bug#39339: 26.3; set-fontset-font does not replace some glyphs

[Eli Zaretskii]

tags 39339 notabug
thanks

> From: ynyaaa@gmail.com
> Cc: 39339@debbugs.gnu.org
> Date: Thu, 30 Jan 2020 15:06:55 +0900
> 
> >> (let ((base-font "Noto Sans Mono")
> >>       (alt-font "Noto Sans Mono CJK JP Regular")
> >>       (l '(#x00A7 #x00B1 #x00B6 #x00D7 #x00F7
> >>                   #x02C7 #x02C9 #x02CA #x02CB #x02D9
> >>                   #x2015 #x2016 #x2020 #x2021 #x2025 #x2027 #x2030
> >>                   #x2035 #x203B #x2042 #x2051
> >>                   #x2100 #x2103 #x2105 #x2109 #x210A #x210F #x2116
> >>                   #x2121 #x2126 #x2127 #x212B #x2135 #x213B
> >>                   #x221A #x221E #x2229 #x2248 #x2261 #x2264 #x2265
> >>                   (#x2500 . #x25A0))))
> >>   (set-frame-font base-font)
> >>   (dolist (c l) (set-fontset-font t c alt-font)))
> >
> > This makes "Noto Sans Mono" the default face's font, so it will be
> > used for any character supported by the font.  If that's what you see,
> > then it's the intended behavior, AFAIK.
> 
> With '(setq use-default-font-for-symbols nil)', the glyphs are displayed
> as I intended.

Good, then there's no bug here.

> How can I find this variable?

But you just did (with a bit of help from me).

If you are asking about its discoverability, then this variable is
supposed to be important in very infrequent situations, so it is not
very visible.  As evidence to that effect, please note that by default
Emacs does display these characters, which is not a small feat for
Emacs to achieve, given that most fonts on Windows are not free.

> And what characters are considered to be symbols and punctuation?

Any character CH for which

  (aref char-script-table CH) => symbol