Re: [gentoo-dev] www-client/chromium and asian fonts, help needed

2011-04-07 Thread Paweł Hajdan, Jr.
On 4/5/11 6:37 PM, Ulrich Mueller wrote:
 Already chromium's dependency on virtual/ttf-fonts is wrong, IMHO.

Oh, I added it to RDEPEND after Raúl Porcel (armin76) asked me on IRC to
do that - I think the issue was that the browser failed to launch
without the fonts installed.

I don't remember all the details here, but do you have some idea for
better solution? I'd like the browser to be launchable out-of-the-box,
an elog message like Your browser will fail to start if you don't have
ttf-fonts wouldn't be OK I think.

 We had the same discussion for Emacs some time ago (see bug 137598),
 and we output a message in pkg_postinst() that Emacs requires fonts.

Yeah, I just added an elog for the Asian fonts here too.



signature.asc
Description: OpenPGP digital signature


Re: [gentoo-dev] www-client/chromium and asian fonts, help needed

2011-04-07 Thread Ulrich Mueller
 On Thu, 07 Apr 2011, Paweł Hajdan, wrote:

 Already chromium's dependency on virtual/ttf-fonts is wrong, IMHO.

 Oh, I added it to RDEPEND after Raúl Porcel (armin76) asked me on
 IRC to do that - I think the issue was that the browser failed to
 launch without the fonts installed.

 I don't remember all the details here, but do you have some idea for
 better solution?

As nirbheek has pointed out to me on IRC, freetype actually does font
rendering on the client side. So I stand corrected, and above
mentioned dependency is o.k.

 Yeah, I just added an elog for the Asian fonts here too.

Here I still think that this is a better solution than adding lots of
dependencies. I would expect that users have fonts for their language
installed in any case.

Ulrich



[gentoo-dev] www-client/chromium and asian fonts, help needed

2011-04-05 Thread Paweł Hajdan, Jr.
There is a bug about www-client/chromium and asian fonts:
http://bugs.gentoo.org/show_bug.cgi?id=359153

First, I'm just wondering whether making the browser RDEPEND on some
fonts would be a correct solution. Maybe, similarly to the icon themes,
we should just suggest some fonts in pkg_postinst.

I have no idea which of the fonts listed in the bug above handle
Chinese, Japanese, or Korean. It seems that it is important to
differentiate between them. Short example: for a Chinese page we need
the Chinese font, not just any CJK font. A Korean font might seem to
render most of the Japanese characters correctly, but it's not 100%
correct. Do you have some ideas how to handle that?



signature.asc
Description: OpenPGP digital signature


Re: [gentoo-dev] www-client/chromium and asian fonts, help needed

2011-04-05 Thread Mike Frysinger
On Tue, Apr 5, 2011 at 11:45 AM, Paweł Hajdan, Jr. wrote:
 First, I'm just wondering whether making the browser RDEPEND on some
 fonts would be a correct solution. Maybe, similarly to the icon themes,
 we should just suggest some fonts in pkg_postinst.

i dont think you should be depending on fonts.  a pkg_postinst elog
sounds fine though.
-mike



Re: [gentoo-dev] www-client/chromium and asian fonts, help needed

2011-04-05 Thread Chí-Thanh Christopher Nguyễn
Paweł Hajdan, Jr. schrieb:
 First, I'm just wondering whether making the browser RDEPEND on some
 fonts would be a correct solution. Maybe, similarly to the icon themes,
 we should just suggest some fonts in pkg_postinst.
   

You could extend the existing virtual/ttf-fonts with a new cjk USE-flag,
which would make it depend on one of the suggested fonts from bug 359153.


Best regards,
Chí-Thanh Christopher Nguyễn




Re: [gentoo-dev] www-client/chromium and asian fonts, help needed

2011-04-05 Thread Ulrich Mueller
 On Tue, 05 Apr 2011, Paweł Hajdan, wrote:

 There is a bug about www-client/chromium and asian fonts:
 http://bugs.gentoo.org/show_bug.cgi?id=359153

 First, I'm just wondering whether making the browser RDEPEND on some
 fonts would be a correct solution. Maybe, similarly to the icon
 themes, we should just suggest some fonts in pkg_postinst.

I think this would be conceptionally wrong. Fonts are a property of
the X server and shouldn't be a dependency of an X client (which may
even run on a different machine).

Already chromium's dependency on virtual/ttf-fonts is wrong, IMHO.

We had the same discussion for Emacs some time ago (see bug 137598),
and we output a message in pkg_postinst() that Emacs requires fonts.

Ulrich