More DSC compliant(papersize info).
URL output(in the PS comments).

And other improvement planned after source freezing:

CID support for CJK.
External metrics
Embedded fonts
Font downloading
Translating into other format(like pdf) - almost a dream
Unicode(UTF-8 or others) printing
Trying to integrate ogonkify...(?)

For current status, please see "Download" and "Open Issues" section.

I need some examples of Latin-2 and Chinese pages. If you know good url of these languages, please mail to me.

Detail Design

(coming soon).....

Download:

Snapshots(only testing purpose)

This patch is for source of Mozilla(Unix/XFE only). You must have original source.

moz-xlate-19980627.diff.gz
- Just porting of Liu's original work into Mozilla source tree.
moz-xlate-19980728-0824.diff.gz
- Based on mozilla-19980728 public source(it can be applied with 19980904 safely)
- ISO8859-2, EUC-KR, EUC-JP, SJIS verified.
- Non-latin1 title printing in xl_annotate_page.
- Give charset info and URL in postscript output.
- Give attribute info in postscript output.
- Give papersize info in DSC comment.
- Be more liberal in parsing AFM files.
- Non-latin1 underline and strikeout decoration enabled.
- Japanese printing charset is fixed to SJIS(in resource file).
- Resource file arranged
This patch is checked in. Just take newest mozilla source from mozilla.org, not from here!

To test it, uncompress the patch, and apply it to the source:

# cd MOZILLA-SOURCE/mozilla
# gunzip /some/dir/moz-xlate-xxxx.diff.gz
# patch < /some/dir/moz-xlate-xxxx.diff

and build Mozilla as usual. You may have to install X resource file(Netscape-export.ad) to appropriate X application resource directory(or one of the paths specified by XAPPLRESDIR variable).

# cp cmd/xfe/.OBJ/Netscape-export.ad \
     /usr/X11R6/lib/X11/app-defaults/Netscape
     (otherwise the directory you can use)

And don't forget to edit the resource file. You can alter the font name, AFM file directory, and other thing. You should uncomment the setup for each encoding to use it. See examples in Netscape.ad.

!
! ISO-8859-2 sample:
!   You can get latin-2 Type 1 fonts from:
!     ftp://ftp.osb.hu/pub/misc/fonts/local/latin-2/
!   Get l2afm005.zip and l2pfb005.zip.
!
*documentFonts.iso-8859-2*pscode:			iso-8859-2
*documentFonts.iso-8859-2.variable.normal*psname:	TimesNewRomanL2
*documentFonts.iso-8859-2.variable.normal*psafmfile:	/usr/local/lib/mozilla/fonts/Type1/tnrl2___.afm
*documentFonts.iso-8859-2.variable.bold*psname:		TimesNewRomanL2-Bold
*documentFonts.iso-8859-2.variable.bold*psafmfile:	/usr/local/lib/mozilla/fonts/Type1/tnrl2b__.afm
*documentFonts.iso-8859-2.variable.italic*psname:	TimesNewRomanL2-Italic
*documentFonts.iso-8859-2.variable.italic*psafmfile:	/usr/local/lib/mozilla/fonts/Type1/tnrl2i__.afm
*documentFonts.iso-8859-2.variable.boldItalic*psname:	TimesNewRomanL2-BoldItalic
*documentFonts.iso-8859-2.variable.boldItalic*psafmfile:/usr/local/lib/mozilla/fonts/Type1/tnrl2bi_.afm
*documentFonts.iso-8859-2.fixed.normal*psname:		CourierNewL2
*documentFonts.iso-8859-2.fixed.normal*psafmfile:	/usr/local/lib/mozilla/fonts/Type1/cnl2____.afm
*documentFonts.iso-8859-2.fixed.bold*psname:		CourierNewL2-Bold
*documentFonts.iso-8859-2.fixed.bold*psafmfile:		/usr/local/lib/mozilla/fonts/Type1/cnl2b___.afm
*documentFonts.iso-8859-2.fixed.italic*psname:		CourierNewL2-Italic
*documentFonts.iso-8859-2.fixed.italic*psafmfile:	/usr/local/lib/mozilla/fonts/Type1/cnl2i___.afm
*documentFonts.iso-8859-2.fixed.boldItalic*psname:	CourierNewL2-BoldItalic
*documentFonts.iso-8859-2.fixed.boldItalic*psafmfile:	/usr/local/lib/mozilla/fonts/Type1/cnl2bi__.afm

!
! EUC-KR sample:
!
*documentFonts.EUC-KR*pscode:				EUC-KR
*documentFonts.EUC-KR.variable.normal*psname:		Munhwa-Regular-KSC-EUC-H
*documentFonts.EUC-KR.variable.normal*psmetric:		1000 1000 990 -10
*documentFonts.EUC-KR.variable.bold*psname:		Munhwa-Regular-KSC-EUC-H
*documentFonts.EUC-KR.variable.bold*psmetric:		1000 1000 990 -10
*documentFonts.EUC-KR.variable.italic*psname:		Munhwa-Regular-KSC-EUC-H
*documentFonts.EUC-KR.variable.italic*psmetric:		1000 1000 990 -10
*documentFonts.EUC-KR.variable.boldItalic*psname:	Munhwa-Regular-KSC-EUC-H
*documentFonts.EUC-KR.variable.boldItalic*psmetric:	1000 1000 990 -10
*documentFonts.EUC-KR.fixed.normal*psname:		MunhwaGothic-Regular-KSC-EUC-H
*documentFonts.EUC-KR.fixed.normal*psmetric:		1000 1000 990 -10
*documentFonts.EUC-KR.fixed.bold*psname:		MunhwaGothic-Regular-KSC-EUC-H
*documentFonts.EUC-KR.fixed.bold*psmetric:		1000 1000 990 -10
*documentFonts.EUC-KR.fixed.italic*psname:		MunhwaGothic-Regular-KSC-EUC-H
*documentFonts.EUC-KR.fixed.italic*psmetric:		1000 1000 990 -10
*documentFonts.EUC-KR.fixed.boldItalic*psname:		MunhwaGothic-Regular-KSC-EUC-H
*documentFonts.EUC-KR.fixed.boldItalic*psmetric:	1000 1000 990 -10

You can edit psname, psmetric, psafmfile for your purpose.

Open Issues:

19980627 snapshot

Japanese printing will be difficult. It follows document's codeset too much: e.g. previously Japanese output is JIS-encoded(almost all Japanese fonts is JIS-encoded) but now it depends on it's original encoding(SJIS or EUC?). I'll check it more. Of course, all encoding have same problem.
Decoration(title, date, etc..) is correct in only Letter-sized paper. When I set it to A4, Title is disappeared.
No check with Latin-2 or other one-byte non-latin1 fonts.

19980824 snapshot

Japanese printing problem is solved. To print, printer fonts charset should be SJIS. It can be adjusted in resource.
I tested it with mozilla built on FreeBSD-2.2-STABLE. Other platform test is required.