Fix mw.ustring.gmatch and patterns with '^'

The Lua manual says this:

 For this function, a '^' at the start of a pattern does not work as an
 anchor, as this would prevent the iteration.

I had interpreted that to mean that a pattern starting with '^' would
never match in gmatch. But further testing reveals that the '^' is just
treated as a literal character: string.gmatch( "foo ^bar baz", "^%a+" )
will match "^bar".

Change-Id: Id91d6ee2db753ce1d6a4f6ae27764691d9e9fdc4

engines/LuaCommon/UstringLibrary.php

@@ -232,7 +232,7 @@ class Scribunto_LuaUstringLibrary extends Scribunto_LuaLibraryBase {
 	}
 
 	/* Convert a Lua pattern into a PCRE regex */
-	private function patternToRegex( $pattern ) {
+	private function patternToRegex( $pattern, $noAnchor = false ) {
 		$pat = preg_split( '//us', $pattern, null, PREG_SPLIT_NO_EMPTY );
 
 		static $charsets = null, $brcharsets = null;
@@ -295,7 +295,7 @@ class Scribunto_LuaUstringLibrary extends Scribunto_LuaLibraryBase {
 			switch ( $pat[$i] ) {
 			case '^':
 				$q = $i;
-				$re .= $q ? '\\^' : '^';
+				$re .= ( $noAnchor || $q ) ? '\\^' : '^';
 				break;
 
 			case '$':
@@ -497,7 +497,7 @@ class Scribunto_LuaUstringLibrary extends Scribunto_LuaLibraryBase {
 		$this->checkString( 'gmatch', $s );
 		$this->checkPattern( 'gmatch', $pattern );
 
-		list( $re, $capt ) = $this->patternToRegex( $pattern );
+		list( $re, $capt ) = $this->patternToRegex( $pattern, true );
 		return array( $re, $capt );
 	}
 

engines/LuaCommon/lualib/mw.ustring.lua

@@ -14,10 +14,6 @@ local function php_gmatch( s, pattern )
 	checkType( 'gmatch', 1, s, 'string' )
 	checkType( 'gmatch', 2, pattern, 'string' )
 
-	if string.sub( pattern, 1, 1 ) == '^' then
-		return function() return nil end, nil, nil
-	end
-
 	local re, capt = gmatch_init( s, pattern )
 	local pos = 0
 	return function()

engines/LuaCommon/lualib/ustring/ustring.lua

@@ -397,10 +397,11 @@ setmetatable( charset_cache, { __weak = 'kv' } )
 -- @param rawpat string  Pattern
 -- @param pattern table  Exploded pattern
 -- @param init int  Starting index
+-- @param noAnchor boolean  True to ignore '^'
 -- @return int starting index of the match
 -- @return int ending index of the match
 -- @return string|int* captures
-local function find( s, cps, rawpat, pattern, init )
+local function find( s, cps, rawpat, pattern, init, noAnchor )
 	local charsets = require 'ustring/charsets'
 	local anchor = false
 	local ncapt, captures
@@ -676,7 +677,7 @@ local function find( s, cps, rawpat, pattern, init )
 	-- match.
 	local sp = init
 	local pp = 1
-	if pattern.codepoints[1] == 0x5e then -- '^': Pattern is anchored
+	if not noAnchor and pattern.codepoints[1] == 0x5e then -- '^': Pattern is anchored
 		anchor = true
 		pp = 2
 	end
@@ -838,15 +839,8 @@ function ustring.gmatch( s, pattern )
 	end
 	local init = 1
 
-	if pat.codepoints[1] == 0x5e then -- '^': Pattern is anchored
-		-- Lua special-cases this to never match
-		return function ()
-			return nil
-		end
-	end
-
 	return function ()
-		local m = { find( s, cps, pattern, pat, init ) }
+		local m = { find( s, cps, pattern, pat, init, true ) }
 		if not m[1] then
 			return nil
 		end

tests/engines/LuaCommon/UstringLibraryTests.lua

@@ -396,8 +396,8 @@ return testframework.getTestProvider( {
 	  type = 'Iterator'
 	},
 	{ name = 'gmatch: anchored', func = mw.ustring.gmatch,
-	  args = { str2, '^f%a+' },
-	  expect = {},
+	  args = { "fóó1 ^fóó2 fóó3 ^fóó4", '^fóó%d+' },
+	  expect = { { "^fóó2" }, { "^fóó4" } },
 	  type = 'Iterator'
 	},