From f109bb918f42823df00b052b72bfa8881d24f91f Mon Sep 17 00:00:00 2001
From: "Reece H. Dunn" <msclrhd@gmail.com>
Date: Sun, 12 Feb 2017 11:11:52 +0000
Subject: [PATCH] isspace: don't include <noBreak> characters.

---
 src/ctype.c        | 10 +++++++++-
 tools/printdata.py |  6 +++++-
 2 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/src/ctype.c b/src/ctype.c
index 9f9b91b9..d9818f90 100644
--- a/src/ctype.c
+++ b/src/ctype.c
@@ -140,10 +140,18 @@ int ucd_isspace(codepoint_t c)
 	{
 	case UCD_CATEGORY_Zl:
 	case UCD_CATEGORY_Zp:
+		return 1;
 	case UCD_CATEGORY_Zs:
+		switch (c) // Exclude characters with the <noBreak> DispositionType
+		{
+		case 0x00A0: // U+00A0 : NO-BREAK SPACE
+		case 0x2007: // U+2007 : FIGURE SPACE
+		case 0x202F: // U+202F : NARROW NO-BREAK SPACE
+			return 0;
+		}
 		return 1;
 	case UCD_CATEGORY_Cc:
-		switch (c) // Some control characters are also whitespace characters:
+		switch (c) // Include control characters marked as White_Space
 		{
 		case 0x09: // U+0009 : CHARACTER TABULATION
 		case 0x0A: // U+000A : LINE FEED
diff --git a/tools/printdata.py b/tools/printdata.py
index 46a2fc1f..7c173add 100755
--- a/tools/printdata.py
+++ b/tools/printdata.py
@@ -50,7 +50,11 @@ def isxdigit(data):
 	return 1 if data['CodePoint'].char() in '0123456789ABCDEFabcdef' else 0
 
 def isspace(data):
-	return data.get('White_Space', 0)
+	if data.get('White_Space', 0):
+		dt = data.get('DecompositionType', '')
+		return 1 if dt == None or not dt.startswith('<noBreak>') else 0
+	else:
+		return 0
 
 def isupper(data):
 	if data.get('LowerCase', null) != null: