urosm has quit [Remote host closed the connection]
urosm has joined #foot
urosm has quit [Ping timeout: 260 seconds]
mtm has quit [Ping timeout: 264 seconds]
mtm has joined #foot
<WhyNotHugo>
For the '一' character, foot uses /usr/share/fonts/noto/NotoSansCJK-Bold.ttc if it is bold, or /usr/share/fonts/noto/NotoSerifTangut-Regular.ttf if it is non-bold.
<WhyNotHugo>
Funny that the chosen fallback font changes depending on bold/regular.
<WhyNotHugo>
This particular unicode point represents different characters depending on the language.
<WhyNotHugo>
So when regular I get a chinese characteer, but when bold I get a character of some other language.
<dnkl>
WhyNotHugo: you can blame that on fontconfig...
<dnkl>
in this case I believe fontconfig falls back to tangut before CJK. But, tangut only has the "regular" font. So, for a bike character it continues to the next fallback font. The CJK font happens to be it
<dnkl>
s/bike/bold
<WhyNotHugo>
Yeah, that seems to be the case. Still trying to figure out how to tell fontconfig to always prefer cjk instead of Tangut.
<WhyNotHugo>
I'd rather blame the unicode consortium here, for using the same codepoint for different characters in different languags. FontConfig has no way of knowing whether I'm writing Chinese of Tangul.
<dnkl>
what's the name of the codepoint?
an3223 has quit [Remote host closed the connection]
<sewn>
what is the project goal of fyi? if it was made to be an alternative, it has to have solved a problem unsolvable in notify-send.
<Consolatis>
application feedback about $things I assume, like the xdg activation token which you only get from notify-send by using hacky debug log options
<Consolatis>
also: great name
<WhyNotHugo>
It's U+4E00. I'm not sure how to figure out the unicode name.
<WhyNotHugo>
If I understand correctly, this unicode block has characters that look similar in various languages, so each codepoint can correspond to any of these (and any language).
<WhyNotHugo>
So you need additional metadata to infer a language and pick the right font.
<Arnavion>
Yes, Unicode has "Han unification" to use the same codepoint for the underlying meaning even if the glyph is different in Mandarin / Japanese / Cantonese / etc, and the renderer is expected to use out-of-band knowledge of the language to pick the correct glpyh
<dnkl>
well that's easy when the text can contain codepoints from many languages...
<WhyNotHugo>
It's a really funny choice considering that а and a have different codepoints but look identical.
<WhyNotHugo>
How is it easy? AFAIK, there's no way a regular text field, or this IRC message can include such metadata.
<dnkl>
WhyNotHugo: I was being sarcastic
<WhyNotHugo>
I was being slow :P
urosm has joined #foot
eqb has quit [Ping timeout: 260 seconds]
urosm has quit [Remote host closed the connection]
<WhyNotHugo>
If I press shift on one foot window, then focus (by moving the mouse) another foot window, the second window reacts as if shift was not pressed.
<WhyNotHugo>
Could this be a compositor issue?
<WhyNotHugo>
Basically if I want to select text on a window, I move my mouse towards that window while I press shift, and sometimes shift is pressed before the mouse has crossed over to the second window.
<vyivel>
does the second window receive wl_keyboard.modifiers?
<Consolatis>
sounds like a compositor issue, the 2nd foot window should receive a list of currently pressed keys when receiving focus