[Konsole-devel] [Bug 53713] double spaced on some of the font

llch at redhat.com llch at redhat.com
Wed Feb 5 13:03:19 UTC 2003


------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
     
http://bugs.kde.org/show_bug.cgi?id=53713     




------- Additional Comments From llch at redhat.com  2003-02-05 14:03 -------
Subject: Re:  double spaced on some of the font         

On 5 Feb 2003, Waldo Bastian wrote:
> Shouldn't it use the maximum? You want that the whole character fits your cell 
> for all (single width) characters. If you average, you pretty much get that 
> half of the chars fit and the other half doesn't.

Popular portorional font like ZYSong18030 and Baekmuk Gulim has width from
7 to 12 even in english/symbol glyphs. So for example when using one of
the font with getting the max. width, the display will seen as double
spaced. I think it is between a trade off of getting max width to have
huge space on some portorional font, or getting avg width to have tight
space on some of portorional font. Of course it would be great if we could
think of other solutions.

> 
> Thanks, something like this is indeed needed. How bad does this suck in terms 
> of performance? 

I haven't notice any performance loses in my PIII machine. Any performance 
measurement recommend for me to do on konsole?

> 
> Note that "use double width if not ascii" is not correct, we keep track 
> internally of characters that are supposed to be double width and that 
> information should be used instead.
> 
> E.g. setImage knows when a character is double width, because the next 
> character will be 0. What might work is to pass fragments to drawAttrStr that 
> consist of either only single width or only double width characters and add a 
> bool isDoubleWidth to drawAttrStr. Then drawAttrStr can do "cell drawing" 
> using the specified width.
> 
> It would be nice if you could make a patch based on the above, if not I will 
> have a go at it myself in the second half of february.
> 

Right agreed. Please see the following patch. I made it to pass a cell at
a time so that it will know when to render a double width characters, as
well then we can omit the loop in drawAttrStr(). So far so good remotely,
however I will get the performance review done locally tomorrow.



More information about the konsole-devel mailing list