From: Fletcher T. Penney <fletcher@fletcherpenney.net>
Date: Wed, 8 Aug 2018 00:27:35 +0000 (-0500)
Subject: FIXED: Use Setext headers when necessary to convert from OPML to text (fixes #138)
X-Git-Tag: 6.4.0^2~5
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=cd6c4902545f941d6a821dfef59bfb65b11c1d63;p=multimarkdown

FIXED: Use Setext headers when necessary to convert from OPML to text (fixes #138)
---

diff --git a/Sources/libMultiMarkdown/opml-lexer.c b/Sources/libMultiMarkdown/opml-lexer.c
index ca92df9..1588ee7 100644
--- a/Sources/libMultiMarkdown/opml-lexer.c
+++ b/Sources/libMultiMarkdown/opml-lexer.c
@@ -1,4 +1,4 @@
-/* Generated by re2c 1.0.3 on Mon Aug  6 17:50:40 2018 */
+/* Generated by re2c 1.0.3 on Tue Aug  7 19:20:45 2018 */
 /**
 
 	MultiMarkdown -- Lightweight markup processor to produce HTML, LaTeX, and more.
@@ -2090,3 +2090,94 @@ yy206:
 	}
 
 }
+
+
+/// Does the string include encoded newline?
+size_t scan_encoded_newline(const char * c, size_t len) {
+	const char * marker = NULL;
+	const char * start = c;
+
+scan:
+
+	if ((*c == '\0') || ((c - start) > len)) {
+		// Not found
+		return -1;
+	}
+
+
+	{
+		unsigned char yych;
+		yych = *c;
+
+		switch (yych) {
+			case '\n':
+				goto yy210;
+
+			case '&':
+				goto yy213;
+
+			default:
+				goto yy211;
+		}
+
+yy210:
+		c = marker;
+		goto yy212;
+yy211:
+		++c;
+yy212: {
+			goto scan;
+		}
+yy213:
+		yych = *(marker = ++c);
+
+		switch (yych) {
+			case '#':
+				goto yy214;
+
+			default:
+				goto yy212;
+		}
+
+yy214:
+		yych = *++c;
+
+		switch (yych) {
+			case '1':
+				goto yy215;
+
+			default:
+				goto yy210;
+		}
+
+yy215:
+		yych = *++c;
+
+		switch (yych) {
+			case '0':
+			case '3':
+				goto yy216;
+
+			default:
+				goto yy210;
+		}
+
+yy216:
+		yych = *++c;
+
+		switch (yych) {
+			case ';':
+				goto yy217;
+
+			default:
+				goto yy210;
+		}
+
+yy217:
+		++c;
+		{
+			return (size_t)(c - start);
+		}
+	}
+
+}
diff --git a/Sources/libMultiMarkdown/opml-lexer.h b/Sources/libMultiMarkdown/opml-lexer.h
index 482d67e..e17ef09 100644
--- a/Sources/libMultiMarkdown/opml-lexer.h
+++ b/Sources/libMultiMarkdown/opml-lexer.h
@@ -150,4 +150,8 @@ size_t scan_note(const char * c);
 /// find end of double quoted value
 size_t scan_double_quoted(const char * c);
 
+/// Does the string include encoded newline?
+size_t scan_encoded_newline(const char * c, size_t len);
+
+
 #endif
diff --git a/Sources/libMultiMarkdown/opml-lexer.re b/Sources/libMultiMarkdown/opml-lexer.re
index 3146a2d..25f9360 100644
--- a/Sources/libMultiMarkdown/opml-lexer.re
+++ b/Sources/libMultiMarkdown/opml-lexer.re
@@ -138,6 +138,8 @@ int opml_scan(Scanner * s, const char * stop) {
 		text_attribute											= WSNL* 'text' WSNL* EQUAL WSNL*;
 		note_attribute											= WSNL* '_note' WSNL* EQUAL WSNL*;
 
+		contains_newline										= "&#10;" | "&#13;";
+
 		'<?xml' [^>\x00]* '>'									{ return OPML_XML; }
 
 		'<opml' [^>\x00]* '>'									{ return OPML_OPML_OPEN; }
@@ -184,7 +186,7 @@ size_t scan_text(const char * c) {
 	const char * start = c;
 
 /*!re2c
-	text_attribute / double_quoted		{ return (size_t)( c - start ); }
+	text_attribute / double_quoted			{ return (size_t)( c - start ); }
 	.?										{ return 0; }
 */	
 }
@@ -196,7 +198,7 @@ size_t scan_note(const char * c) {
 	const char * start = c;
 
 /*!re2c
-	note_attribute / double_quoted	{ return (size_t)( c - start ); }
+	note_attribute / double_quoted			{ return (size_t)( c - start ); }
 	.?										{ return 0; }
 */	
 }
@@ -212,3 +214,22 @@ size_t scan_double_quoted(const char * c) {
 	.?										{ return 0; }
 */	
 }
+
+
+/// Does the string include encoded newline?
+size_t scan_encoded_newline(const char * c, size_t len) {
+	const char * marker = NULL;
+	const char * start = c;
+
+	scan:
+
+	if ((*c == '\0') || ((c - start) > len)) {
+		// Not found
+		return -1;
+	}
+
+/*!re2c
+	contains_newline						{ return (size_t)(c - start); }
+	.										{ goto scan; }
+*/
+}
diff --git a/Sources/libMultiMarkdown/opml-reader.c b/Sources/libMultiMarkdown/opml-reader.c
index db1385c..f373fff 100644
--- a/Sources/libMultiMarkdown/opml-reader.c
+++ b/Sources/libMultiMarkdown/opml-reader.c
@@ -337,18 +337,35 @@ void parse_opml_token_chain(mmd_engine * e, token * chain) {
 							print_const(":\t");
 						} else {
 							// Print header
-							for (int i = 0; i < header_level; ++i) {
-								print_char('#');
-							}
+							if (scan_encoded_newline(&(e->dstr->str[start + 1]), len - 2) == -1) {
+								// ATX header
+								for (int i = 0; i < header_level; ++i) {
+									print_char('#');
+								}
 
-							print_char(' ');
+								print_char(' ');
+							}
 
 							print_opml_text(out, e->dstr->str, start + 1, len - 2);
 
-							print_char(' ');
-
-							for (int i = 0; i < header_level; ++i) {
-								print_char('#');
+							if (scan_encoded_newline(&(e->dstr->str[start + 1]), len - 2) == -1) {
+								// ATX header
+								print_char(' ');
+
+								for (int i = 0; i < header_level; ++i) {
+									print_char('#');
+								}
+							} else {
+								// Print Setext Header
+								switch (header_level) {
+									case 1:
+										print_const("\n======");
+										break;
+
+									default:
+										print_const("\n------");
+										break;
+								}
 							}
 
 							print_const("\n");