Using w3m for HTML->TXT doc conversion

h/t Shin-ichi for the tip

docs/CHANGES.txt

@@ -6,7 +6,7 @@ Copyright (c) 2002-2023
 Various contributors (see below, and AUTHORS.txt)
 https://tuxpaint.org/
 
-2023.July.13 (0.9.31)
+2023.July.15 (0.9.31)
   * New Magic Tools:
     ----------------
     * Loops - Draw loop-the-loops.
@@ -105,6 +105,12 @@ https://tuxpaint.org/
       and keyboard accessibility mode).
       Bill Kendrick <bill@newbreedsoftware.com>
 
+    * Using `w3m` (rather than `links`) to generated plain text
+      alternatives to HTML documentation.  (It handles word-wrapping
+      Japanese text properly.)
+      h/t TOYAMA Shin-ichi
+      Bill Kendrick <bill@newbreedsoftware.com>
+
   * Bug Fixes:
     ----------
     * In some window size / button size combinations, Eraser

docs/Makefile

@@ -1,21 +1,23 @@
 # Makefile for Tux Paint docs
-# 
+#
 # Uses "links" to convert docs from HTML to plain text.
 # (Normally only ran by the developers after updating the HTML, prior to
 # release.)
 #
 # Bill Kendrick
 # bill@newbreedsoftware.com
-# 
-# Sept. 4, 2005 - June 29, 2023
+#
+# Sept. 4, 2005 - July 16, 2023
 
 # FIXME: Japanese does not wordwrap in many cases, leading to very long
 # lines in the TXT output.  Post-processing with `fmt` doesn't look like
 # it would help, because it doesn't know how to wrap Japanese, either.
 # -bjk 2023.05.02
+# HTML2TXT_OPTIONS:=-dump -codepage utf8 -width 80
+# HTML2TXT:=links $(HTML2TXT_OPTIONS)
 
-LINKS_OPTIONS:=-dump -codepage utf8 -width 80
-LINKS:=links $(LINKS_OPTIONS)
+HTML2TXT_OPTIONS:=-dump -cols 80 -no-graph -o pseudo_inlines=f -o display_image=f -T text/html
+HTML2TXT:=./w3m.sh $(HTML2TXT_OPTIONS)
 
 EN_HTMLFILES:=$(wildcard en/html/*.html)
 EN_TEXTFILES:=$(patsubst en/html/%.html,en/%.txt,$(EN_HTMLFILES))
@@ -80,35 +82,34 @@ clean:
 	$(ZH_TW_TEXTFILES)
 
 $(EN_TEXTFILES): en/%.txt: en/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(ES_TEXTFILES): es_ES.UTF-8/%.txt: es_ES.UTF-8/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(FR_TEXTFILES): fr_FR.UTF-8/%.txt: fr_FR.UTF-8/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(GL_TEXTFILES): gl_ES.UTF-8/%.txt: gl_ES.UTF-8/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(IS_TEXTFILES): is_IS.UTF-8/%.txt: is_IS.UTF-8/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(IT_TEXTFILES): it/%.txt: it/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(JA_TEXTFILES): ja_JP.UTF-8/%.txt: ja_JP.UTF-8/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(NL_TEXTFILES): nl/%.txt: nl/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(RU_TEXTFILES): ru/%.txt: ru/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(ZH_CN_TEXTFILES): zh_cn/%.txt: zh_cn/html/%.html
-	$(LINKS) $< > $@
+	$(HTML2TXT) $< > $@
 
 $(ZH_TW_TEXTFILES): zh_tw/%.txt: zh_tw/html/%.html
-	$(LINKS) $< > $@
-
+	$(HTML2TXT) $< > $@

docs/cjkwrap-docs.py

@@ -1,27 +0,0 @@
-#!/usr/bin/env python3
-
-# FIXME: Under construction!
-# Experimental script to try and wrap
-# Japanese plaintext docs, since `links` does not
-# currently do this properly.  -bjk 2023.07.15
- 
-import cjkwrap
-
-with open('ja_JP.UTF-8/README.txt') as f:
-  while True:
-    line = f.readline()
-    if not line:
-      break
-
-    oldlen = len(line)
-    line = line.lstrip()
-    newlen = len(line)
-    indent = oldlen - newlen
-
-    wrappedlines = cjkwrap.wrap(line, 78 - indent)
-    for l in wrappedlines:
-      print(" " * indent, end="")
-      print(l)
-
-f.close()
-

docs/w3m.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+infile=${@: -1}
+args=${@:1:${#}-1}
+
+sed -e "s/<\\/dd>/<br\\/><br\\/><\\/dd>/g" $infile | w3m $args
+