Fontconfig - Nov 2005 - How to disable the "global advance" for CJK fonts?

Ambrose> Is there an "official" way
Ambrose> to disable the "global advance" for CJK fonts? 
http://cvs.mandrakesoft.com/cgi-bin/cvsweb.cgi/SPECS/fontconfig/fontconfig-2.2.96-dual_width.patch

Hi,

sorry if this has been asked before. Is there an "official" way
to disable the "global advance" for CJK fonts? From what I''ve
read in fonts.conf, this should be possible. But when I tried
it in local.conf, I did not get any results. fontconfig version
is 2.3.0.

(In fact, from the looks of how my mozilla renders pages, it
looks like even the default list in fonts.conf is not working)-:

Thanks very much in advance.

regards,
- ambrose

Hi,

On Fri, Mar 18, 2005 at 01:57:56PM -0800, Keith Packard
wrote:
> 
> Feel free to send along more family names that need this treatment and 
> I''ll stick ''em in the next release.
I believe currently (for whatever reason I don''t think this used to
be the case) all the Arphic fonts need to have global advance disabled.
With some others familiar ones, at least the following common fonts
must have "global advance" disabled:

	AR PL KaitiM Big5
	AR PL SungtiL GB
	AR PL Mingti2L Big5
	AR PL KaitiM GB
	Kochi Gothic
	Kochi Mincho
	Ming(for ISO10646)	[i.e., HK government''s reference font]
	MS Hei
	MS Song

There are also other free and freeware fonts that don''t work, e.g.,
Mikachan, Y.OzFont, MT_TARE.

(I don''t know whether the Win2000 equivalents of the above MS fonts
would cause problems; I don''t have them for checking.  MS Gothic,
plus the Baekmuk and Wang fonts, do not seem to be affected.)

But I think the real question we should as is: Why do we need to do
this at all? All these fonts work right out of the box in Windows
(any version from Win95 onwards) and it seems also in Mac OSX, and
many of these are system fonts in Linux and/or Windows. I tend to
feel that we (i.e., the Free Software side) must be doing something
fundamentally wrong, or at least something "de-factoly" wrong. We
seem to be dealing with an "exception" list of unknown length and
the list will be growing and it will never be complete.

Some time ago, iirc, someone else proposed on this list that global
advance be turned off by default for all CJK fonts; from what I
observe that would be an entirely reasonable thing, and I''d like to
know any technical reason why that should not be the case.

Regards,
- Ambrose

Keith> Feel free to send along more family names that need this treatment and
Keith> I''ll stick ''em in the next release.
I''m afraid it applies to all the Chinese fonts. Almost all the Chinese
fonts have built-in half-width and full-width characters. It would be
better fix it at libxft level(bug#2715).

Around 2 o''clock on Mar 19, Ambrose wrote:
> Ambrose> Is there an "official" way
> Ambrose> to disable the "global advance" for CJK fonts? 
In fontconfig 2.3, there are several CJK fonts configured to disable 
global advance already.

Check out the fonts.conf where you''ll find:

<!--
  Some Asian fonts misadvertise themselves as monospaced when
  in fact they are dual-spaced (half and full).  This makes
  FreeType very confused as it forces all widths to match.
  Undo this magic by disabling the width forcing code -->
        <match target="font">
                <test
name="family"><string>GulimChe</string></test>
                <edit
name="globaladvance"><bool>false</bool></edit>
        </match>

        <match target="font">
                <test
name="family"><string>DotumChe</string></test>
                <edit
name="globaladvance"><bool>false</bool></edit>
        </match>

        <match target="font">
                <test
name="family"><string>BatangChe</string></test>
                <edit
name="globaladvance"><bool>false</bool></edit>
        </match>

        <match target="font">
                <test
name="family"><string>GungsuhChe</string></test>
                <edit
name="globaladvance"><bool>false</bool></edit>
        </match>

Feel free to send along more family names that need this treatment and 
I''ll stick ''em in the next release.

-keith


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
Url :
http://lists.freedesktop.org/archives/fontconfig/attachments/20050318/69d03f6e/attachment.pgp

On Fri, Mar 18, 2005 at 11:50:43PM -0500, Ambrose Li
wrote:
> 
> 	AR PL KaitiM Big5
> 	AR PL SungtiL GB
> 	AR PL Mingti2L Big5
> 	AR PL KaitiM GB
> 	Kochi Gothic
> 	Kochi Mincho
> 	Ming(for ISO10646)	[i.e., HK government''s reference font]
> 	MS Hei
> 	MS Song
I forgot to write that MingLiU should be in the list too.

Honestly, in my collection of CJK fonts, the list of "exceptions"
is about the same length as the list of correctly-handled fonts
(if not longer).  When you have the number of exceptions the same
as the number of non-exceptions, I don''t think the exceptions can
be rightly called exceptions at all.

Around 0 o''clock on Mar 19, Ambrose Li wrote:
> Honestly, in my collection of CJK fonts, the list of "exceptions"
> is about the same length as the list of correctly-handled fonts
> (if not longer).
I labeled it as an exceptional case because FreeType did (not to disclaim 
responsibility, but to notice that I used no first-hand knowledge of fonts 
when making this decision).
>From what people explained to me at the time, the font itself says whether 
to use global advance and this actually instructs FreeType to disregard 
that instruction.

Is there some sensible way to automatically tell whether global advance is 
giving bogus values?  I don''t even care how expensive it is to compute
as
we can do it at font scan time instead of font load time.

-keith


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
Url :
http://lists.freedesktop.org/archives/fontconfig/attachments/20050318/b2b42bb7/attachment.pgp

Hi there,

Just trying to clear my confusion. In original mail, the modification of
fonts.conf shows an edit clause on "globaladvance" with 
binding="strong". I
don''t know what effect will the strong-binding bring. I dug source
codes of
fontconfig some time ago, AFAIK, strong-binding only make family name
priorer than lang. Is things changed a lot?

Thanks for any hint.

Xie Qian

Open System and Chinese Information Processing Center,
Institute of Software, Chinese Academy of Sciences,
P.O. Box 8718, Beijing 100080, CHINA

----- Original Message -----
From: "Mike FABIAN" <mfabian@suse.de>
To: "Ambrose Li" <acli@ada.dhs.org>
Cc: "Keith Packard" <keithp@keithp.com>;
<fontconfig@fontconfig.org>
Sent: Wednesday, June 29, 2005 9:58 PM
Subject: [Fontconfig] Re: How to disable the "global advance" for CJK
fonts?

> Ambrose Li <acli@ada.dhs.org> $B$5$s$O=q$-$^$7$?(B:
>
> > Some time ago, iirc, someone else proposed on this list that global
> > advance be turned off by default for all CJK fonts; from what I
> > observe that would be an entirely reasonable thing, and I''d
like to
> > know any technical reason why that should not be the case.
>
> Currently I use the following workaround, which seems to work OK:
>
> <!--
>   The dual-width Asian fonts (spacing=dual) are not rendered correctly,
>   apparently FreeType forces all widths to match. Trying to disable the
>   width forcing code by setting globaladvance=false alone doesn''t 
help.
>   As a brute force workaround, also set spacing=proportional, i.e. handle
>   them as proportional fonts:
> -->
>
>         <match target="font">
>                 <test name="lang"
compare="contains">
>                         <string>ja</string>
>                         <string>zh</string>
>                         <string>ko</string>
>                 </test>
>                 <test name="spacing"
compare="eq">
>                         <const>dual</const>
>                 </test>
>                 <edit name="spacing">
>                         <const>proportional</const>
>                 </edit>
>                 <edit name="globaladvance"
binding="strong">
>                         <bool>false</bool>
>                 </edit>
>         </match>
>
> --
> Mike FABIAN   <mfabian@suse.de>   http://www.suse.de/~mfabian
> $B?gL2ITB-$O$$$$;E;v$NE($@!#(B
>
> _______________________________________________
> fontconfig mailing list
> fontconfig@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/fontconfig
>

Keith> Is global advance *ever* useful?  Should the default in Xft be to
ignore
Keith> global advance unless someone specifically sets
Keith> the flag? It seems like this should be the case as Fontconfig detects
Keith> monospaced fonts by disabling global advance and comparing glyph
widths
Keith> without it.
https://bugs.freedesktop.org/show_bug.cgi?id=2715

Ambrose Li <acli@ada.dhs.org> ????????:
> Some time ago, iirc, someone else proposed on this list that global
> advance be turned off by default for all CJK fonts; from what I
> observe that would be an entirely reasonable thing, and I''d like
to
> know any technical reason why that should not be the case.
Currently I use the following workaround, which seems to work OK:

<!--
  The dual-width Asian fonts (spacing=dual) are not rendered correctly,
  apparently FreeType forces all widths to match. Trying to disable the
  width forcing code by setting globaladvance=false alone doesn''t 
help.
  As a brute force workaround, also set spacing=proportional, i.e. handle
  them as proportional fonts:
-->

        <match target="font">
                <test name="lang" compare="contains">
                        <string>ja</string>
                        <string>zh</string>
                        <string>ko</string>
                </test>
                <test name="spacing" compare="eq">
                        <const>dual</const>
                </test>
                <edit name="spacing">
                        <const>proportional</const>
                </edit>
                <edit name="globaladvance"
binding="strong">
                        <bool>false</bool>
                </edit>
        </match>

-- 
Mike FABIAN   <mfabian@suse.de>   http://www.suse.de/~mfabian
?????????????

On Wed, 2005-06-29 at 15:58 +0200, Mike FABIAN wrote:
> Ambrose Li <acli@ada.dhs.org> ????????:
> 
> > Some time ago, iirc, someone else proposed on this list that global
> > advance be turned off by default for all CJK fonts; from what I
> > observe that would be an entirely reasonable thing, and I''d
like to
> > know any technical reason why that should not be the case.
Is global advance *ever* useful?  Should the default in Xft be to ignore
global advance unless someone specifically sets
the flag? It seems like this should be the case as Fontconfig detects
monospaced fonts by disabling global advance and comparing glyph widths
without it.
> Currently I use the following workaround, which seems to work OK:
> 
> <!--
>   The dual-width Asian fonts (spacing=dual) are not rendered correctly,
>   apparently FreeType forces all widths to match. Trying to disable the
>   width forcing code by setting globaladvance=false alone doesn''t 
help.
>   As a brute force workaround, also set spacing=proportional, i.e. handle
>   them as proportional fonts:
> -->
FreeType isn''t responsible for interpreting the spacing value,
that''s
purely in Xft. That change is in Xft version 2.1.7, which was released
2005-3-29.

-keith

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url :
http://lists.freedesktop.org/archives/fontconfig/attachments/20050629/08f9f252/attachment.pgp