From 7f2a2075423fa409e5cbc0903260a899bc47fb9a Mon Sep 17 00:00:00 2001
From: Syndamia <kamen@syndamia.com>
Date: Sun, 2 Mar 2025 23:29:25 +0200
Subject: feat(parser)!: Rework

---
 md-parser.awk | 381 +++++++++++++++++++++++++++++++++++++---------------------
 1 file changed, 247 insertions(+), 134 deletions(-)

(limited to 'md-parser.awk')

diff --git a/md-parser.awk b/md-parser.awk
index c063b4e..7f3519b 100644
--- a/md-parser.awk
+++ b/md-parser.awk
@@ -23,40 +23,9 @@
 #
 # No character classes have been used, since they aren't supported on every system
 
-function empty(value) {
-	return value == ""
-}
-
-function PNappnd(value) {
-	if (emptyP1) P2 = P2 ORS value
-	else         P1 = P1 ORS value
-}
-
-function _next() {
-	__buffered_print()
-	not_paragraph = 1
-	next
-}
-
-function sub_match(replacement) {
-	$0 = substr($0, 1, RSTART - 1) replacement substr($0, RSTART + RLENGTH)
-}
-
-function char(string, ind) {
-	if (1 > ind || ind > length(string)) {
-		return "";
-	}
-	return substr(string, ind, 1)
-}
-
-function backup_match() {
-	backup_rstart = RSTART
-	backup_rlength = RLENGTH
-}
-function restore_match() {
-	RSTART = backup_rstart
-	RLENGTH = backup_rlength
-}
+# Naming conventions:
+# all caps - constants, do not assign to variable (directly)
+# begins with "__" - internal, do not use
 
 BEGIN {
 	## This is used to convert email autolinks into a series of hex values
@@ -77,11 +46,11 @@ BEGIN {
 }
 
 {
-	not_paragraph_p1 = not_paragraph
-	not_paragraph = 0
+	STATE_assign(STATE_PARAGRAPH_P1, STATE_PARAGRAPH)
+	STATE_set(STATE_PARAGRAPH)
+
 	empty0  = empty($0)
 	emptyP1 = empty(P1)
-	list_started = 0
 }
 
 # Output is buffered on multiple lines, held in these variables:
@@ -90,96 +59,116 @@ BEGIN {
 # $0 = current line
 # N1 = next new line (N1 is always set by the user)
 
+#== Block markup ==#
+
 # Inline HTML
-/^[ \t]*<\/?[0-9A-Za-z]+/ {
-	match($0, /^[ \t]*/)
-	if (last_html || RLENGTH - 2 * list_level < 4) {
-		last_html = 1
-		_next()
+/^[ \t]?[ \t]?[ \t]?<\/?[0-9A-Za-z]+/ {
+	# if it isn't inside a list
+	if (indentation() - 2 * list_level < 4) {
+		STATE_set(STATE_HTML)
+		soft_next()
 	}
 }
-{
-	last_html = 0
+STATE_isset(STATE_HTML) {
+	if (/^[ \t]*$/) {
+		STATE_unset(STATE_HTML)
+	}
+	else {
+		soft_next()
+	}
 }
 
 # Fenced code blocks
-fenced_codeblock {
-	$0 = "    " $0
-}
-
-# Code blocks (ending)
-inside_codeblock && (/    ```/ || ! /^(    |\t|[ \t]*$)/) {
-	PNappnd("</code></pre>")
-	inside_codeblock = 0
-	emptyP1 = 1
-	not_paragraph = 1
-	if (fenced_codeblock) {
-		fenced_codeblock = 0
+STATE_isset(STATE_FENCED_CODEBLOCK) {
+	if (fenced_class) {
+		$0 = fenced_class $0
 		fenced_class = ""
-		next
+		soft_next()
 	}
+	$0 = "    " $0
 }
 
-# Link definition
-!inside_codeblock && /^\[[^\]]+\]:[ \t]+.*/ {
-	next
-}
-
-# Escapes
-!inside_codeblock && /\\[-\\`*_{}\[\]()#+.!^~]/ {
-	while (match($0, /\\[-\\`*_{}\[\]()#+.!^~]/)) {
-		sub_match(escapes[char($0, RSTART + 1)])
+# Code blocks (ending condition)
+STATE_isset(STATE_CODEBLOCK) {
+	# If ended codeblock
+	if ((STATE_isset(STATE_FENCED_CODEBLOCK) && /    ```$/) ||
+	    ! /^(    |\t|[ \t]*$)/ ||
+	    (blockquote_level > 0 && empty0))
+	{
+		PNappnd("</code></pre>")
+		STATE_unset(STATE_CODEBLOCK)
+		emptyP1 = 1
+		if (STATE_thenunset(STATE_FENCED_CODEBLOCK)) {
+			next
+		}
+	}
+	else if (!STATE_isset(STATE_BLOCKQUOTE)) {
+		handle_codeblock()
+		soft_next()
 	}
-}
-
-# Horizontal rule
-/^(---+|- - -( -)*|\*\*\*+|\* \* \*( \*)*|___+|_ _ _( _)*)$/ {
-	$0 = "<hr/>"
-	not_paragraph = 1
 }
 
 # Blockquotes
 blockquote_level > 0 && empty0 {
-	if (inside_codeblock) {
-		P1 = P1 ORS "</code></pre>"
-		inside_codeblock = 0
-	}
+	STATE_unset(STATE_PARAGRAPH_P1)
 	for (P1 = P1 ORS; blockquote_level > 0; --blockquote_level) {
 		P1 = P1 "</blockquote>"
 	}
-	not_paragraph_p1 = 1
 }
 /^>[ \t]/ {
 	match($0, /^[> \t]*> /)
 	level = substr($0, RSTART, RLENGTH)
+
 	$0 = substr($0, RSTART + RLENGTH)
 	gsub(/[ \t]/, "", level)
 	level = length(level)
 
+	STATE_set(STATE_BLOCKQUOTE)
 	for (;blockquote_level < level; ++blockquote_level) {
 		P1 = P1 "<blockquote>"
 	}
-	if (blockquote_level > level) {
-		if (inside_codeblock) {
-			P2 = P2 ORS "</code></pre>"
-			inside_codeblock = 0
-		}
-		P2 = P2 ORS
-	}
 	for (;blockquote_level > level; --blockquote_level) {
 		P2 = P2 "</blockquote>"
 	}
-	not_paragraph = 1
 }
-/^>$/ { $0 = ""; not_paragraph = 1 }
+STATE_isset(STATE_BLOCKQUOTE) && $0 == ">" { $0 = "" }
+
+# Code block (end)
+STATE_isset(STATE_CODEBLOCK) { ## when in blockquote
+	handle_codeblock()
+	soft_next()
+}
+
+# Link definition
+/^\[[^\]]+\]:[ \t]+.*/ { # line starts link definition
+	next
+}
+
+# Escapes
+/\\[-\\`*_{}\[\]()#+.!^~]/ { # has escaped char
+	while (match($0, /\\[-\\`*_{}\[\]()#+.!^~]/)) {
+		sub_match(escapes[char($0, RSTART + 1)])
+	}
+}
+
+# Horizontal rule
+/^(---+|- - -( -)*|\*\*\*+|\* \* \*( \*)*|___+|_ _ _( _)*)$/ {
+	STATE_unset(STATE_PARAGRAPH)
+	$0 = "<hr/>"
+}
 
 # Lists
 function end_list() {
+	STATE_unset(STATE_LIST)
+
 	if (list_level == 0) return
 	PNappnd("</li></" (lists[list_level] == "." ? "o" : "u") "l>")
 	--list_level
 }
 function start_list(level, type, shift) {
+	STATE_set(STATE_LIST)
+	brand_new = TRUE
+
 	## Every two columns we define a new level
 	level = int((level + 2) / 2)
 	## If list definition is the first line inside a code block
@@ -207,22 +196,20 @@ function start_list(level, type, shift) {
 		PNappnd("</li>")
 		P1 = P1 ORS "<li>"
 	}
-	not_paragraph = 1
-	list_started = 1
 }
-!inside_codeblock && /^[ \t]*[*+-] / { ## Unordered list
-	match($0, /^[ \t]*[*+-] /)
-	level = RLENGTH - 2
-	type  = char($0, RSTART + RLENGTH - 2)
+/^[ \t]*[*+-] / { ## Unordered list
+	level = indentation()
+
 	match($0, /^[^*+-]*[*+-] /)
+	type  = char($0, RSTART + RLENGTH - 2)
 	shift = RLENGTH
 	sub_match("")
 
 	start_list(level, type, shift)
 }
-!inside_codeblock && /^[ \t]*[0-9]+\. / { ## Ordered list
-	match($0, /^[ \t]*/)
-	level = RLENGTH
+/^[ \t]*[0-9]+\. / { ## Ordered list
+	level = indentation()
+
 	type  = "."
 	match($0, /^[^.]+\. /)
 	shift = RLENGTH
@@ -230,39 +217,45 @@ function start_list(level, type, shift) {
 
 	start_list(level, type, shift)
 }
-list_level > 0 && !list_started && /[^ \t]/ { ## Non-list line
-	match($0, /^[ \t]*/)
-	level = int((RLENGTH + 2) / 2)
-	not_paragraph = 1
+!brand_new && list_level > 0 && /[^ \t]/ { ## Non-list line
+	level = int((indentation() + 2) / 2)
 	sub("^[ \t]{1," lists_shift[list_level] "}", "")
 
 	while (list_level >= level && emptyP1) {
 		end_list()
-		not_paragraph = 0
 	}
 }
+list_level > 0 {
+	STATE_unset(STATE_PARAGRAPH)
+	brand_new = FALSE
+}
 
 # Checklists
-list_started && /^\[[x ]\][ \t]/ {
+STATE_isset(STATE_LIST) && /^\[[x ]\][ \t]/ {
 	checked = (char($0, 2) == "x") ? " checked=\"checked\"" : "";
 	$0 = "<input type=\"checkbox\" disabled=\"disabled\"" checked \
 	     "><label>" substr($0, 5) "</label>"
 }
 
 # Fenced code blocks
-!inside_codeblock && /^[ \t]*```/ {
-	fenced_codeblock = 1
+/^[ \t]*```/ {
+	STATE_set(STATE_FENCED_CODEBLOCK)
+	STATE_set(STATE_CODEBLOCK)
 
 	match(P1, /```[^ \t]*/)
 	if (RLENGTH > 3) {
 		fenced_class = " class=\"language-" tolower(substr(P1, RSTART + 3, RLENGTH - 3)) "\""
 	}
+	else {
+		fenced_class = ""
+	}
 
+	fenced_class = "<pre><code" fenced_class ">"
 	next
 }
 
 # Code blocks
-/^(    |\t)/ {
+function handle_codeblock() {
 	## Autoescape & and <
 	for (i = 1; match(substr($0, i), /&/); i += RSTART) {
 		RSTART += i - 1
@@ -271,34 +264,42 @@ list_started && /^\[[x ]\][ \t]/ {
 	gsub(/</, "\\&lt;")
 	gsub(/>/, "\\&gt;")
 
-	if (fenced_codeblock) {
+	if (STATE_isset(STATE_FENCED_CODEBLOCK)) {
 		$0 = substr($0, 5)
 	}
-	if (!inside_codeblock) {
-		$0 = "<pre><code" fenced_class ">" $0
-		inside_codeblock = 1
-	}
-
-	_next()
+}
+/^(    |\t)/ && /[^ \t]/ { ## New codeblock
+	STATE_set(STATE_CODEBLOCK)
+	handle_codeblock()
+	$0 = "<pre><code>" $0
+	soft_next()
 }
 
 # Definition lists
-inside_definition_list == 2 && !emptyP1 && ! /^: / {
-	inside_definition_list = 0
-	P2 = P2 "</dl>"
-}
-inside_definition_list == 1 {
-	## If empty, get ready to close dl
-	if ($0 !~ /[^ \t]/) {
-		++inside_definition_list
+STATE_isset(STATE_DEFINITION_LIST) {
+	if (!empty0) {
+		STATE_unset(STATE_PARAGRAPH)
+	}
+	else if (def_started) {
+		P1 = P1 "</dd>"
 	}
 
-	## If nonempty, we're continuing the dd
-	if (!empty0) {
-		not_paragraph = 1
+	if ((!empty0 &&  emptyP1 && /^: /) ||
+	    ( empty0 && !emptyP1 && !def_started && P1 !~ /^<dd>/) ||
+	    (!empty0 && !emptyP1 && !def_started && !/^: /)) {
+		STATE_unset(STATE_DEFINITION_LIST)
+		P2 = P2 "</dl>"
+
+		STATE_set(STATE_PARAGRAPH_P1)
+		STATE_set(STATE_PARAGRAPH)
+		P1 = "<p>" P1
 	}
-	else {
-		PNappnd("</dd>")
+
+	if (/^: /) {
+		def_started = TRUE
+	}
+	else if (empty0) {
+		def_started = FALSE
 	}
 }
 /^: / && !emptyP1 {
@@ -306,16 +307,18 @@ inside_definition_list == 1 {
 	$0 = substr($0, 3)
 
 	P1 = "<dt>" P1 "</dt>"
-	if (!inside_definition_list) {
+	if (STATE_thenset(STATE_DEFINITION_LIST)) {
 		P1 = "<dl>" P1
 	}
 	$0 = "<dd>" $0
+	def_started = TRUE
 
-	inside_definition_list = 1
-	not_paragraph = 1
-	not_paragraph_p1 = 1
+	STATE_unset(STATE_PARAGRAPH)
+	STATE_unset(STATE_PARAGRAPH_P1)
 }
 
+#== Inline markup ==#
+
 # Automatic links
 /<[^ \t<>]+>/ {
 	## URL: <uri_scheme:...>
@@ -367,7 +370,7 @@ inside_definition_list == 1 {
 		gsub(/[ \t]*#+[ \t]*$/, "")
 
 		$0 = "<h" RLENGTH ">" $0 "</h" RLENGTH ">"
-		not_paragraph = 1
+		STATE_unset(STATE_PARAGRAPH)
 	}
 }
 
@@ -376,7 +379,7 @@ inside_definition_list == 1 {
 	hlevel = 1 + ($0 ~ /-/)
 	P1 = "<h" hlevel ">" substr(P1, 4) "</h" hlevel ">"
 	$0 = ""
-	not_paragraph = 1
+	STATE_unset(STATE_PARAGRAPH)
 }
 
 # Emphasis
@@ -420,11 +423,17 @@ function gsurround(brace, left, right, blen) {
 	gsurround("\\^", "<sup>"   , "</sup>", 1)
 }
 
+# Footnote references
+/\[\^[^\]]+\]/ {
+	while (match($0, /\[\^[\]]+\]/)) {
+	}
+}
+
 # Links
 /!?\[[^\]]*\] ?[(\[][^)\]]*[\])]/ {
 	while (match($0, /!?\[[^\]]*\] ?[(\[][^)\]]*[\])]/)) {
 		link = substr($0, RSTART, RLENGTH)
-		backup_match()
+		save_match()
 
 		## ](xxxxx)
 		match(link, /\] ?[(\[][^)\]]*[\])]$/)
@@ -460,8 +469,8 @@ function gsurround(brace, left, right, blen) {
 }
 
 # Paragraph
-emptyP1 && !empty0 && !not_paragraph { $0 = "<p>" $0 }
-!emptyP1 && empty0 && !not_paragraph_p1 { P1 = P1 "</p>" }
+emptyP1 && !empty0 && STATE_isset(STATE_PARAGRAPH) { $0 = "<p>" $0 }
+!emptyP1 && empty0 && STATE_isset(STATE_PARAGRAPH_P1) { P1 = P1 "</p>" }
 
 # After all $0 parsing
 {
@@ -514,7 +523,7 @@ function __buffered_print() {
 }
 
 END {
-	if (inside_definition_list) {
+	if (STATE_isset(STATE_DEFINITION_LIST)) {
 		PNappnd("</dl>")
 	}
 
@@ -527,3 +536,107 @@ END {
 	print
 	if (!empty(N1)) { print N1 }
 }
+
+#== Boolean "constants" ==#
+BEGIN {
+	FALSE = 0
+	TRUE = 1
+}
+
+#== States system ==#
+# To convey the current state of parsing, we're using "states"
+# These are nothing more than variables and an array, storing "booleans"
+# The main benefit is semantics and syntax usage
+# (especially since we're using variables, editors will offer autocompletion)
+
+BEGIN {
+	# State definitions
+	STATE_HTML = 0
+	STATE_FENCED_CODEBLOCK = 1
+	STATE_CODEBLOCK = 2
+	STATE_BLOCKQUOTE = 3
+	STATE_PARAGRAPH = 4
+	STATE_PARAGRAPH_P1 = 5
+	STATE_LIST = 6
+	STATE_DEFINITION_LIST = 7
+}
+function STATE_set(name) {
+	STATE_unset(STATE_PARAGRAPH)
+	__state[name] = TRUE
+}
+function STATE_assign(name, other) {
+	__state[name] = __state[other]
+}
+function STATE_unset(name) {
+	__state[name] = FALSE
+}
+function STATE_isset(name) {
+	return __state[name]
+}
+function STATE_thenset(name) {
+	if (!__state[name]) {
+		STATE_set(name)
+		return TRUE
+	}
+	return FALSE
+}
+function STATE_thenunset(name) {
+	if (__state[name]) {
+		STATE_unset(STATE_PARAGRAPH)
+		__state[name] = FALSE
+		return TRUE
+	}
+	return FALSE
+}
+
+#== Line information ==#
+
+function indentation(    len) {
+	save_match()
+		match($0, /^[ \t]*/)
+		len = RLENGTH
+	restore_match()
+	return len
+}
+
+#== Match handling ==#
+
+function save_match() {
+	backup_rstart = RSTART
+	backup_rlength = RLENGTH
+}
+function restore_match() {
+	RSTART = backup_rstart
+	RLENGTH = backup_rlength
+}
+
+
+#== Other ==#
+
+function empty(value) {
+	return value == ""
+}
+
+function PNappnd(value) {
+	if (emptyP1) P2 = P2 ORS value
+	else         P1 = P1 ORS value
+}
+
+function soft_next() {
+	__buffered_print()
+
+	STATE_unset(STATE_PARAGRAPH)
+	next
+}
+
+function sub_match(replacement) {
+	$0 = substr($0, 1, RSTART - 1) replacement substr($0, RSTART + RLENGTH)
+}
+
+function char(string, ind) {
+	if (1 > ind || ind > length(string)) {
+		return "";
+	}
+	return substr(string, ind, 1)
+}
+
-- 
cgit v1.3.1-23-gd82d