Bug 578051

Summary:	lang-specific overrides rule doesn't work as expected
Product:	[Fedora] Fedora	Reporter:	Akira TAGOH <tagoh>
Component:	wqy-zenhei-fonts	Assignee:	Peng Wu <pwu>
Status:	CLOSED ERRATA	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	medium	Docs Contact:
Priority:	low
Version:	rawhide	CC:	fangqq, i18n-bugs, petersen, phuang, pwu
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2010-05-04 02:20:48 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:	578015
Bug Blocks:	507684

Description Akira TAGOH 2010-03-30 04:00:37 UTC

Description of problem:
All of detailed information is available on my post at the fonts list:
http://lists.fedoraproject.org/pipermail/fonts/2010-March/001117.html

binding="same" in the fontconfig config file prevents to apply the rule for the
specific
language only properly. As a result, fonts is used for non-targetted languages
and it may gives different look and feel in some cases.

I'd propose to get rid of binding="same" from:
 65-wqy-zenhei.conf

Comment 1 Peng Wu 2010-04-07 02:32:43 UTC

I tested the changes on Fedora 13, and the result is as following:
(font config file is from http://pwu.fedorapeople.org/fonts-conf/).

1. Before removing binding="same" or binding="strong":
+ fc-match sans
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "中等"
+ fc-match sans:lang=zh
DejaVuSans.ttf: "DejaVu Sans" "Book"
+ fc-match sans:lang=zh-cn
DejaVuSans.ttf: "DejaVu Sans" "Book"
+ fc-match sans:lang=zh-tw
DejaVuSans.ttf: "DejaVu Sans" "Book"
+ fc-match sans:lang=zh-hk
DejaVuSans.ttf: "DejaVu Sans" "Book"
+ fc-match sans:lang=en
DejaVuSans.ttf: "DejaVu Sans" "Book"
+ fc-match sans:lang=en-us
DejaVuSans.ttf: "DejaVu Sans" "Book"

2. After removing binding="same" or binding="strong":
+ fc-match sans
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "中等"
+ fc-match sans:lang=zh
uming.ttc: "AR PL UMing TW" "Light"
+ fc-match sans:lang=zh-cn
uming.ttc: "AR PL UMing TW" "Light"
+ fc-match sans:lang=zh-tw
uming.ttc: "AR PL UMing TW" "Light"
+ fc-match sans:lang=zh-hk
uming.ttc: "AR PL UMing TW" "Light"
+ fc-match sans:lang=en
DejaVuSans.ttf: "DejaVu Sans" "Book"
+ fc-match sans:lang=en-us
DejaVuSans.ttf: "DejaVu Sans" "Book"

It seems that removing the binding="same" helps split between English and Chinese fonts.

Comment 2 Qianqian Fang 2010-04-07 18:02:03 UTC

similar to Bug#568613, I am taking myself off and let Peng to handle these bugs.

Comment 3 Peng Wu 2010-04-19 06:14:30 UTC

Fixed in wqy-zenhei-fonts-0.9.45-2.fc13.

Comment 4 Fedora Update System 2010-04-19 06:34:22 UTC

wqy-zenhei-fonts-0.9.45-2.fc13 has been submitted as an update for Fedora 13.
http://admin.fedoraproject.org/updates/wqy-zenhei-fonts-0.9.45-2.fc13

Comment 5 Fedora Update System 2010-04-20 13:13:27 UTC

wqy-zenhei-fonts-0.9.45-2.fc13 has been pushed to the Fedora 13 testing repository.  If problems still persist, please make note of it in this bug report.
 If you want to test the update, you can install it with 
 su -c 'yum --enablerepo=updates-testing update wqy-zenhei-fonts'.  You can provide feedback for this update here: http://admin.fedoraproject.org/updates/wqy-zenhei-fonts-0.9.45-2.fc13

Comment 6 Akira TAGOH 2010-04-26 05:50:59 UTC

Testing result:
# rpm -qa wqy-zenhei-fonts
wqy-zenhei-fonts-0.9.45-2.fc13.noarch
# fc-match monospace:lang=en
DejaVuSansMono.ttf: "DejaVu Sans Mono" "Book"
# fc-match monospace:lang=en-us
DejaVuSansMono.ttf: "DejaVu Sans Mono" "Book"
# fc-match monospace:lang=zh
wqy-microhei.ttc: "WenQuanYi Micro Hei" "Regular"
# fc-match monospace:lang=zh-cn
wqy-zenhei.ttc: "文泉驛正黑" "Regular"
# fc-match monospace:lang=zh-hk
hanazono.ttf: "HanaMin" "Regular"
# fc-match monospace:lang=zh-sg
wqy-zenhei.ttc: "文泉驛正黑" "Regular"
# fc-match monospace:lang=zh-tw
hanazono.ttf: "HanaMin" "Regular"

Though the above result looks good, this isn't by 65-wqy-zenhei.conf. but by the implicit rules in fontconfig that happens on when no appropriate fonts matched by the rules. so

  <test name="lang">
    <string>zh</string>
  </test>

  should be:

  <test name="lang">
    <string>zh-cn</string>
    <string>zh-sg</string>
  </test>

# fc-match sans:lang=en
DejaVuSans.ttf: "DejaVu Sans" "Book"
# fc-match sans:lang=en-us
DejaVuSans.ttf: "DejaVu Sans" "Book"
# fc-match sans:lang=zh
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"
# fc-match sans:lang=zh-cn
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"
# fc-match sans:lang=zh-hk
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"
# fc-match sans:lang=zh-sg
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"
# fc-match sans:lang=zh-tw
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"

I'd add same comment above for this one as well. also that would be good to rename 65-wqy-zenhei.conf to 65-0- or something like that to apply this prior to 65-nonlatin.conf.

# fc-match serif:lang=en
DejaVuSerif.ttf: "DejaVu Serif" "Book"
# fc-match serif:lang=en-us
DejaVuSerif.ttf: "DejaVu Serif" "Book"
# fc-match serif:lang=zh
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"
# fc-match serif:lang=zh-cn
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"
# fc-match serif:lang=zh-hk
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"
# fc-match serif:lang=zh-sg
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"
# fc-match serif:lang=zh-tw
wqy-zenhei.ttc: "WenQuanYi Zen Hei" "Regular"

No explicit rules for serif in wqy-zenhei-fonts. this would be a kind of side-effect of 65-nonlatin.conf.

# for i in $(ls /usr/share/locale|grep -v -E "^(zh)$"); do fc-match sans:lang=$i|grep -E "^(WenQuanYi Zen Hei)$"; [ $? = 0 ] && echo $i; done
#

No affects to other languages.