On 03/09/2012 02:04 PM, Nicolas Mailhot wrote:
No the issue is that you're putting the font stack in a situation
where it's guessing and can't produce reliable results.
There is nothing wrong with guessing. In fact, for a mature
OS designed for the mass, fault-tolerance and legitimate
guess/default values are almost required. A robust system
should also function reasonably even under unexpected
situations (and I believe it is fair to inspect the font
rendering pipelines under these more stringent requirements
given the profound impact).
All I was trying to say was that the current guess is a bad one.
If it can be improved, it can save many people's time and
frustration. Tell what's the downside for correcting this?
Have your Chinese pages properly declare they are in Chinese as the
w3c recommends for i18n sites and you won't have any problems.
There is going to be no criminals if we are all law-abiding
citizens, but unfortunately, that's not our world.