high CPU load under zh_CN-UTF-8 locale

[bxuefeng]

On 09:25 Fri 20 Jan 2006, Marc Lehmann wrote:
> > > No idea. Do you use any input methods? What happens when you don't use any
> > > input methods?
> > 
> > I think the input method is causing the problem. Now the question is perhaps
> > how to make SCIM compatible with urxvt.
> 
> SCIM is compatible and works fine. What takes this long is the input
> method extension searching for a suitable font, which is almost completely
> out of rxvt-unicodes control.

I don't understand. Why under en_US.UTF-8 locale, there is no slow down and
why suddenly under Chinese locale, it began to search a suitable font?

> You can try setting a suitable font via the "imFont" locale, this might or
> might not improve the time required to find a font (-set).

Can you give me an example on how to set a font via 'imFont'? Shall I use the
font name from xfontsel? I don't really know how to do this. Again, I want to
know why urxvt does not fuss with en locales. 

> In any case, using urxvtd is likely to reduce this overhead immensely except
> for the first start.
Actually I did it. But in both way, when the locale is zh, the process X
still takes 99% of the CPU time. 

－－
regards
phyrster

You can never trust a woman; she may be true to you.