From 20a276ea451cae1f07009486b622302a7b832673 Mon Sep 17 00:00:00 2001
From: yuin <yuin@inforno.net>
Date: Sat, 16 Oct 2021 19:39:51 +0900
Subject: [PATCH] #248 - 4

---
 _test/extra.txt      | 16 ++++++++++++++++
 parser/html_block.go |  4 ++--
 parser/raw_html.go   |  5 +++--
 3 files changed, 21 insertions(+), 4 deletions(-)

diff --git a/_test/extra.txt b/_test/extra.txt
index 9b90cf2..b5246d3 100644
--- a/_test/extra.txt
+++ b/_test/extra.txt
@@ -510,3 +510,19 @@ _a[b_c_](d)
 //- - - - - - - - -//
 <p>d</p>
 //= = = = = = = = = = = = = = = = = = = = = = = =//
+
+40: Invalid HTML tag names
+//- - - - - - - - -//
+<1>
+
+<a:>
+
+<a\f>
+
+< p>
+//- - - - - - - - -//
+<p>&lt;1&gt;</p>
+<p>&lt;a:&gt;</p>
+<p>&lt;a\f&gt;</p>
+<p>&lt; p&gt;</p>
+//= = = = = = = = = = = = = = = = = = = = = = = =//
diff --git a/parser/html_block.go b/parser/html_block.go
index db2cf11..531d2fb 100644
--- a/parser/html_block.go
+++ b/parser/html_block.go
@@ -91,9 +91,9 @@ var htmlBlockType4Close = []byte{'>'}
 var htmlBlockType5OpenRegexp = regexp.MustCompile(`^[ ]{0,3}<\!\[CDATA\[`)
 var htmlBlockType5Close = []byte{']', ']', '>'}
 
-var htmlBlockType6Regexp = regexp.MustCompile(`^[ ]{0,3}</?([a-zA-Z0-9]+)(?:\s.*|>.*|/>.*|)\n?$`)
+var htmlBlockType6Regexp = regexp.MustCompile(`^[ ]{0,3}<(?:/[ ]*)?([a-zA-Z]+[a-zA-Z0-9\-]*)(?:[ ].*|>.*|/>.*|)\n?$`)
 
-var htmlBlockType7Regexp = regexp.MustCompile(`^[ ]{0,3}<(/)?\s*([a-zA-Z0-9\-]+)(` + attributePattern + `*)\s*(:?>|/>)\s*\n?$`)
+var htmlBlockType7Regexp = regexp.MustCompile(`^[ ]{0,3}<(/[ ]*)?([a-zA-Z]+[a-zA-Z0-9\-]*)(` + attributePattern + `*)[ ]*(?:>|/>)[ ]*\n?$`)
 
 type htmlBlockParser struct {
 }
diff --git a/parser/raw_html.go b/parser/raw_html.go
index 7fd696c..43bbab7 100644
--- a/parser/raw_html.go
+++ b/parser/raw_html.go
@@ -48,8 +48,9 @@ func (s *rawHTMLParser) Parse(parent ast.Node, block text.Reader, pc Context) as
 }
 
 var tagnamePattern = `([A-Za-z][A-Za-z0-9-]*)`
-var attributePattern = `(?:\s+[a-zA-Z_:][a-zA-Z0-9:._-]*(?:\s*=\s*(?:[^\"'=<>` + "`" + `\x00-\x20]+|'[^']*'|"[^"]*"))?)`
-var openTagRegexp = regexp.MustCompile("^<" + tagnamePattern + attributePattern + `*\s*/?>`)
+
+var attributePattern = `(?:[\n \t]+[a-zA-Z_:][a-zA-Z0-9:._-]*(?:[\n \t]*=[\n \t]*(?:[^\"'=<>` + "`" + `\x00-\x20]+|'[^']*'|"[^"]*"))?)`
+var openTagRegexp = regexp.MustCompile("^<" + tagnamePattern + attributePattern + `*[ \t]*/?>`)
 var closeTagRegexp = regexp.MustCompile("^</" + tagnamePattern + `\s*>`)
 var commentRegexp = regexp.MustCompile(`^<!---->|<!--(?:-?[^>-])(?:-?[^-])*-->`)
 var processingInstructionRegexp = regexp.MustCompile(`^(?:<\?).*?(?:\?>)`)