Macintosh-SW/tenfourfox
1
0
Fork 0
mirror of https://github.com/classilla/tenfourfox.git synced 2024-06-06 06:29:31 +00:00
Code Issues Projects Releases Wiki Activity

#393, Bug 1135377 - Part 9: Use RegExp unicode flag in String.prototype.{match,replace,split}. r=till, f=anba

Browse Source
This commit is contained in:
Tooru Fujisawa 2015-08-07 08:13:37 +09:00 committed by Cameron Kaiser
parent c05db4075d
commit 8636a96b78
2 changed files with 95 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

60
js/src/jsstr.cpp
View File

@ -2276,6 +2276,41 @@ DoMatchLocal(JSContext* cx, const CallArgs& args, RegExpStatics* res, HandleLine
return true;
}
/* ES6 21.2.5.2.3. */
static size_t
AdvanceStringIndex(HandleLinearString input, size_t length, size_t index, bool unicode)
{
/* Steps 1-3 (implicit). */
/* Step 4: If input is latin1, there is no surrogate pair. */
if (!unicode || input->hasLatin1Chars())
return index + 1;
JS::AutoCheckCannotGC nogc;
const char16_t* S = input->twoByteChars(nogc);
/* Step 6. */
if (index + 1 >= length)
return index + 1;
/* Step 7. */
char16_t first = S[index];
/* Step 8. */
if (!unicode::IsLeadSurrogate(first))
return index + 1;
/* Step 9. */
char16_t second = S[index + 1];
/* Step 10. */
if (!unicode::IsTrailSurrogate(second))
return index + 1;
/* Step 11. */
return index + 2;
}
/* ES5 15.5.4.10 step 8. */
static bool
DoMatchGlobal(JSContext* cx, const CallArgs& args, RegExpStatics* res, HandleLinearString input,
@ -2331,6 +2366,7 @@ DoMatchGlobal(JSContext* cx, const CallArgs& args, RegExpStatics* res, HandleLin
ScopedMatchPairs matches(&cx->tempLifoAlloc());
size_t charsLen = input->length();
RegExpShared& re = g.regExp();
bool unicode = re.unicode();
for (size_t searchIndex = 0; searchIndex <= charsLen; ) {
if (!CheckForInterrupt(cx))
return false;
@ -2348,7 +2384,9 @@ DoMatchGlobal(JSContext* cx, const CallArgs& args, RegExpStatics* res, HandleLin
MatchPair& match = matches[0];
// Steps 8f(iii)(1-3).
searchIndex = match.isEmpty() ? match.limit + 1 : match.limit;
searchIndex = match.isEmpty()
? AdvanceStringIndex(input, charsLen, match.limit, unicode)
: match.limit;
// Step 8f(iii)(4-5).
JSLinearString* str = NewDependentString(cx, input, match.start, match.length());
@ -2614,6 +2652,7 @@ static bool
DoMatchForReplaceGlobal(JSContext* cx, RegExpStatics* res, HandleLinearString linearStr,
RegExpShared& re, ReplaceData& rdata, size_t* rightContextOffset)
{
bool unicode = re.unicode();
size_t charsLen = linearStr->length();
ScopedMatchPairs matches(&cx->tempLifoAlloc());
for (size_t count = 0, searchIndex = 0; searchIndex <= charsLen; ++count) {
@ -2628,7 +2667,9 @@ DoMatchForReplaceGlobal(JSContext* cx, RegExpStatics* res, HandleLinearString li
break;
MatchPair& match = matches[0];
searchIndex = match.isEmpty() ? match.limit + 1 : match.limit;
searchIndex = match.isEmpty()
? AdvanceStringIndex(linearStr, charsLen, match.limit, unicode)
: match.limit;
*rightContextOffset = match.limit;
if (!res->updateFromMatchPairs(cx, linearStr, matches))
@ -3228,6 +3269,7 @@ StrReplaceRegexpRemove(JSContext* cx, HandleString str, RegExpShared& re)
size_t lazyIndex = 0; /* Index before last successful match. */
/* Accumulate StringRanges for unmatched substrings. */
bool unicode = re.unicode();
while (startIndex <= charsLen) {
if (!CheckForInterrupt(cx))
return nullptr;
@ -3248,7 +3290,9 @@ StrReplaceRegexpRemove(JSContext* cx, HandleString str, RegExpShared& re)
lazyIndex = lastIndex;
lastIndex = match.limit;
startIndex = match.isEmpty() ? match.limit + 1 : match.limit;
startIndex = match.isEmpty()
? AdvanceStringIndex(linearStr, charsLen, match.limit, unicode)
: match.limit;
/* Non-global removal executes at most once. */
if (!re.global())
@ -3631,7 +3675,7 @@ class SplitMatchResult {
template<class Matcher>
static JSObject*
SplitHelper(JSContext* cx, HandleLinearString str, uint32_t limit, const Matcher& splitMatch,
HandleObjectGroup group)
HandleObjectGroup group, bool unicode)
{
size_t strLength = str->length();
SplitMatchResult result;
@ -3696,7 +3740,7 @@ SplitHelper(JSContext* cx, HandleLinearString str, uint32_t limit, const Matcher
/* Step 13(c)(ii). */
if (endIndex == lastEndIndex) {
index++;
index = AdvanceStringIndex(str, strLength, index, unicode);
continue;
}
@ -3925,14 +3969,14 @@ js::str_split(JSContext* cx, unsigned argc, Value* vp)
aobj = CharSplitHelper(cx, linearStr, limit, group);
} else {
SplitStringMatcher matcher(cx, sepstr);
aobj = SplitHelper(cx, linearStr, limit, matcher, group);
aobj = SplitHelper(cx, linearStr, limit, matcher, group, false);
}
} else {
RegExpStatics* res = cx->global()->getRegExpStatics(cx);
if (!res)
return false;
SplitRegExpMatcher matcher(*re, res);
aobj = SplitHelper(cx, linearStr, limit, matcher, group);
aobj = SplitHelper(cx, linearStr, limit, matcher, group, re->unicode());
}
if (!aobj)
return false;
@ -3960,7 +4004,7 @@ js::str_split_string(JSContext* cx, HandleObjectGroup group, HandleString str, H
return CharSplitHelper(cx, linearStr, limit, group);
SplitStringMatcher matcher(cx, linearSep);
return SplitHelper(cx, linearStr, limit, matcher, group);
return SplitHelper(cx, linearStr, limit, matcher, group, false);
}
/*

43
js/src/tests/ecma_6/String/AdvanceStringIndex.js Normal file
View File

@ -0,0 +1,43 @@
var BUGNUMBER = 1135377;
var summary = "Implement RegExp unicode flag -- AdvanceStringIndex in global match and replace.";
print(BUGNUMBER + ": " + summary);
// ==== String.prototype.match ====
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".match(/\uD83D|X|/gu),
["", "", "X", "", ""]);
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".match(/\uDC38|X|/gu),
["", "", "X", "", ""]);
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".match(/\uD83D\uDC38|X|/gu),
["\uD83D\uDC38", "", "X", "", ""]);
// ==== String.prototype.replace ====
// empty string replacement (optimized)
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".replace(/\uD83D|X|/gu, ""),
"\uD83D\uDC38\uD83D\uDC39\uD83D\uDC3A");
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".replace(/\uDC38|X|/gu, ""),
"\uD83D\uDC38\uD83D\uDC39\uD83D\uDC3A");
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".replace(/\uD83D\uDC38|X|/gu, ""),
"\uD83D\uDC39\uD83D\uDC3A");
// non-empty string replacement
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".replace(/\uD83D|X|/gu, "x"),
"x\uD83D\uDC38x\uD83D\uDC39xx\uD83D\uDC3Ax");
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".replace(/\uDC38|X|/gu, "x"),
"x\uD83D\uDC38x\uD83D\uDC39xx\uD83D\uDC3Ax");
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".replace(/\uD83D\uDC38|X|/gu, "x"),
"xx\uD83D\uDC39xx\uD83D\uDC3Ax");
// ==== String.prototype.split ====
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".split(/\uD83D|X|/u),
["\uD83D\uDC38", "\uD83D\uDC39", "\uD83D\uDC3A"]);
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".split(/\uDC38|X|/u),
["\uD83D\uDC38", "\uD83D\uDC39", "\uD83D\uDC3A"]);
assertEqArray("\uD83D\uDC38\uD83D\uDC39X\uD83D\uDC3A".split(/\uD83D\uDC38|X|/u),
["", "\uD83D\uDC39", "\uD83D\uDC3A"]);
if (typeof reportCompare === "function")
reportCompare(true, true);