From d8314539daa86b000de0ad603b967e91563a07f8 Mon Sep 17 00:00:00 2001
From: Brad Jorsch <bjorsch@wikimedia.org>
Date: Wed, 3 Jul 2013 11:41:25 -0400
Subject: [PATCH] Fix mw.ustring edge cases

A few edge cases were being incorrectly handled:
* mw.ustring.sub( 'abc', 1, 0 ) returned 'a', not ''.
* mw.ustring.codepoint( 'abc', 1, 0 ) returned 97, not no results.
* mw.ustring.codepoint( 'abc', 4, 4 ) returned 99, not no results.
* mw.ustring.gcodepoint had the same issues as mw.ustring.codepoint.

Change-Id: Ib8c0ef5a8073106eb7d90d0aa0513be4525dca08
---
 engines/LuaCommon/UstringLibrary.php          | 10 ++++--
 engines/LuaCommon/lualib/ustring/ustring.lua  | 10 ++++--
 .../engines/LuaCommon/UstringLibraryTests.lua | 36 ++++++++++++++++++-
 3 files changed, 51 insertions(+), 5 deletions(-)

diff --git a/engines/LuaCommon/UstringLibrary.php b/engines/LuaCommon/UstringLibrary.php
index fddffa83..3c164864 100644
--- a/engines/LuaCommon/UstringLibrary.php
+++ b/engines/LuaCommon/UstringLibrary.php
@@ -137,8 +137,11 @@ class Scribunto_LuaUstringLibrary extends Scribunto_LuaLibraryBase {
 		if ( $j < 0 ) {
 			$j = $l + $j + 1;
 		}
-		$i = max( 1, min( $i, $l ) );
-		$j = max( 1, min( $j, $l ) );
+		if ( $j < $i ) {
+			return array();
+		}
+		$i = max( 1, min( $i, $l + 1 ) );
+		$j = max( 1, min( $j, $l + 1 ) );
 		$s = mb_substr( $s, $i - 1, $j - $i + 1, 'UTF-8' );
 		return unpack( 'N*', mb_convert_encoding( $s, 'UTF-32BE', 'UTF-8' ) );
 	}
@@ -203,6 +206,9 @@ class Scribunto_LuaUstringLibrary extends Scribunto_LuaLibraryBase {
 		if ( $j < 0 ) {
 			$j = $len + $j + 1;
 		}
+		if ( $j < $i ) {
+			return array( '' );
+		}
 		$i = max( 1, min( $i, $len + 1 ) );
 		$j = max( 1, min( $j, $len + 1 ) );
 		$s = mb_substr( $s, $i - 1, $j - $i + 1, 'UTF-8' );
diff --git a/engines/LuaCommon/lualib/ustring/ustring.lua b/engines/LuaCommon/lualib/ustring/ustring.lua
index 87f3b4ab..55766243 100644
--- a/engines/LuaCommon/lualib/ustring/ustring.lua
+++ b/engines/LuaCommon/lualib/ustring/ustring.lua
@@ -241,8 +241,11 @@ function ustring.codepoint( s, i, j )
 	if j < 0 then
 		j = cps.len + j + 1
 	end
-	i = math.max( 1, math.min( i, cps.len ) )
-	j = math.max( 1, math.min( j, cps.len ) )
+	if j < i then
+		return -- empty result set
+	end
+	i = math.max( 1, math.min( i, cps.len + 1 ) )
+	j = math.max( 1, math.min( j, cps.len + 1 ) )
 	return unpack( cps.codepoints, i, j )
 end
 
@@ -352,6 +355,9 @@ function ustring.sub( s, i, j )
 	if j < 0 then
 		j = cps.len + j + 1
 	end
+	if j < i then
+		return ''
+	end
 	i = math.max( 1, math.min( i, cps.len + 1 ) )
 	j = math.max( 1, math.min( j, cps.len + 1 ) )
 	return sub( s, cps, i, j )
diff --git a/tests/engines/LuaCommon/UstringLibraryTests.lua b/tests/engines/LuaCommon/UstringLibraryTests.lua
index c0e57600..4b0c0391 100644
--- a/tests/engines/LuaCommon/UstringLibraryTests.lua
+++ b/tests/engines/LuaCommon/UstringLibraryTests.lua
@@ -91,11 +91,22 @@ return testframework.getTestProvider( {
 	  args = { str1, 1, -1 },
 	  expect = { 0, 0x7f, 0x80, 0x7ff, 0x800, 0xffff, 0x10000, 0x10ffff }
 	},
-
 	{ name = 'codepoint: substring', func = mw.ustring.codepoint,
 	  args = { str1, 5, -2 },
 	  expect = { 0x800, 0xffff, 0x10000 }
 	},
+	{ name = 'codepoint: (5,4)', func = mw.ustring.codepoint,
+	  args = { str1, 5, 4 },
+	  expect = {}
+	},
+	{ name = 'codepoint: (1,0)', func = mw.ustring.codepoint,
+	  args = { str1, 1, 0 },
+	  expect = {}
+	},
+	{ name = 'codepoint: (9,9)', func = mw.ustring.codepoint,
+	  args = { str1, 9, 9 },
+	  expect = {}
+	},
 
 	{ name = 'char: basic test', func = mw.ustring.char,
 	  args = { 0, 0x7f, 0x80, 0x7ff, 0x800, 0xffff, 0x10000, 0x10ffff },
@@ -151,10 +162,18 @@ return testframework.getTestProvider( {
 	  args = { str1, 4, 3 },
 	  expect = { "" }
 	},
+	{ name = 'sub: (1,0)', func = mw.ustring.sub,
+	  args = { str2, 1, 0 },
+	  expect = { "" }
+	},
 	{ name = 'sub: (5,5)', func = mw.ustring.sub,
 	  args = { str1, 5, 5 },
 	  expect = { "\224\160\128" }
 	},
+	{ name = 'sub: (9,9)', func = mw.ustring.sub,
+	  args = { str1, 9, 9 },
+	  expect = { "" }
+	},
 	{ name = 'sub: empty string', func = mw.ustring.sub,
 	  args = { '', 5 },
 	  expect = { "" }
@@ -456,6 +475,21 @@ return testframework.getTestProvider( {
 	  expect = { { 0x7ff }, { 0x800 }, { 0xffff }, { 0x10000 } },
 	  type = 'Iterator'
 	},
+	{ name = 'gcodepoint: (4, 3)', func = mw.ustring.gcodepoint,
+	  args = { str1, 4, 3 },
+	  expect = {},
+	  type = 'Iterator'
+	},
+	{ name = 'gcodepoint: (1, 0)', func = mw.ustring.gcodepoint,
+	  args = { str1, 1, 0 },
+	  expect = {},
+	  type = 'Iterator'
+	},
+	{ name = 'gcodepoint: (9, 9)', func = mw.ustring.gcodepoint,
+	  args = { str1, 9, 9 },
+	  expect = {},
+	  type = 'Iterator'
+	},
 
 	{ name = 'gmatch: test string 1', func = mw.ustring.gmatch,
 	  args = { str2, 'f%a+' },