bug#30377: 27.0.50; Missing and misspelled SHR/CSS colors

bug#30377: 27.0.50; Missing and misspelled SHR/CSS colors

[Basil L. Contovounesios]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: 0001-Fix-shr-and-CSS-4-color-maps.patch --]
[-- Type: text/x-diff, Size: 2196 bytes --]

From abb1b62bb2cda93cc902d06dd438a37fdcb1a7cf Mon Sep 17 00:00:00 2001
From: "Basil L. Contovounesios" <contovob@tcd.ie>
Date: Wed, 7 Feb 2018 02:49:08 +0000
Subject: [PATCH] Fix shr and CSS 4 color maps

* lisp/net/shr-color.el (shr-color-html-colors-alist): Fix typos.
Add RebeccaPurple.
* lisp/textmodes/css-mode.el (css--color-map): Add fuchsia-magenta
and aqua-cyan aliases.
---
 lisp/net/shr-color.el      | 5 +++--
 lisp/textmodes/css-mode.el | 2 ++
 2 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/lisp/net/shr-color.el b/lisp/net/shr-color.el
index 31f3d46ed6..6303141c89 100644
--- a/lisp/net/shr-color.el
+++ b/lisp/net/shr-color.el
@@ -137,7 +137,7 @@ shr-color-html-colors-alist
     ("MediumAquaMarine" . "#66CDAA")
     ("MediumBlue" . "#0000CD")
     ("MediumOrchid" . "#BA55D3")
-    ("MediumPurple" . "#9370D8")
+    ("MediumPurple" . "#9370DB")
     ("MediumSeaGreen" . "#3CB371")
     ("MediumSlateBlue" . "#7B68EE")
     ("MediumSpringGreen" . "#00FA9A")
@@ -158,7 +158,7 @@ shr-color-html-colors-alist
     ("PaleGoldenRod" . "#EEE8AA")
     ("PaleGreen" . "#98FB98")
     ("PaleTurquoise" . "#AFEEEE")
-    ("PaleVioletRed" . "#D87093")
+    ("PaleVioletRed" . "#DB7093")
     ("PapayaWhip" . "#FFEFD5")
     ("PeachPuff" . "#FFDAB9")
     ("Peru" . "#CD853F")
@@ -166,6 +166,7 @@ shr-color-html-colors-alist
     ("Plum" . "#DDA0DD")
     ("PowderBlue" . "#B0E0E6")
     ("Purple" . "#800080")
+    ("RebeccaPurple" . "#663399")
     ("Red" . "#FF0000")
     ("RosyBrown" . "#BC8F8F")
     ("RoyalBlue" . "#4169E1")
diff --git a/lisp/textmodes/css-mode.el b/lisp/textmodes/css-mode.el
index 55c21f8acb..727bc18ebb 100644
--- a/lisp/textmodes/css-mode.el
+++ b/lisp/textmodes/css-mode.el
@@ -499,6 +499,7 @@ css--color-map
     ("red" . "#ff0000")
     ("purple" . "#800080")
     ("fuchsia" . "#ff00ff")
+    ("magenta" . "#ff00ff")
     ("green" . "#008000")
     ("lime" . "#00ff00")
     ("olive" . "#808000")
@@ -507,6 +508,7 @@ css--color-map
     ("blue" . "#0000ff")
     ("teal" . "#008080")
     ("aqua" . "#00ffff")
+    ("cyan" . "#00ffff")
     ("orange" . "#ffa500")
     ("aliceblue" . "#f0f8ff")
     ("antiquewhite" . "#faebd7")
-- 
2.15.1


[-- Attachment #2: Type: text/plain, Size: 2050 bytes --]


I attach a patch to fix the following minor issues.

1. The variable shr-color-html-colors-alist from lisp/net/shr-color.el
   is missing the CSS 4 colour RebeccaPurple and contains typos in the
   hex value of colours MediumPurple and PaleVioletRed.

2. The variable css--color-map from lisp/textmodes/css-mode.el is
   missing the magenta and cyan aliases of fuchsia and aqua,
   respectively.

I have taken the liberty of CCing the authors of these features both for
comment on this patch, but more importantly to address Eli's question
from bug#25525 anew:

> I wonder whether it would make sense to reuse some existing data and
> code here?

In that bug report, which introduced css--color-map, Eli was referring
to the list of colours in lisp/term/tty-colors.el.  I would instead like
to draw attention to the overlap between css--color-map and
shr-color-html-colors-alist.

In particular, applying the patch proposed herein satisfies the
following condition:

  (seq-set-equal-p
   css--color-map
   shr-color-html-colors-alist
   (lambda (a b)
     (and (eq t (compare-strings (car a) nil nil (car b) nil nil t))
          (eq t (compare-strings (cdr a) nil nil (cdr b) nil nil t)))))

In other words, css--color-map and shr-color-html-colors-alist become
identical, bar letter case.  I have additionally cross-checked the
updated maps with the latest CSS spec[1] and did not find any other
omissions.

[1] https://www.w3.org/TR/css-color-4/.

So, would it be reasonable for one mode to reuse the map of the other?
A (very) cursory glance through bug#25525 and lisp/net/shr-color.el
suggests there may be some further code duplication pertaining to colour
distances; is this so?

Thanks (and apologies if the noise is unjustified),

-- 
Basil

In GNU Emacs 27.0.50 (build 4, x86_64-pc-linux-gnu, X toolkit, Xaw3d scroll bars)
 of 2018-02-03 built on thunk
Repository revision: 84c9dba4cee052b68b194c3a2e5c297a94d8c8af
Windowing system distributor 'The X.Org Foundation', version 11.0.11906000
System Description: Debian GNU/Linux buster/sid

bug#30377: 27.0.50; Missing and misspelled SHR/CSS colors

[Simen Heggestøyl]

[-- Attachment #1: Type: text/plain, Size: 267 bytes --]

Thanks Basil, good catches. The patch looks good to me. I suppose it's
OK for emacs-26, Eli?

I agree that it looks like the two color lists can easily be unified
later and that there seems to be more opportunities for reuse by
css-mode.el in shr-color.el.

-- Simen

[-- Attachment #2: Type: text/html, Size: 393 bytes --]

bug#30377: 27.0.50; Missing and misspelled SHR/CSS colors

[Tom Tromey]

>>>>> "Basil" == Basil L Contovounesios <contovob@tcd.ie> writes:

Basil> * lisp/textmodes/css-mode.el (css--color-map): Add fuchsia-magenta
Basil> and aqua-cyan aliases.

Thanks.  According to MDN these are the only non-grey/gray aliases.
Maybe they weren't on whatever list I started with for that reason.

Basil> In that bug report, which introduced css--color-map, Eli was referring
Basil> to the list of colours in lisp/term/tty-colors.el.  I would instead like
Basil> to draw attention to the overlap between css--color-map and
Basil> shr-color-html-colors-alist.

Yes, I wasn't aware of this list from shr.  It seems to me that shr
ought to use exactly the CSS colors, so the lists can be shared.
Though, I recommend contacting Lars before doing anything.

Basil> A (very) cursory glance through bug#25525 and lisp/net/shr-color.el
Basil> suggests there may be some further code duplication pertaining to colour
Basil> distances; is this so?

It seems so but I've mostly forgotten whatever I knew about this area.

What I do remember is checking my original algorithm against the one Eli
suggested, by comparing all the CSS named colors with their backgrounds;
and concluding that the one Eli suggested (built in to Emacs) was just
as good.  So perhaps shr is doing too much work here as well.

Tom

bug#30377: 27.0.50; Missing and misspelled SHR/CSS colors

[Eli Zaretskii]

> Date: Wed, 07 Feb 2018 10:45:02 +0100
> From: Simen Heggestøyl <simenheg@gmail.com>
> Cc: 30377@debbugs.gnu.org, Tom Tromey <tom@tromey.com>, Julien Danjou
> 	<julien@danjou.info>, Eli Zaretskii <eliz@gnu.org>
> 
> Thanks Basil, good catches. The patch looks good to me. I suppose it's
> OK for emacs-26, Eli?

Yes, thanks.

bug#30377: 27.0.50; Missing and misspelled SHR/CSS colors

[Simen Heggestøyl]

[-- Attachment #1: Type: text/plain, Size: 190 bytes --]

I've installed the patch, thanks again, Basil.

I'm closing this ticket, but I think the discussion regarding color
distances could continue naturally at the still-open Bug#30295.

-- Simen

[-- Attachment #2: Type: text/html, Size: 296 bytes --]