Wide symbols in Kate

[tkzv]

Is there any way to make all symbols in Kate strictly either single or double width? Right now, European:CJK width ratio is something like 3:5, and I'd like 1:2, like in Konsole.

I was told that this may be possible with a specific CJK font, but he didn't know if such font exist. Is there one?

[tkzv]

Kate is set to use "Monospace" font. Which package is that?

How do I find, which fonts Kate uses for Japanese symbols?

I got the impression, that Kate uses different fonts for digits in lines "秋23456789" and "あ23456789". Or even different digits left and right of "a" in "あ123a123". How do I check that?

After changing settings in "eselect fontconfig" how do I apply changes?
Do I need "fc-cache -fv" ?
Is it enough to restart Kate?
"kwin_x11 --replace" ?
Exit X Window and start it again?

[flexibeast]

Kate is set to use "Monospace" font. Which package is that?

How do I find, which fonts Kate uses for Japanese symbols?

Assuming Kate does indeed use Fontconfig (which it probably does, check the output of `ldd /usr/bin/kate | grep fontconfig`) the wiki's "Fontconfig" page has information about this, some of which i describe below.

"Monospace" in this context is not a particular package, but the first appropriate result returned by Fontconfig when Kate asks for a monospace font. This can be checked by asking Fontconfig directly, via fc-match(1):

Code: Select all

fc-match -s 'mono' | head -n1

The `-s` option sorts the output from 'best' to 'worst' match.

Then, to find which font will be returned in the context of Japanese, add `lang` to the query:

Code: Select all

fc-match -s 'mono:lang=jp' | head -n1

Remove `| head -n1` from both commands to list all fonts that match.

[tkzv]

ldd /usr/bin/kate | grep fontconfig

Thanks, it does use fontconfig

Code: Select all

fc-match -s 'mono' | head -n1

The `-s` option sorts the output from 'best' to 'worst' match.

Code: Select all

fc-match -s 'mono:lang=jp' | head -n1

The "best" for both queries is `DejaVuSansMono.ttf: "DejaVu Sans Mono" "Book"` It has no hieroglyphs.
How does this best-to-worst-matching work? Do programs scan this list and use the first font that has the required symbol? If yes, the topmost with CJK is `wqy-zenhei.ttc: "WenQuanYi Zen Hei Mono" "Regular"`

Somehow, it appears in Kate's font selection dialog only when "Show only monospaced fonts" is unchecked. Bug?

But yes, it seems that the line "あ23456789" uses digits from "WenQuanYi Zen Hei Mono" and the line "秋23456789" uses digits from "DejaVu Sans Mono".
The line "あ234f6789" uses WenQuanYi for "234" and DejaVu for "6789".

At least one question sorted out. Thanks.
But how does Kate pick them?
Does it assume that every symbol after Japanese-exclusive "あ" is also Japanese until it encounters end-of-line or a European letter? Can I disable this behavior?
And why does it behave correctly when I use "Droid" font instead of "Monospace"?

Could you answer several more Fontconfig questions?

I disabled all 3 DejaVu configs in "eselect fontconfig" and refreshed the cache. What should be the result? I still get DejaVu on top of fontconfig queries.

I've got several CJK fonts. How do I set the preferred one? With something like

Code: Select all

<test name="lang">
            <string>ja</string>
</test>
<prefer>
                        <family>WenQuanYi Zen Hei Mono</family>
</prefer>

in ~/.config/fontconfig/fonts.conf ?

Is it possible to use fontconfig to set double symbol width e.g. when "WenQuanYi Zen Hei Mono" supplies symbols missing in "Droid Sans Mono Dotted"? Preferrably, by increasing spacing between symbols rather than stretching them.