Tries to retrieve all post_match_families in the font to support non-ascii fo...

Tries to retrieve all post_match_families in the font file to support non-ascii fonts. This patch should fix CHECK failures like this:

yusukes@yusukes-desktop:~/chromium2/src/build$ ../sconsbuild/Release/chrome
[20666:20666:544922424350:FATAL:/home/yusukes/chromium2/src/app/gfx/font_skia.cc(90)]
Check failed: tf. Could not find font: IPA モナー Pゴシック

Since some fonts have multiple font family names, the FontConfigDirect::Match() function should try to compare post_config_family with _all_ family names in the font, rather than with only the first (usually English) family name. For example, a popular Japanese font called ipagpmona.ttf has two family names, an English name and an internationalized (Japanese) one, and post_config_family for the font can be either the English name or the Japanese name depending on the situation.

$ showttf /usr/share/fonts/truetype/ttf-ipamonafont/ipagp-mona.ttf | egrep -A 1 -e
"platform=1 .* Family" 
   platform=1 plat spec encoding=0 language=0 name=1 Family
    strlen=14  stroff=288 IPAMonaPGothic
--
   platform=1 plat spec encoding=1 language=b name=1 Family
    strlen=20  stroff=871 IPA <83><82><83>i<81>[ P<83>S<83>V<83>b<83>N

BUG=12530
TEST=none

[agl, 2009-07-18 20:39:59]

You should copy more of your comment #77 into the change log for this patch.
Otherwise, and with a couple of style nits, LGTM. Thanks.

http://codereview.chromium.org/155751/diff/5/1004
File skia/ext/SkFontHost_fontconfig_direct.cpp (right):

http://codereview.chromium.org/155751/diff/5/1004#newcode158
Line 158: while (id < 255) {
this should be a for() loop.

http://codereview.chromium.org/155751/diff/5/1004#newcode171
Line 171: if ((!family.empty()) && (!family_names_match)) {
remove superfluous parens

[Yusuke Sato, 2009-07-19 00:30:36]

Thanks for the review. I've revised the CL description.
Please take another look.

On 2009/07/18 20:39:59, agl wrote:
> You should copy more of your comment #77 into the change log for this patch.
> Otherwise, and with a couple of style nits, LGTM. Thanks.
> 
> http://codereview.chromium.org/155751/diff/5/1004
> File skia/ext/SkFontHost_fontconfig_direct.cpp (right):
> 
> http://codereview.chromium.org/155751/diff/5/1004#newcode158
> Line 158: while (id < 255) {
> this should be a for() loop.
> 
> http://codereview.chromium.org/155751/diff/5/1004#newcode171
> Line 171: if ((!family.empty()) && (!family_names_match)) {
> remove superfluous parens

[Yusuke Sato, 2009-07-19 00:30:54]

http://codereview.chromium.org/155751/diff/5/1004
File skia/ext/SkFontHost_fontconfig_direct.cpp (right):

http://codereview.chromium.org/155751/diff/5/1004#newcode158
Line 158: while (id < 255) {
On 2009/07/18 20:39:59, agl wrote:
> this should be a for() loop.

Done.

http://codereview.chromium.org/155751/diff/5/1004#newcode171
Line 171: if ((!family.empty()) && (!family_names_match)) {
On 2009/07/18 20:39:59, agl wrote:
> remove superfluous parens

Done.

[agl, 2009-07-19 01:12:31]

LGTM

[Yusuke Sato, 2009-07-19 02:41:42]

Landed in r21063. Thanks.
http://codereview.chromium.org/159061

--Yusuke

On 2009/07/19 01:12:31, agl wrote:
> LGTM