Remove (linktext) links, add <link> links

1 files changed, 40 insertions, 34 deletions

diff --git a/emparse b/emparse
index 3184488..5ddb585 100755
--- a/emparse
+++ b/emparse
@@ -2,19 +2,14 @@
 
 # aux/emparse -- parse em source
 
-BEGIN {
-	expectblock = 1
-	itemlevel = 1
-	linktext = ENVIRON["linktext"]
-	if (!linktext) linktext = "link"
-}
+BEGIN { expectblock = 1; itemlevel = 1 }
 END { breakblock() }
 
 /^$/							{ breakblock(); getline }
 expectblock && /^  /			{ newblock("table") }
 expectblock && /^ - /			{ newblock("ul") }
-expectblock && /^ [0-9a-z]+\. /	{ newblock("ol") }
-expectblock && /^ \[[0-9a-z]+\] /	{ newblock("nl") }
+expectblock && /^ [0-9]+\. /		{ newblock("ol") }
+expectblock && /^ \[[0-9]+\] /	{ newblock("nl") }
 expectblock && /^ .*: /			{ newblock("dl") }
 expectblock && /^	/			{ newblock("pre") }
 expectblock && /^---$/			{ expectblock = 0; printf "<hr/>\n"; next }
@@ -28,36 +23,43 @@ expectblock { newblock("p") }
 
 openblock == "pre" { sub("^	", ""); $0 = escape($0); printf "%s\n", $0; next }
 
-openblock == "ul" && /^ - /					{ item(1, "ul", line) }
-openblock == "ol" && /^ [0-9a-z]+\. /			{ item(1, "ol", line) }
-openblock == "nl" && /^ \[[0-9a-z]+\] [^ 	]+$/	{ next }			# hyperlink reference
-openblock == "nl" && /^ \[[0-9a-z]+\] /			{ item(1, "nl", line) }	# text reference
-openblock == "dl" && /^ .*: /					{ term(line) }
-
-(openblock == "ul" || openblock == "ol") && /^  -/				{ item(2, "ul", line) }
-(openblock == "ul" || openblock == "ol") && /^  [0-9a-z]+\./		{ item(2, "ol", line) }
-(openblock == "ul" || openblock == "ol") && /^   -/				{ item(3, "ul", line) }
-(openblock == "ul" || openblock == "ol") && /^   [0-9a-z]+\./		{ item(3, "ol", line) }
-(openblock == "ul" || openblock == "ol") && /^    -/				{ item(4, "ul", line) }
-(openblock == "ul" || openblock == "ol") && /^    [0-9a-z]+\./		{ item(4, "ol", line) }
-(openblock == "ul" || openblock == "ol") && /^     -/				{ item(5, "ul", line) }
-(openblock == "ul" || openblock == "ol") && /^     [0-9a-z]+\./		{ item(5, "ol", line) }
-(openblock == "ul" || openblock == "ol") && /^      -/			{ item(6, "ul", line) }
-(openblock == "ul" || openblock == "ol") && /^      [0-9a-z]+\./	{ item(6, "ol", line) }
+openblock == "ul" && /^ - /				{ item(1, "ul", line) }
+openblock == "ol" && /^ [0-9]+\. /			{ item(1, "ol", line) }
+#openblock == "nl" && /^ \[[0-9]+\] [^ 	]+$/	{ next }			# hyperlink reference
+openblock == "nl" && /^ \[[0-9]+\] /		{ item(1, "nl", line) }	# text reference
+openblock == "dl" && /^ .*: /				{ term(line) }
+
+(openblock == "ul" || openblock == "ol") && /^  -/			{ item(2, "ul", line) }
+(openblock == "ul" || openblock == "ol") && /^  [0-9]+\./		{ item(2, "ol", line) }
+(openblock == "ul" || openblock == "ol") && /^   -/			{ item(3, "ul", line) }
+(openblock == "ul" || openblock == "ol") && /^   [0-9]+\./	{ item(3, "ol", line) }
+(openblock == "ul" || openblock == "ol") && /^    -/			{ item(4, "ul", line) }
+(openblock == "ul" || openblock == "ol") && /^    [0-9]+\./	{ item(4, "ol", line) }
+(openblock == "ul" || openblock == "ol") && /^     -/			{ item(5, "ul", line) }
+(openblock == "ul" || openblock == "ol") && /^     [0-9]+\./	{ item(5, "ol", line) }
+(openblock == "ul" || openblock == "ol") && /^      -/		{ item(6, "ul", line) }
+(openblock == "ul" || openblock == "ol") && /^      [0-9]+\./	{ item(6, "ol", line) }
 
 { format($0) } # inline formatting
 
 function format(line) {
 	line = escape(line)
-	n = split(line, w, "[ 	]")
+	n = split(line, w, "[ 	]") # TODO: this removes spaces...
 	for (i = 0; i <= n; i++) {
 		if (w[i] == "") continue;
 
-		if (match(w[i], "^\\[[0-9a-z]+\\]$"))
+		if (match(w[i], "^&lt;(\\./.*|\\.\\./.*|.*/.*)&gt;$"))
+			w[i] = link(substr(w[i], 5, RLENGTH-8))
+		else if (match(w[i], "^&lt;(\\./.*|\\.\\./.*|.*/.*)&gt;[.,:;?!)]$"))
+			w[i] = link(substr(w[i], 5, RLENGTH-9)) substr(w[i], RLENGTH)
+		else if (match(w[i], "^&lt;(\\./.*|\\.\\./.*|.*/.*)&gt;\\)[.,:;?!]$"))
+			w[i] = link(substr(w[i], 5, RLENGTH-10)) substr(w[i], RLENGTH-1)
+
+		if (match(w[i], "^\\[[0-9]+\\]$"))
 			w[i] = ref(substr(w[i], 2, RLENGTH-2))
-		else if (match(w[i], "^\\[[0-9a-z]+\\][.,:;?!)]$"))
+		else if (match(w[i], "^\\[[0-9]+\\][.,:;?!)]$"))
 			w[i] = ref(substr(w[i], 2, RLENGTH-3)) substr(w[i], RLENGTH)
-		else if (match(w[i], "^\\[[0-9a-z]+\\]\\)[.,:;?!]$"))
+		else if (match(w[i], "^\\[[0-9]+\\]\\)[.,:;?!]$"))
 			w[i] = ref(substr(w[i], 2, RLENGTH-4)) substr(w[i], RLENGTH-1)
 
 		if (!openformat) {
@@ -112,7 +114,7 @@ function format(line) {
 }
 
 function opentag(name) {
-	if (name == "nl") printf "<ol class=\"reflist\" style=\"font-size: small;\">\n"
+	if (name == "nl") printf "<ol class=\"reflist\">\n"
 	else printf "<%s>\n", name
 }
 
@@ -175,15 +177,15 @@ function item(level, type, line) {
 		printf "<li>"
 	}
 	if (type == "ol") {
-		match($0, "[0-9a-z]+\.")
+		match($0, "[0-9]+\.")
 		v = substr($0, RSTART, RLENGTH-1)
-		sub("^ +[0-9a-z]+\. ", "")
+		sub("^ +[0-9]+\. ", "")
 		printf "<li value=\"%s\">", v
 	}
 	if (type == "nl") {
-		match($0, "\\[[0-9a-z]+\\]")
+		match($0, "\\[[0-9]+\\]")
 		v = substr($0, RSTART+1, RLENGTH-2)
-		sub("^ +\\[[0-9a-z]+\\] ", "")
+		sub("^ +\\[[0-9]+\\] ", "")
 		printf "<li value=\"%s\" id=\"ref%s\">", v, v
 	}
 }
@@ -201,10 +203,14 @@ function term(line, t) { # t is a local variable
 
 function ref(v) {
 	if (ENVIRON["ref" v] != "")
-		return "(<a href=\"" ENVIRON["ref" v] "\">" linktext "</a>)"
+		return "[<a href=\"" ENVIRON["ref" v] "\">" v "</a>]"
 	return "[<a href=\"#ref" v "\">" v "</a>]"
 }
 
+function link(h) {
+	return "<a href=\"" h "\">" h "</a>"
+}
+
 function escape(s) {
 	gsub("&", "\\&amp;", s)
 	gsub("<", "\\&lt;", s)