Re: Suggest installing more fonts?

[Gregory Heytings]

> > > [ FWIW, I don't think replacing those tofu chars with some "ugly 
> > > default" is a good option: it's more work with very little benefit if 
> > > any, since the problem still remains largely unchanged. ]
> >
> > FWIW, I do not agree with this.
>
> I knid of guessed ;-)

Life would be boring with quick agreements ;-)

> > It is much better from a end-user viewpoint to see something they can 
> > decypher, even if it has a non-optimal rendering, than to see a hex 
> > code in a box.
>
> [ What you describe is the "very little benefit" I referred to.  ]

I don't understand why you consider this as a "very little benefit".  The 
Unifont is not unreadable, it is certainly not a nice modern font with 
antialiasing etc. , because it is on purpose as small as possible (8x16 or 
16x16 black and white bitmaps), but it is useable.

> I think most users won't be happy with the result in both cases, so 
> they'll still want to find a fix for the problem [ This is "the problem 
> still remains largely unchanged" that I referred to. ]

Here I agree with you: the problem remains largely unchanged, but at least 
Emacs would have done its best (at a reasonable cost) to help the user. 
And Emacs would do better than most other editors, which just show an 
empty box or a question mark in that case.

> > > - Does the problem only affect Emacs and not other applications? If so 
> > > why?  If not, then what do other applications do about it?
> >
> > This problem affects all applications.  If happens whenever an 
> > application tries to display a character for which the font in use has 
> > no glyph. What they do varies, most display a question mark in a box, 
> > or an empty box.
>
> I have seen several questions about Emacs showing tofu where the user 
> points out that such-and-such other application displays that same text 
> just fine on the same system.

This depends on many parameters, most of which are outside Emacs' control. 
And you will probably not see questions about the opposite situation, 
where other applications show tofus when Emacs displays the text just fine 
;-)  So I'll give you two examples (and I'm sure there are many, many 
more):

- On my computer, with a recent version of Chromium, the sentence "Can we 
see the full text of your speech before Tuesday?" 
(我们能在周二前看到你演讲的全文吗？) on 
https://dictionary.cambridge.org/dictionary/english-chinese-simplified/text 
contains three tofus (们讲吗), which Chromium displays as empty boxes. 
The same three tofus appear on Firefox, but Firefox displays them with 
four hex digits in a box.  Emacs displays that sentence fine, both in a 
GUI because I happen to have a certain "-isas-song 
ti-medium-r-normal--24-240-72-72-c-240-gb2312.1980-0" font installed, and 
in a terminal because my default font is DejaVu Sans, which contains all 
these characters.

- On my computer, two of the three missing "smiling face" emojis in DejaVu 
Sans are displayed fine by VS Code, because I happen to have the 
"fonts-noto-color-emoji" Debian package installed.  But the third one 
(0x1f972) is displayed as a tofu (an empty box) by VS Code, even with that 
package installed.  If I remove that package, VS Code displays these three 
emojis as tofus, as Emacs does.