feat(single)!: Rewrite using buffered output

This should provide a significantly simpler codebase.
Implemented paragraphs, line break, headers and blockquotes.

1 files changed, 99 insertions, 298 deletions

diff --git a/single.awk b/single.awk
index 78ae873..5ddcb94 100644
--- a/single.awk
+++ b/single.awk
@@ -1,34 +1,34 @@
 # Status:
 # [.] Inline HTML
-# [x] Autoescape
+# [ ] Autoescape
 # [.] Block elements
 #     [x] Paragraph
 #     [x] Line break
 #     [x] Headers (atx and setex)
 #     [x] Blockquotes
-#     [x] Lists
+#     [ ] Lists
 #     [ ] Code blocks
-#     [x] Horizontal rules
-# [x] Span elements
-#     [x] Links
-#         [x] Inline
-#         [x] Reference
-#     [x] Emphasis
-#     [x] Code
-#     [x] Images
-#         [x] Inline
-#         [x] Reference
+#     [ ] Horizontal rules
+# [ ] Span elements
+#     [ ] Links
+#         [ ] Inline
+#         [ ] Reference
+#     [ ] Emphasis
+#     [ ] Code
+#     [ ] Images
+#         [ ] Inline
+#         [ ] Reference
 # [ ] Other
 #     [ ] Autolinks
 #     [ ] Escapes
 #
-# [.] Non-standard features
-#     [x] Strikethrough ~~...~~
-#     [x] Superscript ^...^
-#     [x] Subscript ~...~
+# [ ] Non-standard features
+#     [ ] Strikethrough ~~...~~
+#     [ ] Superscript ^...^
+#     [ ] Subscript ~...~
 #     [ ] Footnote [^...] (with user customizations)
-#     [x] Task list - [ ] (readonly)
-#     [.] Definition lists
+#     [ ] Task list - [ ] (readonly)
+#     [ ] Definition lists
 #         [x] Single line
 #         [ ] Multi-line
 #     [ ] Fenced code blocks ```lang ```
@@ -46,94 +46,56 @@
 # - emphasis inside inline code block is still converted to HTML
 # - backticks inside `` emphasis blocks aren't escaped
 
-BEGIN {
-	prev_blankline = blankline = 1
-	skipprint = 1
-	newStack("bq-level")
-	newStack("list-level")
-}
-
-function println(line) {
-	if (0 == line) line = $0
-	print prev
-	prev = line
-}
-
-function autoinsert(arrname, isbefore) {
-	if (isbefore == "") isbefore = 0
-	for (confstr in arrname) {
-		split(confstr, conf, SUBSEP)
-
-		## FILENAME regex, ".*" (default)
-		if (2 in conf && FILENAME !~ conf[2]) { continue }
-		## Text regex, ".*" (default)
-		if (3 in conf && text !~ conf[3]) { continue }
-
-		if (isbefore) prev = arrname[confstr] prev
-		else          prev = prev arrname[confstr]
-
-		## When, FIRST | ALL
-		if (conf[1] == FIRST) { delete arrname[confstr] }
-	}
+function empty(value) {
+	return value == ""
 }
 
 {
-	prev_blankline = blankline
-	blankline = 1
+	not_paragraph_p1 = not_paragraph
+	not_paragraph = 0
+	empty0  = empty($0)
+	emptyP1 = empty(P1)
 }
 
-/[^ \t]/ {
-	blankline = 0
-}
+# Output is buffered on multiple lines, held in these variables:
+# P2 = previous non-empty line
+# P1 = previous line
+# $0 = current line
+# N1 = next new line (N1 is always set by the user)
 
 # Inline HTML
 /^[ \t]*<\/?[0-9A-Za-z]+/ {
-	println()
-	next
+	not_paragraph = 1
 }
 
-# Link definition
-/^\[.+\]:/ {
-	blankline = 1
-	next
+# Blockquotes
+blockquote_level > 0 && empty0 {
+	for (P1 = P1 ORS; blockquote_level > 0; --blockquote_level) {
+		P1 = P1 "</blockquote>"
+	}
+	not_paragraph_p1 = 1
 }
+/^>[ \t]/ {
+	match($0, /^[> \t]+[ \t]+/)
+	level = substr($0, RSTART, RLENGTH)
+	$0 = substr($0, RSTART + RLENGTH)
+	gsub(/[ \t]/, "", level)
+	level = length(level)
 
-# Auto-escape &
-/&/ {
-	ampstart = 0
-	## gsub(/&(?![[:alnum:]#]+;)/, "&amp;")
-	while (match(substr($0, ampstart + 1), /&/)) {
-		ampstart += RSTART
-		if (match(substr($0, ampstart), /&[0-9A-Za-z#]+;/)) { continue }
-
-		$0 = substr($0, 1, ampstart - 1) "&amp;" substr($0, ampstart + 1)
+	for (;blockquote_level < level; ++blockquote_level) {
+		P1 = P1 "<blockquote>"
 	}
+	if (blockquote_level > level) { P2 = P2 ORS }
+	for (;blockquote_level > level; --blockquote_level) {
+		P2 = P2 "</blockquote>"
+	}
+	not_paragraph = 1
 }
-# Auto-escape <
-/</ {
-	gsub(/</, "\\&lt;")
-}
-
-# Blockquote Part 1/2
-/^>[ \t]*$/ { next }
-/^> / {
-	level = 0
-	while (match($0, /^> [ \t]*/)) {
-		$0 = substr($0, RLENGTH + 1)
-		level += RSTART
-		while (topStack("bq-level") > level) {
-			println("</blockquote>")
-			autoinsert(user_BLOCKQUOTE_AFTER)
+/^>$/ { $0 = ""; not_paragraph = 1 }
 
-			popStack("bq-level")
-		}
-		if (topStack("bq-level") < level) {
-			autoinsert(user_BLOCKQUOTE_BEFORE, 1)
-			println("<blockquote>")
-
-			pushStack("bq-level", level)
-		}
-	}
+# Line break
+/  $/ {
+	sub(/[ \t]*  $/, "</br>")
 }
 
 # atx-style headers
@@ -144,232 +106,71 @@ function autoinsert(arrname, isbefore) {
 		gsub(/^[ \t]*#+[ \t]*/, "")
 		gsub(/[ \t]*#+[ \t]*$/, "")
 
-		     if (RLENGTH == 1) autoinsert(user_H1_BEFORE, 1)
-		else if (RLENGTH == 2) autoinsert(user_H2_BEFORE, 1)
-		else if (RLENGTH == 3) autoinsert(user_H3_BEFORE, 1)
-		else if (RLENGTH == 4) autoinsert(user_H4_BEFORE, 1)
-		else if (RLENGTH == 5) autoinsert(user_H5_BEFORE, 1)
-		else if (RLENGTH == 6) autoinsert(user_H6_BEFORE, 1)
-		autoinsert(user_H_BEFORE, 1)
-
-		println("<h" RLENGTH ">" $0 "</h" RLENGTH ">")
-
-		autoinsert(user_H_AFTER)
-		     if (RLENGTH == 1) autoinsert(user_H1_AFTER)
-		else if (RLENGTH == 2) autoinsert(user_H2_AFTER)
-		else if (RLENGTH == 3) autoinsert(user_H3_AFTER)
-		else if (RLENGTH == 4) autoinsert(user_H4_AFTER)
-		else if (RLENGTH == 5) autoinsert(user_H5_AFTER)
-		else if (RLENGTH == 6) autoinsert(user_H6_AFTER)
-
-		blankline = 1
-		next
+		$0 = "<h" RLENGTH ">" $0 "</h" RLENGTH ">"
+		not_paragraph = 1
 	}
 }
+
 # setex-style headers
-!prev_blankline && /^[ \t]*(=|-)+[ \t]*$/ {
-	sub(/^<p>/, "", prev) ## from paragraph logic
+!emptyP1 && /^[ \t]*(=|-)+[ \t]*$/ {
 	hlevel = 1 + ($0 ~ /-/)
-
-	prev =  "<h" hlevel ">" prev "</h" hlevel ">"
-
-	if (hlevel == 1) autoinsert(user_H1_BEFORE, 1)
-	else             autoinsert(user_H2_BEFORE, 1)
-	autoinsert(user_H_BEFORE, 1)
-
-	autoinsert(user_H_AFTER)
-	if (hlevel == 1) autoinsert(user_H1_AFTER)
-	else             autoinsert(user_H2_AFTER)
-
-	blankline = 1
-	next
+	P1 = "<h" hlevel ">" substr(P1, 4) "</h" hlevel ">"
+	$0 = ""
+	not_paragraph = 1
 }
 
-# Horizontal rules
-/^(-|\*|_) ?(-|\*|_) ?(-|\*|_|[ \t])+$/ {
-	autoinsert(user_HR_BEFORE, 1)
-	println("<hr>")
-	autoinsert(user_HR_AFTER)
-	blankline = 1
-	next
-}
+# Paragraph
+emptyP1 && !empty0 && !not_paragraph { $0 = "<p>" $0 }
+!emptyP1 && empty0 && !not_paragraph_p1 { P1 = P1 "</p>" }
 
-# Line break
-/  $/ {
-	sub(/[ \t]*  $/, "</br>")
-}
-
-
-# Code blocks
-/^(    |\t)[ \t]*/ {
-}
+# After all $0 parsing
+{
+	empty0  = empty($0)
+	emptyP1 = empty(P1)
 
-# Check boxes Part 1/2
-/^[ \t]*- \[[ x]\][ \t]/ {
-	if (checkbox_level <= 0) {
-		println("<ul class=\"checkboxed\">")
-		checkbox_level = 1
+	if (empty0 && emptyP1) {
+		++P1_repeats
 	}
-	match($0, /- \[[ x]\][ \t]/)
-	if (substr($0, RSTART + 3, 1) == "x") checked = " checked=\"checked\""
-	else                                  checked = ""
-	$0 = "<li><input type=\"checkbox\" disabled=\"disabled\"" checked "><label>" substr($0, RSTART + RLENGTH) "</label></li>"
-}
-
-# Unordered lists
-function abs(x) { return ((x < 0) ? -x : x) }
-!blankline && lengthStack("list-level") > 0 {
-	match($0, /^ */)
-	indent = RLENGTH
-	match($0, /(\*|\+|-|[0-9]+\.) /)
-	if (RLENGTH == -1) RLENGTH = 0
-	## indent + RLENGTH = length("  * " or " 1. ")
-	while (indent + RLENGTH < abs(topStack("list-level"))) {
-		if (topStack("list-level") < 0) {
-			println("</ol>")
-			autoinsert(user_L_AFTER)
-			autoinsert(user_OL_AFTER)
-		}
-		else {
-			println("</ul>")
-			autoinsert(user_L_AFTER)
-			autoinsert(user_UL_AFTER)
-		}
-		popStack("list-level")
+	else if (!empty0 && emptyP1) {
+		P1 = $0
 	}
-}
-/^[ \t]*(\*|\+|-|[0-9]+\.) / {
-	match($0, /(\*|\+|-|[0-9]+\.) /)
-	if (abs(topStack("list-level")) < RSTART) {
-		if (RLENGTH > 2) {
-			xl = "<ol>"; mult = -1
-			autoinsert(user_OL_BEFORE, 1)
+	else {
+		print P2
+		while (P1_repeats > 0) {
+			print ""
+			--P1_repeats
 		}
-		else {
-			xl = "<ul>"; mult = 1
-			autoinsert(user_UL_BEFORE, 1)
+		P2 = P1
+		P1 = $0
+		if (empty0) {
+			P1_repeats = 1
 		}
-		autoinsert(user_L_BEFORE, 1)
-		println(xl "<li>")
-		pushStack("list-level", mult * (RSTART + RLENGTH - 1))
-	}
-	else {
-		println("</li><li>")
 	}
-}
-!blankline && lengthStack("list-level") > 0 {
-	$0 = substr($0, abs(topStack("list-level")) + 1)
-}
-
-# Paragraphs
-prev_blankline && !blankline && lengthStack("bq-level") == 0 && lengthStack("list-level") == 0 && checkbox_level == 0 { $0 = "<p>" $0 }
-!prev_blankline && blankline && lengthStack("bq-level") == 0 && lengthStack("list-level") == 0 && checkbox_level == 0 { prev = prev "</p>" }
-
-# Blockquote Part 2/2
-blankline {
-	while (lengthStack("bq-level") > 0) {
-		println("</blockquote>")
-		autoinsert(user_BLOCKQUOTE_AFTER)
-		popStack("bq-level")
-	}
-}
-
-# Check boxes Part 2/2
-checkbox_level > 0 && ! /^<li><input type="checkbox" disabled="disabled"/ {
-	println("</ul>")
-	checkbox_level = 0
-}
-
-# Links
-/!?\[[^\]]*\] ?[(\[][^)\]]*[\])]/ {
-	isimg = 0
-	while (match($0, /!?\[[^\]]*\] ?[(\[]/)) {
-		if (substr($0, RSTART, 1) == "!") {
-			isimg = 1
-			RSTART++
-			RLENGTH--
-		}
-		astart = RSTART ## astart = indexof("[")
-		has_space = substr($0, RSTART + RLENGTH - 2, 1) == " "
-		inner_anchor = substr($0, RSTART + 1, RLENGTH - 3 - has_space) ## ...[xxxxx](...)...
-
-		aend = RSTART + RLENGTH - 1
-		match(substr($0, astart), /\] ?[(\[][^)\]]*[\])]/)
-		aend += RLENGTH - 2 - has_space ## aend = indexof(")")
-		href = substr($0, astart + RSTART + 1 + has_space, RLENGTH - 3 - has_space) ## ...[...](xxxxx)...
 
-		title = "\"\""
-		if (substr($0, astart + RSTART + has_space, 1) == "[") { ## is reference link
-			if (href == "") href = inner_anchor
-			title = reference[href, "title"]
-			href = reference[href, "href"]
+	if (!empty(N1)) {
+		if (emptyP1) {
+			P1 = N1
 		}
-		else if (match(href, /[ \t]"[^"]*"[ \t]*$/)) { ## has title
-			title = substr(href, RSTART + 1, RLENGTH - 1)
-			href  = substr(href, 1, RSTART - 1)
+		else {
+			print P2
+			while (P1_repeats > 0) {
+				print ""
+				--P1_repeats
+			}
+			P2 = P1
+			P1 = N1
+			N1 = ""
 		}
-		sub(/[ \t]+/, "", href)
-
-		if (isimg) elem = "<img src=\"" href "\" alt=\"" inner_anchor "\" title=" title ">"
-		else       elem = "<a href=\"" href "\" title=" title ">" inner_anchor "</a>"
-
-		$0 = substr($0, 1, astart - 1 - isimg) elem substr($0, aend + 1)
-		isimg = 0
 	}
 }
 
-# Emphasis
-function surround(brace, opening, closing, blen, single_surround) {
-	if (blen == "") { blen = length(brace) }
-
-	while (match($0, brace "[^" brace "]+" brace)) {
-		elem = opening substr($0, RSTART + blen, RLENGTH - 2 * blen) closing
-		$0 = substr($0, 1, RSTART - 1) elem substr($0, RSTART + RLENGTH)
-
-		if (single_surround != "") { return; }
-	}
-}
-{
-	## Make sure <em> is on the inside for long chains
-	## E.g. *****a***** will be the same as ****_a_****
-	surround("___"      , "<strong><em>", "</em></strong>", 3, "oneshot")
-	surround("\\*\\*\\*", "<strong><em>", "</em></strong>", 3, "oneshot")
-
-	surround("__"    , "<strong>", "</strong>")
-	surround("_"     , "<em>"    , "</em>")
-	surround("\\*\\*", "<strong>", "</strong>", 2)
-	surround("\\*"   , "<em>"    , "</em>", 1)
-	surround("~~"    , "<s>"     , "</s>")
-	surround("~"     , "<sub>"   , "</sub>")
-	surround("\\^"   , "<sup>"   , "</sup>", 1)
-	surround("``"    , "<code>"  , "</code>")
-	surround("`"     , "<code>"  , "</code>")
-}
-
-# Definition lists
-!prev_blankline && /^: / {
-	sub(/^<p>/, "", prev)
-	sub(/^: /, "")
-	prev = "<dl><dt>" ORS prev ORS "</dt><dd>" ORS $0 ORS "</dd></dl>"
-	blankline = 1
-	next
-}
-
-{
-	println()
-}
-
 END {
-	print prev
-	if (!blankline && lengthStack("bq-level") == 0) { print "</p>" }
-	## Flush stacks
-	while (lengthStack("bq-level") > 0) {
-		print "</blockquote>"
-		popStack("bq-level")
-	}
-	while (lengthStack("list-level") > 0) {
-		if (topStack("list-level") < 0) print "</ol>"
-		else                            print "</ul>"
-		popStack("list-level")
+	print P2
+	while (P1_repeats > 0) {
+		print ""
+		--P1_repeats
 	}
+	if (!empty(P1)) { print P1 }
+	print
+	if (!empty(N1)) { print N1 }
 }