min-colors 88

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

I keep seeing 88 for min-colors.  I wonder if a different number would be more appropriate, as it seems 
that 88 is quite arbitrary.

Re: min-colors 88

[Yuri Khan]

On Thu, 27 Jan 2022 at 13:14, fatiparty--- via Users list for the GNU
Emacs text editor <help-gnu-emacs@gnu.org> wrote:

> I keep seeing 88 for min-colors.  I wonder if a different number would be more appropriate, as it seems
> that 88 is quite arbitrary.

Historically, there existed terminals that supported a maximum of 8
colors, 16 colors, 88 colors (basic 16 + 4×4×4 color cube + 8 grays?),
then 256 colors (basic 16 + 6×6×6 cube + 24 grays), and finally 24-bit
color.

Unless you intend to make a color theme for wide distribution, I
suggest configuring your own terminal emulators for 24-bit color and
always assuming full color range.

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

Yuri Khan wrote:

>> I keep seeing 88 for min-colors. I wonder if a different
>> number would be more appropriate, as it seems that 88 is
>> quite arbitrary.
>
> Historically, there existed terminals that supported
> a maximum of 8 colors, 16 colors, 88 colors (basic 16 +
> 4×4×4 color cube + 8 grays?), then 256 colors (basic 16 +
> 6×6×6 cube + 24 grays), and finally 24-bit color.

24-bit color, which confusingly gives you an 8-bit image, since

  $ echo $(( (2**8)**3 )) = $(( 256**3 )) # 16 777 216
              ^^^^^^^^ one byte per each RGB colors

> Unless you intend to make a color theme for wide
> distribution, I suggest configuring your own terminal
> emulators for 24-bit color and always assuming full
> color range.

The sky is the limit ...

-- 
underground experts united
https://dataswamp.org/~incal

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

[-- Attachment #1: Type: text/plain, Size: 837 bytes --]

Yuri Khan wrote:

>> I keep seeing 88 for min-colors. I wonder if a different
>> number would be more appropriate, as it seems that 88 is
>> quite arbitrary.
>
> Historically, there existed terminals that supported
> a maximum of 8 colors, 16 colors, 88 colors (basic 16 +
> 4×4×4 color cube + 8 grays?), then 256 colors (basic 16 +
> 6×6×6 cube + 24 grays), and finally 24-bit color.

Were there no monochrome terminals?

My first computer was the Mac Plus:

  A compact Mac, the Plus has a 9-inch (23 cm) 512x342 pixel
  monochrome display with a resolution of 72 PPI, identical to
  that of previous Macintosh models. [1]

This didn't prevent it from having awesome graphics,
sometimes:

  https://dataswamp.org/~incal/pimgs/dc.png

1986 game!

You can run on Linux with the Mini vMac emulator [2] - mail me
if you need ROMs ...


[-- Attachment #2: dc.png --]
[-- Type: image/png, Size: 26930 bytes --]

[-- Attachment #3: Type: text/plain, Size: 164 bytes --]


[1] https://en.wikipedia.org/wiki/Macintosh_Plus
[2] https://www.gryphel.com/c/minivmac/dnld_std.html

-- 
underground experts united
https://dataswamp.org/~incal

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 27, 2022, 18:33 by yuri.v.khan@gmail.com:

> On Thu, 27 Jan 2022 at 13:14, fatiparty--- via Users list for the GNU
> Emacs text editor <help-gnu-emacs@gnu.org> wrote:
>
>> I keep seeing 88 for min-colors.  I wonder if a different number would be more appropriate, as it seems
>> that 88 is quite arbitrary.
>>
>
> Historically, there existed terminals that supported a maximum of 8
> colors, 16 colors, 88 colors (basic 16 + 4×4×4 color cube + 8 grays?),
> then 256 colors (basic 16 + 6×6×6 cube + 24 grays), and finally 24-bit
> color.
>

I can understand 88 now


> Unless you intend to make a color theme for wide distribution, I
> suggest configuring your own terminal emulators for 24-bit color and
> always assuming full color range.
>
Does one still use (min-colors 88) or should this be changed (or not-used, neglected).

Re: min-colors 88

[Eli Zaretskii]

> Date: Thu, 27 Jan 2022 07:14:15 +0100 (CET)
> From: fatiparty--- via Users list for the GNU Emacs text editor <help-gnu-emacs@gnu.org>
> 
> I keep seeing 88 for min-colors.  I wonder if a different number would be more appropriate, as it seems 
> that 88 is quite arbitrary.

It is not arbitrary: xterm (and its workalikes) has a mode where it
supports 88 colors.  That number is enough for Emacs to express all of
the possible colors from the X repertoire, so we use that as a
threshold for "any color is supported".

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 27, 2022, 18:46 by eliz@gnu.org:

>> Date: Thu, 27 Jan 2022 07:14:15 +0100 (CET)
>> From: fatiparty--- via Users list for the GNU Emacs text editor <help-gnu-emacs@gnu.org>
>>
>> I keep seeing 88 for min-colors.  I wonder if a different number would be more appropriate, as it seems 
>> that 88 is quite arbitrary.
>>
>
> It is not arbitrary: xterm (and its workalikes) has a mode where it
> supports 88 colors.  That number is enough for Emacs to express all of
> the possible colors from the X repertoire, so we use that as a
> threshold for "any color is supported".
>
Would you know what colour bit would the 88 colours correspond to?

Re: min-colors 88

[Eli Zaretskii]

> Date: Thu, 27 Jan 2022 09:43:53 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> > It is not arbitrary: xterm (and its workalikes) has a mode where it
> > supports 88 colors.  That number is enough for Emacs to express all of
> > the possible colors from the X repertoire, so we use that as a
> > threshold for "any color is supported".
> >
> Would you know what colour bit would the 88 colours correspond to?

Sorry, I don't understand: what do you mean by "colour bit" here?

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 27, 2022, 21:56 by eliz@gnu.org:

>> Date: Thu, 27 Jan 2022 09:43:53 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> > It is not arbitrary: xterm (and its workalikes) has a mode where it
>> > supports 88 colors.  That number is enough for Emacs to express all of
>> > the possible colors from the X repertoire, so we use that as a
>> > threshold for "any color is supported".
>> >
>> Would you know what colour bit would the 88 colours correspond to?
>>
>
> Sorry, I don't understand: what do you mean by "colour bit" here?
>
 
I mean the pixel depth

4-bit, 16 colours
 8-bit, 256 colours
16-bit, 65,000 colours

Re: min-colors 88

[Bastian Beranek]

On Thu, Jan 27, 2022 at 11:46 AM fatiparty--- via Users list for the
GNU Emacs text editor <help-gnu-emacs@gnu.org> wrote:
>
> Jan 27, 2022, 21:56 by eliz@gnu.org:
>
> >> Date: Thu, 27 Jan 2022 09:43:53 +0100 (CET)
> >> From: fatiparty@tutanota.com
> >> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> >>
> >> > It is not arbitrary: xterm (and its workalikes) has a mode where it
> >> > supports 88 colors.  That number is enough for Emacs to express all of
> >> > the possible colors from the X repertoire, so we use that as a
> >> > threshold for "any color is supported".
> >> >
> >> Would you know what colour bit would the 88 colours correspond to?
> >>
> >
> > Sorry, I don't understand: what do you mean by "colour bit" here?
> >
>
> I mean the pixel depth
>
> 4-bit, 16 colours
>  8-bit, 256 colours
> 16-bit, 65,000 colours

6.459-bit, 88 colours, clearly :-)

>
>

Re: min-colors 88

[Eli Zaretskii]

> Date: Thu, 27 Jan 2022 11:44:14 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> >> Would you know what colour bit would the 88 colours correspond to?
> >>
> >
> > Sorry, I don't understand: what do you mean by "colour bit" here?
> >
>  
> I mean the pixel depth
> 
> 4-bit, 16 colours
>  8-bit, 256 colours
> 16-bit, 65,000 colours

There's no meaningful number of pixels that correspond to 88 colors,
AFAIK.  The number of colors supported by text-mode terminals doesn't
necessarily depend on the pixel depth of the color display.

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 27, 2022, 23:01 by eliz@gnu.org:

>> Date: Thu, 27 Jan 2022 11:44:14 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> >> Would you know what colour bit would the 88 colours correspond to?
>> >>
>> >
>> > Sorry, I don't understand: what do you mean by "colour bit" here?
>> >
>>  
>> I mean the pixel depth
>>
>> 4-bit, 16 colours
>>  8-bit, 256 colours
>> 16-bit, 65,000 colours
>>
>
> There's no meaningful number of pixels that correspond to 88 colors,
> AFAIK.  The number of colors supported by text-mode terminals doesn't
> necessarily depend on the pixel depth of the color display.
>
So one uses whatever bit format (e.g. 24-Bit Format for #RRGGBB) needed
 to store the maximum colours supported?


 

Re: min-colors 88

[Eli Zaretskii]

> Date: Thu, 27 Jan 2022 12:43:04 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> Jan 27, 2022, 23:01 by eliz@gnu.org:
> 
> >> 4-bit, 16 colours
> >>  8-bit, 256 colours
> >> 16-bit, 65,000 colours
> >>
> >
> > There's no meaningful number of pixels that correspond to 88 colors,
> > AFAIK.  The number of colors supported by text-mode terminals doesn't
> > necessarily depend on the pixel depth of the color display.
> >
> So one uses whatever bit format (e.g. 24-Bit Format for #RRGGBB) needed
>  to store the maximum colours supported?

Yes.  And you can use #RRGGBB (or any variant of it) regardless of the
pixel depth, Emacs will convert as needed.

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

Eli Zaretskii wrote:

>>>>  4-bit, 16 colours
>>>>  8-bit, 256 colours
>>>> 16-bit, 65 000 colours
>>>
>>> There's no meaningful number of pixels that correspond to
>>> 88 colors, AFAIK. The number of colors supported by
>>> text-mode terminals doesn't necessarily depend on the
>>> pixel depth of the color display.
>>
>> So one uses whatever bit format (e.g. 24-Bit Format for
>> #RRGGBB) needed to store the maximum colours supported?
>
> Yes. And you can use #RRGGBB (or any variant of it)
> regardless of the pixel depth, Emacs will convert as needed.

Compare ...

Got to the ttys/Linux VTs/the console ...

$ tput colors
8

I actually there are 16 colors in the VTs, 8 is just what
is reported so it tells you that.

See "test-colors", line 56:

  https://dataswamp.org/~incal/conf/.zsh/colors

screenshot:

  https://dataswamp.org/~incal/figures/shell/test-colors.png

It is 8 colors _and_ 8 supposedly bright versions of the
corresponding colors, but they can be defined to whatever, and
it happens the same way, so there is in practice 16 colors.
(Only when the background comes into play it differs, since
the 8 bright version cannot, I think, be used as backgrounds.)

I have a very old tutorial somewhere on this ...

PS. BTW Double-O Seven, in the computer world it is spelled
    "color".

-- 
underground experts united
https://dataswamp.org/~incal

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

fatiparty--- via Users list for the GNU Emacs text editor wrote:

> So one uses whatever bit format (e.g. 24-Bit Format for
> #RRGGBB) needed to store the maximum colours supported?

Total number of colors expressible by the hexadecimal #RRGGBB
notation:

  $ echo $(( ((16 - 1)*16**1 + (16 - 1)*16**0 + 1)**3 ))
  16 777 216

So the max value is:

  $ dec2bin $(( 16 777 216 - 1 )) # [1]
  11111111 11111111 11111111

which is 3*8 = 24 bits.

(Spaces added for clarity.)

[1] https://dataswamp.org/~incal/conf/.zsh/math

-- 
underground experts united
https://dataswamp.org/~incal

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 27, 2022, 23:01 by eliz@gnu.org:

>> Date: Thu, 27 Jan 2022 11:44:14 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> >> Would you know what colour bit would the 88 colours correspond to?
>> >>
>> >
>> > Sorry, I don't understand: what do you mean by "colour bit" here?
>> >
>>  
>> I mean the pixel depth
>>
>> 4-bit, 16 colours
>>  8-bit, 256 colours
>> 16-bit, 65,000 colours
>>
>
> There's no meaningful number of pixels that correspond to 88 colors,
> AFAIK.  The number of colors supported by text-mode terminals doesn't
> necessarily depend on the pixel depth of the color display.
>
For the case of RGB, would that mean 88 values for each base colour (R,G,B)? 

Re: min-colors 88

[Eli Zaretskii]

> Date: Thu, 27 Jan 2022 12:47:30 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> For the case of RGB, would that mean 88 values for each base colour (R,G,B)? 

No, it means 88 different colors overall.

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

Eli Zaretskii wrote:

>> For the case of RGB, would that mean 88 values for each
>> base colour (R,G,B)?
>
> No, it means 88 different colors overall.

Unless they are all set to the same color ...

-- 
underground experts united
https://dataswamp.org/~incal

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 27, 2022, 23:50 by eliz@gnu.org:

>> Date: Thu, 27 Jan 2022 12:47:30 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> For the case of RGB, would that mean 88 values for each base colour (R,G,B)?
>>
>
> No, it means 88 different colors overall.
>
Would you know the bit length using for each of the base colours when 88 different colours
started to be used? Perhaps three 5-Bit numbers? 

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

fatiparty--- via Users list for the GNU Emacs text editor wrote:

>> No, it means 88 different colors overall.
>
> Would you know the bit length using for each of the base
> colours when 88 different colours started to be used?
> Perhaps three 5-Bit numbers?

007, see:

  There has also been a similar but incompatible 88-color
  encoding using the same escape sequence, seen in rxvt and
  xterm-88color. Not much is known about the scheme besides
  the color codes. It uses a 4×4×4 color cube. [1]

AKA the "Big Eric" mode.

[1] https://en.wikipedia.org/wiki/ANSI_escape_code

-- 
underground experts united
https://dataswamp.org/~incal

Re: min-colors 88

[Eli Zaretskii]

> Date: Thu, 27 Jan 2022 14:03:54 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> > No, it means 88 different colors overall.
> >
> Would you know the bit length using for each of the base colours when 88 different colours
> started to be used? Perhaps three 5-Bit numbers? 

You are thinking about this in the wrong direction, AFAIU.  On
text-mode terminals, colors are just indices into a table, and the
table itself is not exposed.  How the indices are mapped to the actual
colors is entirely up to the terminal emulator.  So it is not useful
to think about these colors in terms of bots per pixel.

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 28, 2022, 03:14 by eliz@gnu.org:

>> Date: Thu, 27 Jan 2022 14:03:54 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> > No, it means 88 different colors overall.
>> >
>> Would you know the bit length using for each of the base colours when 88 different colours
>> started to be used? Perhaps three 5-Bit numbers?
>>
>
> You are thinking about this in the wrong direction, AFAIU.  On
> text-mode terminals, colors are just indices into a table, and the
> table itself is not exposed.  How the indices are mapped to the actual
> colors is entirely up to the terminal emulator.  So it is not useful
> to think about these colors in terms of bots per pixel.
>
I could be mistaking the interpretation of the hex-codes to the number of colours
available.

 

Re: min-colors 88

[Eli Zaretskii]

> Date: Thu, 27 Jan 2022 19:09:34 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> > You are thinking about this in the wrong direction, AFAIU.  On
> > text-mode terminals, colors are just indices into a table, and the
> > table itself is not exposed.  How the indices are mapped to the actual
> > colors is entirely up to the terminal emulator.  So it is not useful
> > to think about these colors in terms of bots per pixel.
> >
> I could be mistaking the interpretation of the hex-codes to the number of colours
> available.

Which hex-codes?

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

Eli Zaretskii wrote:

> Which hex-codes?

#RRGGBB

-- 
underground experts united
https://dataswamp.org/~incal

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 28, 2022, 06:30 by eliz@gnu.org:

>> Date: Thu, 27 Jan 2022 19:09:34 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> > You are thinking about this in the wrong direction, AFAIU.  On
>> > text-mode terminals, colors are just indices into a table, and the
>> > table itself is not exposed.  How the indices are mapped to the actual
>> > colors is entirely up to the terminal emulator.  So it is not useful
>> > to think about these colors in terms of bots per pixel.
>> >
>> I could be mistaking the interpretation of the hex-codes to the number of colours
>> available.
>>
>
> Which hex-codes?
>
For instance #RRGGBB, #RRRRGGGGBBBB.  As I understand there is no relation
between thu possibilities provided by the hex-codes and the lookup table of actual
colours that can be rendered.

Re: min-colors 88

[Eli Zaretskii]

> Date: Thu, 27 Jan 2022 19:46:09 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> For instance #RRGGBB, #RRRRGGGGBBBB.  As I understand there is no relation
> between thu possibilities provided by the hex-codes and the lookup table of actual
> colours that can be rendered.

The relation is established at startup, you can see the code in
xterm.el.

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

>>> You are thinking about this in the wrong direction, AFAIU.
>>> On text-mode terminals, colors are just indices into
>>> a table, and the table itself is not exposed. How the
>>> indices are mapped to the actual colors is entirely up to
>>> the terminal emulator. So it is not useful to think about
>>> these colors in terms of bots per pixel.
>>
>> I could be mistaking the interpretation of the hex-codes to
>> the number of colours available.

There are

  $ echo $(( (16**2)**3 ))
  16 777 216

to choose the 16 (or 88 in your case perhaps) from.

-- 
underground experts united
https://dataswamp.org/~incal

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 28, 2022, 06:47 by help-gnu-emacs@gnu.org:

>>>> You are thinking about this in the wrong direction, AFAIU.
>>>> On text-mode terminals, colors are just indices into
>>>> a table, and the table itself is not exposed. How the
>>>> indices are mapped to the actual colors is entirely up to
>>>> the terminal emulator. So it is not useful to think about
>>>> these colors in terms of bots per pixel.
>>>>
>>>
>>> I could be mistaking the interpretation of the hex-codes to
>>> the number of colours available.
>>>
>
> There are
>
>  $ echo $(( (16**2)**3 ))
>  16 777 216
>
> to choose the 16 (or 88 in your case perhaps) from.
>

I could then come up with a hex-code where the colour in not  supported.


> -- 
> underground experts united
> https://dataswamp.org/~incal
>

Re: min-colors 88

[Eli Zaretskii]

> Date: Fri, 28 Jan 2022 08:58:13 +0100 (CET)
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> From: fatiparty--- via Users list for the GNU Emacs text editor <help-gnu-emacs@gnu.org>
> 
> I could then come up with a hex-code where the colour in not  supported.

There's no such thing when colors on TTY frames are considered:
Emacs automatically and transparently finds the closest color that is
supported, and uses that.

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 28, 2022, 20:07 by eliz@gnu.org:

>> Date: Fri, 28 Jan 2022 08:58:13 +0100 (CET)
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>> From: fatiparty--- via Users list for the GNU Emacs text editor <help-gnu-emacs@gnu.org>
>>
>> I could then come up with a hex-code where the colour in not  supported.
>>
>
> There's no such thing when colors on TTY frames are considered:
> Emacs automatically and transparently finds the closest color that is
> supported, and uses that.
>
For 8-Bit  (2^8=256) there are defined 16 (Basic) + 6×6×6 (Cube) + 24 (Grey)

What's the situation for 24-bit RGB Codes?

Re: min-colors 88

[Eli Zaretskii]

> Date: Fri, 28 Jan 2022 09:59:34 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> Jan 28, 2022, 20:07 by eliz@gnu.org:
> 
> >> I could then come up with a hex-code where the colour in not  supported.
> >>
> >
> > There's no such thing when colors on TTY frames are considered:
> > Emacs automatically and transparently finds the closest color that is
> > supported, and uses that.
> >
> For 8-Bit  (2^8=256) there are defined 16 (Basic) + 6×6×6 (Cube) + 24 (Grey)
> 
> What's the situation for 24-bit RGB Codes?

What do you mean by "defined" in this context?

The colors "defined" for TTY frames are set up in tty-colors.el, and
they don't depend on the number of colors that the terminal actually
supports.

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 28, 2022, 23:36 by eliz@gnu.org:

>> Date: Fri, 28 Jan 2022 09:59:34 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> Jan 28, 2022, 20:07 by eliz@gnu.org:
>>
>> >> I could then come up with a hex-code where the colour in not  supported.
>> >>
>> >
>> > There's no such thing when colors on TTY frames are considered:
>> > Emacs automatically and transparently finds the closest color that is
>> > supported, and uses that.
>> >
>> For 8-Bit  (2^8=256) there are defined 16 (Basic) + 6×6×6 (Cube) + 24 (Grey)
>>
>> What's the situation for 24-bit RGB Codes?
>>
>
> What do you mean by "defined" in this context?
>

For instance I can invent a code #d166ff
 ( ((class color) (min-colors 88) (background dark))
       :foreground "#d166ff" )

That would not mean that #d166ff actually exists for 256 colour availability.


> The colors "defined" for TTY frames are set up in tty-colors.el, and
> they don't depend on the number of colors that the terminal actually
> supports.
>

Re: min-colors 88

[Eli Zaretskii]

> Date: Fri, 28 Jan 2022 12:54:49 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> For instance I can invent a code #d166ff
>  ( ((class color) (min-colors 88) (background dark))
>        :foreground "#d166ff" )
> 
> That would not mean that #d166ff actually exists for 256 colour availability.

It doesn't matter.  Emacs will find a color that is the closest to
that RGB value, and use it.  See tty-colors.el.

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 29, 2022, 00:16 by eliz@gnu.org:

>> Date: Fri, 28 Jan 2022 12:54:49 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> For instance I can invent a code #d166ff
>>  ( ((class color) (min-colors 88) (background dark))
>>        :foreground "#d166ff" )
>>
>> That would not mean that #d166ff actually exists for 256 colour availability.
>>
>
> It doesn't matter.  Emacs will find a color that is the closest to
> that RGB value, and use it.  See tty-colors.el.
>

Are there more colours available than those in the color-name-rgb-alist ?

Re: min-colors 88

[Eli Zaretskii]

> Date: Fri, 28 Jan 2022 13:55:39 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> >> That would not mean that #d166ff actually exists for 256 colour availability.
> >>
> >
> > It doesn't matter.  Emacs will find a color that is the closest to
> > that RGB value, and use it.  See tty-colors.el.
> >
> 
> Are there more colours available than those in the color-name-rgb-alist ?

On some displays, perhaps.  But why is that important?

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 29, 2022, 00:55 by help-gnu-emacs@gnu.org:

> Jan 29, 2022, 00:16 by eliz@gnu.org:
>
>>> Date: Fri, 28 Jan 2022 12:54:49 +0100 (CET)
>>> From: fatiparty@tutanota.com
>>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>>
>>> For instance I can invent a code #d166ff
>>>  ( ((class color) (min-colors 88) (background dark))
>>>        :foreground "#d166ff" )
>>>
>>> That would not mean that #d166ff actually exists for 256 colour availability.
>>>
>>
>> It doesn't matter.  Emacs will find a color that is the closest to
>> that RGB value, and use it.  See tty-colors.el.
>>
>
> Are there more colours available than those in the color-name-rgb-alist ?
>

I have called xdpyinfo, getting (1, 4, 8, 15, 16, 24, 32)
 
xdpyinfo | awk '$1=="depth" && sub(/,$/, "", $2) {print $2}'

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

fatiparty--- via Users list for the GNU Emacs text editor wrote:

> I have called xdpyinfo, getting (1, 4, 8, 15, 16, 24, 32)
> 
> xdpyinfo | awk '$1=="depth" && sub(/,$/, "", $2) {print $2}'

I have something similar,

  $ xdpyinfo -d ":0" | awk '/of root/{print $5}'
  24

I'll add yours to it :)

https://dataswamp.org/~incal/conf/.zsh/x

-- 
underground experts united
https://dataswamp.org/~incal

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 28, 2022, 23:36 by eliz@gnu.org:

>> Date: Fri, 28 Jan 2022 09:59:34 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> Jan 28, 2022, 20:07 by eliz@gnu.org:
>>
>> >> I could then come up with a hex-code where the colour in not  supported.
>> >>
>> >
>> > There's no such thing when colors on TTY frames are considered:
>> > Emacs automatically and transparently finds the closest color that is
>> > supported, and uses that.
>> >
>> For 8-Bit  (2^8=256) there are defined 16 (Basic) + 6×6×6 (Cube) + 24 (Grey)
>>
>> What's the situation for 24-bit RGB Codes?
>>
>
> What do you mean by "defined" in this context?
>
> The colors "defined" for TTY frames are set up in tty-colors.el, and
> they don't depend on the number of colors that the terminal actually
> supports.
>
Looking at tty-colors.el, a 16-bit values is being used for each of the three colours 
(R[0-65535] , G[0-65535] , B[0-65535]).  Giving more possibilities than 24-bit total values
(with 8-bit values for each of the three colours.

tty-color-define (w32console.el) passes each supported color, its index, and its RGB values.

In tty-color-define, each one of the RGB components is a number between 0 and 65535 (16bit).

But unsure how many get passed to tty-color-define.  There is nothing specified in the .c files
about this.



 

Re: min-colors 88

[Eli Zaretskii]

> Date: Fri, 28 Jan 2022 13:53:10 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> > The colors "defined" for TTY frames are set up in tty-colors.el, and
> > they don't depend on the number of colors that the terminal actually
> > supports.
> >
> Looking at tty-colors.el, a 16-bit values is being used for each of the three colours 
> (R[0-65535] , G[0-65535] , B[0-65535]).  Giving more possibilities than 24-bit total values
> (with 8-bit values for each of the three colours.
> 
> tty-color-define (w32console.el) passes each supported color, its index, and its RGB values.
> 
> In tty-color-define, each one of the RGB components is a number between 0 and 65535 (16bit).
> 
> But unsure how many get passed to tty-color-define.  There is nothing specified in the .c files
> about this.

What exactly is the question you want to ask?  Or what is the problem
you are trying to solve?  And how are the above observations related
to those questions/problems?

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 29, 2022, 01:17 by eliz@gnu.org:

>> Date: Fri, 28 Jan 2022 13:53:10 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> > The colors "defined" for TTY frames are set up in tty-colors.el, and
>> > they don't depend on the number of colors that the terminal actually
>> > supports.
>> >
>> Looking at tty-colors.el, a 16-bit values is being used for each of the three colours 
>> (R[0-65535] , G[0-65535] , B[0-65535]).  Giving more possibilities than 24-bit total values
>> (with 8-bit values for each of the three colours.
>>
>> tty-color-define (w32console.el) passes each supported color, its index, and its RGB values.
>>
>> In tty-color-define, each one of the RGB components is a number between 0 and 65535 (16bit).
>>
>> But unsure how many get passed to tty-color-define.  There is nothing specified in the .c files
>> about this.
>>
>
> What exactly is the question you want to ask?  Or what is the problem
> you are trying to solve?  And how are the above observations related
> to those questions/problems?
>
You have clarified a lot of things.  I am trying to figure out the actual list of properly defined
code values for colours, and how many exist if one counts them.  Is there some command 
that actually lists them?

Does emacs only know of the named colour list, or can it use other ones which are unnamed?

Re: min-colors 88

[Eli Zaretskii]

> Date: Fri, 28 Jan 2022 14:49:11 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> > What exactly is the question you want to ask?  Or what is the problem
> > you are trying to solve?  And how are the above observations related
> > to those questions/problems?
> >
> You have clarified a lot of things.  I am trying to figure out the actual list of properly defined
> code values for colours, and how many exist if one counts them.  Is there some command 
> that actually lists them?

Which set of colors do you want to count?  Does "M-x list-colors-display"
give you what you want?  If not, why not?

> Does emacs only know of the named colour list, or can it use other ones which are unnamed?

If you mean specifying colors by their RGB values, that is exactly
what tty-colors.el does.  But I'm not sure I understood the question.

min-colors 88

[fatiparty--- via Users list for the GNU Emacs text editor]

Jan 29, 2022, 02:34 by eliz@gnu.org:

>> Date: Fri, 28 Jan 2022 14:49:11 +0100 (CET)
>> From: fatiparty@tutanota.com
>> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
>>
>> > What exactly is the question you want to ask?  Or what is the problem
>> > you are trying to solve?  And how are the above observations related
>> > to those questions/problems?
>> >
>> You have clarified a lot of things.  I am trying to figure out the actual list of properly defined
>> code values for colours, and how many exist if one counts them.  Is there some command 
>> that actually lists them?
>>
>
> Which set of colors do you want to count?  Does "M-x list-colors-display"
> give you what you want?  If not, why not?
>
You mentioned that Emacs will find a color that is the closest to the RGB value supplied.
I am referring to the list of colours that Emacs uses to color-match the arbitrary RGB user-value supplied.Supposedly there is a list of colours that emacs uses for matching against.  Is the full list defined by
color-name-rgb-alist.  Or are they supplied by hardware calls, or additional ones with no specific name? 


>> Does emacs only know of the named colour list, or can it use other ones which are unnamed?
>>
>
> If you mean specifying colors by their RGB values, that is exactly
> what tty-colors.el does.  But I'm not sure I understood the question.
>

Re: min-colors 88

[Eli Zaretskii]

> Date: Fri, 28 Jan 2022 16:12:23 +0100 (CET)
> From: fatiparty@tutanota.com
> Cc: Help Gnu Emacs <help-gnu-emacs@gnu.org>
> 
> You mentioned that Emacs will find a color that is the closest to the RGB value supplied.
> I am referring to the list of colours that Emacs uses to color-match the arbitrary RGB user-value supplied.Supposedly there is a list of colours that emacs uses for matching against.  Is the full list defined by
> color-name-rgb-alist.

Yes, color-name-rgb-alist is it.

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

fatiparty--- via Users list for the GNU Emacs text editor wrote:

>>> I could be mistaking the interpretation of the hex-codes to
>>> the number of colours available.
>>
>> There are
>>
>>  $ echo $(( (16**2)**3 ))
>>  16 777 216
>>
>> to choose the 16 (or 88 in your case perhaps) from.
>
> I could then come up with a hex-code where the colour in
> not supported.

Usually when you use them you refer to the 16 (or 88) assigned
colors, not hex #RRGGBB numbers. But if you do, either it
won't work or it'll have to be rounded to a color that _is_
assigned and available.

-- 
underground experts united
https://dataswamp.org/~incal

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

fatiparty--- via Users list for the GNU Emacs text editor wrote:

> I could be mistaking the interpretation of the hex-codes to
> the number of colours available.

Notation: #RRGGBB

Each channel: 1 byte = 8 bits

e.g., say "winter blue" is #385d7d, that is binary

  111000 1011101 1111101

As you see, there are

  (16^2)^3 = 16 777 216

colors available!

(or computed from the 24-bit binary,

  2^24 = the same)

However, in _another_ sense there are (or can be) just 8
colors available, and 8 bright versions of them! Here are, for
example, mine! I have them in a tailored file so they are easy
to see and edit, they are then processed to the less-palatable
original Linux format, and the colors appear in the
ttys/Linux VTs/the console.

So the file:

   normal                         bright
   bk  r   g   y  bl   m   c   w  bk    r   g   y  bl   m   c   w
r  0 255   0 190 100 180   0 150  115 255   0 255 125 250  90 210
g  0  50 155 190 100 105 180 150  115  80 185 127 125  90 255 180
b  0  50   0   0 255   0 180 150  115  80   0   0 255 250 255 140

See, there are 8 colors! ........ OK, 16 then!

But you also see that each channel is 0-255. So then we are
back at

  256^3 = 16 777 216

colors available!

-- 
underground experts united
https://dataswamp.org/~incal

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

Eli Zaretskii wrote:

> There's no meaningful number of pixels that correspond to 88
> colors, AFAIK. The number of colors supported by text-mode
> terminals doesn't necessarily depend on the pixel depth of
> the color display.

Here is a reference to 88:

  https://manualzz.com/doc/2019435/amx-nxd-cv7-car-video-system-user-manual

It says:

  Colors can be used to set the colors on buttons, sliders, and
  pages. The lowest color number represents the lightest
  color-specific display; the highest number represents the
  darkest display. For example, 0 represents light red, and 5 is
  dark red.
  RGB triplets and names for basic 88 colors
  RGB Values for all 88 Basic Colors
  Index No.
  Name
  Red
  Green
  Blue
  00
  Very Light Red
  255
  0
  0
  01
  Light Red
  223
  0
  0
  02
  Red
  191
  0
  0
  03
  Medium Red
  159
  0
  0
  04
  Dark Red
  127
  0
  0
  05
  Very Dark Red
  95
  0
  0
  06
  Very Light Orange
  255
  128
  0
  07
  Light Orange
  223
  112
  0
  08
  Orange
  191
  96
  0
  09
  Medium Orange
  159
  80
  0
  10
  Dark Orange
  127
  64
  0
  11
  Very Dark Orange
  95
  48
  0
  12
  Very Light Yellow
  255
  255
  0
  13
  Light Yellow
  223
  223
  0
  14
  Yellow
  191
  191
  0
  15
  Medium Yellow
  159
  159
  0
  16
  Dark Yellow
  127
  127
  0
  17
  Very Dark Yellow
  95
  95
  0
  18
  Very Light Lime
  128
  255
  0
  19
  Light Lime
  112
  223
  0
  20
  Lime
  96
  191
  0
  21
  Medium Lime
  80
  159
  0
  22
  Dark Lime
  64
  127
  0
  23
  Very Dark Lime
  48
  95
  0
  24
  Very Light Green
  0
  255
  0
  25
  Light Green
  0
  223
  0
  26
  Green
  0
  191
  0
  27
  Medium Green
  0
  159
  0
  28
  Dark Green
  0
  127
  0
  29
  Very Dark Green
  0
  95
  0
  30
  Very Light Mint
  0
  255
  128
  31
  Light Mint
  0
  223
  112
  32
  Mint
  0
  191
  96
  33
  Medium Mint
  0
  159
  80
  34
  Dark Mint
  0
  127
  64
  35
  Very Dark Mint
  0
  95
  48
  7" Modero Widescreen Touch Panels
  119
  Programming
  RGB Values for all 88 Basic Colors (Cont.)
  120
  Index No.
  Name
  Red
  Green
  Blue
  36
  Very Light Cyan
  0
  255
  255
  37
  Light Cyan
  0
  223
  223
  38
  Cyan
  0
  191
  191
  39
  Medium Cyan
  0
  159
  159
  40
  Dark Cyan
  0
  127
  127
  41
  Very Dark Cyan
  0
  95
  95
  42
  Very Light Aqua
  0
  128
  255
  43
  Light Aqua
  0
  112
  223
  44
  Aqua
  0
  96
  191
  45
  Medium Aqua
  0
  80
  159
  46
  Dark Aqua
  0
  64
  127
  47
  Very Dark Aqua
  0
  48
  95
  48
  Very Light Blue
  0
  0
  255
  49
  Light Blue
  0
  0
  223
  50
  Blue
  0
  0
  191
  51
  Medium Blue
  0
  0
  159
  52
  Dark Blue
  0
  0
  127
  53
  Very Dark Blue
  0
  0
  95
  54
  Very Light Purple
  128
  0
  255
  55
  Light Purple
  112
  0
  223
  56
  Purple
  96
  0
  191
  57
  Medium Purple
  80
  0
  159
  58
  Dark Purple
  64
  0
  127
  59
  Very Dark Purple
  48
  0
  95
  60
  Very Light Magenta
  255
  0
  255
  61
  Light Magenta
  223
  0
  223
  62
  Magenta
  191
  0
  191
  63
  Medium Magenta
  159
  0
  159
  64
  Dark Magenta
  127
  0
  127
  65
  Very Dark Magenta
  95
  0
  95
  66
  Very Light Pink
  255
  0
  128
  67
  Light Pink
  223
  0
  112
  68
  Pink
  191
  0
  96
  69
  Medium Pink
  159
  0
  80
  70
  Dark Pink
  127
  0
  64
  71
  Very Dark Pink
  95
  0
  48
  72
  White
  255
  255
  255
  73
  Grey1
  238
  238
  238
  74
  Grey3
  204
  204
  204
  75
  Grey5
  170
  170
  170
  76
  Grey7
  136
  136
  136
  77
  Grey9
  102
  102
  102
  78
  Grey4
  187
  187
  187
  79
  Grey6
  153
  153
  153
  7" Modero Widescreen Touch Panels
  Programming
  RGB Values for all 88 Basic Colors (Cont.)
  Index No.
  Name
  Red
  Green
  Blue
  80
  Grey8
  119
  119
  119
  81
  Grey10
  85
  85
  85
  82
  Grey12
  51
  51
  51
  83
  Grey13
  34
  34
  34
  84
  Grey2
  221
  221
  221
  85
  Grey11
  68
  68
  68
  86
  Grey14
  17
  17
  17
  87
  Black
  0
  0
  0
  255
  TRANSPARENT
  99
  53
  99

-- 
underground experts united
https://dataswamp.org/~incal

Re: min-colors 88

[Emanuel Berg via Users list for the GNU Emacs text editor]

Eli Zaretskii wrote:

>> I keep seeing 88 for min-colors. I wonder if a different
>> number would be more appropriate, as it seems that 88 is
>> quite arbitrary.
>
> It is not arbitrary: xterm (and its workalikes) has a mode
> where it supports 88 colors. That number is enough for Emacs
> to express all of the possible colors from the X repertoire,
> so we use that as a threshold for "any color is supported".

88 was useful to people who went to South America after WW2 as
well ...

-- 
underground experts united
https://dataswamp.org/~incal