Re: [dev] [st] possible bug with unicode character

From: Marc Collin <marc.collin7_AT_gmail.com>
Date: Tue, 12 Apr 2016 16:31:15 -0300

So this is not a st bug and not a libc bug.
Just a font bug?
I'll test a few fonts and check which doesn't work, then I'm emailing
the authors of the fonts.
How can we easily describe the problem so the font authors can fix it?
Thanks.



On Tue, Apr 12, 2016 at 3:46 PM, Jochen Sprickerhof
<dwm_AT_jochen.sprickerhof.de> wrote:
> I found the same last December and had a look into it, but my conclusion was
> that it's a bug in the font/libc, so I didn't reported my findings here.
>
> From my notes back then:
> - st expects chars to have a fixed width (see [1]).
> - It uses wcwidth(3) to get the length and reserves as much space on the
> screen.
> - For a lot of UTF-8 emojis and symbols wcwidth return -1 (at least on
> my Debian unstable libc6 2.22-6) and st assumes a width of 1 for them
> (st.c:2937 in current HEAD).
> - If the symbol in your font is actually bigger, parts of it are
> overdrawn.
>
> So it depends on your wcwidth and font to get this effect. Patches for
> both welcome ;).
>
> Cheers Jochen
>
> [1] https://en.wikipedia.org/wiki/Halfwidth_and_fullwidth_forms
Received on Tue Apr 12 2016 - 21:31:15 CEST

This archive was generated by hypermail 2.3.0 : Tue Apr 12 2016 - 21:36:12 CEST