MarcusJohnson91 created this revision.
MarcusJohnson91 added reviewers: aaron.ballman, efriedma.
MarcusJohnson91 added a project: clang.
MarcusJohnson91 requested review of this revision.
Split from D103426 <https://reviews.llvm.org/D103426>
Repository:
rG LLVM Github Monorepo
https://reviews.llvm.org/D106756
Files:
clang/include/clang/AST/FormatString.h
clang/lib/AST/FormatString.cpp
clang/lib/AST/PrintfFormatString.cpp
clang/lib/AST/ScanfFormatString.cpp
clang/test/Sema/format-strings-int-typedefs.c
Index: clang/test/Sema/format-strings-int-typedefs.c
===================================================================
--- clang/test/Sema/format-strings-int-typedefs.c
+++ clang/test/Sema/format-strings-int-typedefs.c
@@ -12,6 +12,10 @@
printf("%td", 42.0); // expected-warning {{format specifies type 'ptrdiff_t' (aka 'int')}}
printf("%lc", 42.0); // expected-warning {{format specifies type 'wint_t' (aka 'int')}}
printf("%ls", 42.0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
+ printf("%l16c", 42.0); // expected-warning {{format specifies type 'char16_t' (aka 'int')}}
+ printf("%l16s", 42.0); // expected-warning {{format specifies type 'char16_t *' (aka 'int *')}}
+ printf("%l32c", 42.0); // expected-warning {{format specifies type 'char32_t' (aka 'int')}}
+ printf("%l32s", 42.0); // expected-warning {{format specifies type 'char32_t *' (aka 'int *')}}
printf("%S", 42.0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
printf("%C", 42.0); // expected-warning {{format specifies type 'wchar_t' (aka 'int')}}
@@ -21,6 +25,10 @@
wprintf(L"%td", 42.0); // expected-warning {{format specifies type 'ptrdiff_t' (aka 'int')}}
wprintf(L"%lc", 42.0); // expected-warning {{format specifies type 'wint_t' (aka 'int')}}
wprintf(L"%ls", 42.0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
+ wprintf(L"%l16c", 42.0); // expected-warning {{format specifies type 'char16_t' (aka 'int')}}
+ wprintf(L"%l16s", 42.0); // expected-warning {{format specifies type 'char16_t *' (aka 'int *')}}
+ wprintf(L"%l32c", 42.0); // expected-warning {{format specifies type 'char32_t' (aka 'int')}}
+ wprintf(L"%l32s", 42.0); // expected-warning {{format specifies type 'char32_t *' (aka 'int *')}}
wprintf(L"%S", 42.0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
wprintf(L"%C", 42.0); // expected-warning {{format specifies type 'wchar_t' (aka 'int')}}
@@ -30,6 +38,10 @@
scanf("%td", 0); // expected-warning {{format specifies type 'ptrdiff_t *' (aka 'int *')}}
scanf("%lc", 0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
scanf("%ls", 0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
+ scanf("%l16c", 0); // expected-warning {{format specifies type 'char16_t *' (aka 'int *')}}
+ scanf("%l16s", 0); // expected-warning {{format specifies type 'char16_t *' (aka 'int *')}}
+ scanf("%l32c", 0); // expected-warning {{format specifies type 'char32_t *' (aka 'int *')}}
+ scanf("%l32s", 0); // expected-warning {{format specifies type 'char32_t *' (aka 'int *')}}
scanf("%S", 0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
scanf("%C", 0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
@@ -39,6 +51,10 @@
wscanf("%td", 0); // expected-warning {{format specifies type 'ptrdiff_t *' (aka 'int *')}}
wscanf("%lc", 0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
wscanf("%ls", 0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
+ wscanf("%l16c", 0); // expected-warning {{format specifies type 'char16_t *' (aka 'int *')}}
+ wscanf("%l16s", 0); // expected-warning {{format specifies type 'char16_t *' (aka 'int *')}}
+ wscanf("%l32c", 0); // expected-warning {{format specifies type 'char32_t *' (aka 'int *')}}
+ wscanf("%l32s", 0); // expected-warning {{format specifies type 'char32_t *' (aka 'int *')}}
wscanf("%S", 0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
wscanf("%C", 0); // expected-warning {{format specifies type 'wchar_t *' (aka 'int *')}}
Index: clang/lib/AST/ScanfFormatString.cpp
===================================================================
--- clang/lib/AST/ScanfFormatString.cpp
+++ clang/lib/AST/ScanfFormatString.cpp
@@ -261,6 +261,8 @@
case LengthModifier::AsInt32:
case LengthModifier::AsInt3264:
case LengthModifier::AsWide:
+ case LengthModifier::AsUTF16:
+ case LengthModifier::AsUTF32:
case LengthModifier::AsShortLong:
return ArgType::Invalid();
}
@@ -302,6 +304,8 @@
case LengthModifier::AsInt32:
case LengthModifier::AsInt3264:
case LengthModifier::AsWide:
+ case LengthModifier::AsUTF16:
+ case LengthModifier::AsUTF32:
case LengthModifier::AsShortLong:
return ArgType::Invalid();
}
@@ -329,14 +333,18 @@
// Char, string and scanlist.
case ConversionSpecifier::cArg:
- case ConversionSpecifier::sArg:
+ case ConversionSpecifier::CArg:
case ConversionSpecifier::ScanListArg:
switch (LM.getKind()) {
case LengthModifier::None:
return ArgType::PtrTo(ArgType::AnyCharTy);
case LengthModifier::AsLong:
case LengthModifier::AsWide:
- return ArgType::PtrTo(ArgType(Ctx.getWideCharType(), "wchar_t"));
+ return ArgType::PtrTo(ArgType::WCStrTy);
+ case LengthModifier::AsUTF16:
+ return ArgType(ArgType::Char16Ty);
+ case LengthModifier::AsUTF32:
+ return ArgType(ArgType::Char32Ty);
case LengthModifier::AsAllocate:
case LengthModifier::AsMAllocate:
return ArgType::PtrTo(ArgType::CStrTy);
@@ -347,13 +355,17 @@
default:
return ArgType::Invalid();
}
- case ConversionSpecifier::CArg:
+ case ConversionSpecifier::sArg:
case ConversionSpecifier::SArg:
// FIXME: Mac OS X specific?
switch (LM.getKind()) {
case LengthModifier::None:
case LengthModifier::AsWide:
- return ArgType::PtrTo(ArgType(Ctx.getWideCharType(), "wchar_t"));
+ return ArgType::PtrTo(ArgType(Ctx.getWideCharType(), "wchar_t *"));
+ case LengthModifier::AsUTF16:
+ return ArgType::PtrTo(ArgType(Ctx.getChar16Type(), "char16_t *"));
+ case LengthModifier::AsUTF32:
+ return ArgType::PtrTo(ArgType(Ctx.getChar32Type(), "char32_t *"));
case LengthModifier::AsAllocate:
case LengthModifier::AsMAllocate:
return ArgType::PtrTo(ArgType(ArgType::WCStrTy, "wchar_t *"));
@@ -398,6 +410,8 @@
case LengthModifier::AsInt32:
case LengthModifier::AsInt3264:
case LengthModifier::AsWide:
+ case LengthModifier::AsUTF16:
+ case LengthModifier::AsUTF32:
case LengthModifier::AsShortLong:
return ArgType::Invalid();
}
@@ -435,10 +449,14 @@
return false;
// Pointer to a character.
- if (PT->isAnyCharacterType()) {
+ if (PT->isAnyCharacterType(LangOpt)) {
CS.setKind(ConversionSpecifier::sArg);
if (PT->isWideCharType())
LM.setKind(LengthModifier::AsWideChar);
+ else if (PT->isChar16Type(LangOpt))
+ LM.setKind(LengthModifier::AsUTF16);
+ else if (PT->isChar32Type(LangOpt))
+ LM.setKind(LengthModifier::AsUTF32);
else
LM.setKind(LengthModifier::None);
Index: clang/lib/AST/PrintfFormatString.cpp
===================================================================
--- clang/lib/AST/PrintfFormatString.cpp
+++ clang/lib/AST/PrintfFormatString.cpp
@@ -494,6 +494,10 @@
case LengthModifier::AsLong:
case LengthModifier::AsWide:
return ArgType(ArgType::WIntTy, "wint_t");
+ case LengthModifier::AsUTF16:
+ return ArgType(ArgType::Char16Ty, "char16_t");
+ case LengthModifier::AsUTF32:
+ return ArgType(ArgType::Char32Ty, "char32_t");
case LengthModifier::AsShort:
if (Ctx.getTargetInfo().getTriple().isOSMSVCRT())
return Ctx.IntTy;
@@ -535,6 +539,8 @@
case LengthModifier::AsAllocate:
case LengthModifier::AsMAllocate:
case LengthModifier::AsWide:
+ case LengthModifier::AsUTF16:
+ case LengthModifier::AsUTF32:
return ArgType::Invalid();
}
@@ -567,6 +573,8 @@
case LengthModifier::AsPtrDiff:
return ArgType::makePtrdiffT(
ArgType(Ctx.getUnsignedPointerDiffType(), "unsigned ptrdiff_t"));
+ case LengthModifier::AsUTF16:
+ case LengthModifier::AsUTF32:
case LengthModifier::AsAllocate:
case LengthModifier::AsMAllocate:
case LengthModifier::AsWide:
@@ -618,6 +626,8 @@
case LengthModifier::AsInt3264:
case LengthModifier::AsInt64:
case LengthModifier::AsWide:
+ case LengthModifier::AsUTF16:
+ case LengthModifier::AsUTF32:
return ArgType::Invalid();
case LengthModifier::AsShortLong:
llvm_unreachable("only used for OpenCL which doesn not handle nArg");
@@ -632,9 +642,15 @@
"const unichar *");
return ArgType(ArgType::WCStrTy, "wchar_t *");
}
- if (LM.getKind() == LengthModifier::AsWide)
+ if (LM.getKind() == LengthModifier::AsWide) {
return ArgType(ArgType::WCStrTy, "wchar_t *");
- return ArgType::CStrTy;
+ }
+ if (LM.getKind() == LengthModifier::AsUTF16)
+ return ArgType(ArgType::Char16Ty, "char16_t *");
+ if (LM.getKind() == LengthModifier::AsUTF32)
+ return ArgType(ArgType::Char32Ty, "char32_t *");
+ else
+ return ArgType::CStrTy;
case ConversionSpecifier::SArg:
if (IsObjCLiteral)
return ArgType(Ctx.getPointerType(Ctx.UnsignedShortTy.withConst()),
@@ -642,13 +658,22 @@
if (Ctx.getTargetInfo().getTriple().isOSMSVCRT() &&
LM.getKind() == LengthModifier::AsShort)
return ArgType::CStrTy;
+ if (LM.getKind() == LengthModifier::AsUTF16)
+ return ArgType(ArgType::Char16Ty, "char16_t *");
+ if (LM.getKind() == LengthModifier::AsUTF32)
+ return ArgType(ArgType::Char32Ty, "char32_t *");
return ArgType(ArgType::WCStrTy, "wchar_t *");
+ case ConversionSpecifier::cArg:
case ConversionSpecifier::CArg:
if (IsObjCLiteral)
return ArgType(Ctx.UnsignedShortTy, "unichar");
if (Ctx.getTargetInfo().getTriple().isOSMSVCRT() &&
LM.getKind() == LengthModifier::AsShort)
return Ctx.IntTy;
+ if (LM.getKind() == LengthModifier::AsUTF16)
+ return ArgType(ArgType::Char16Ty, "char16_t");
+ if (LM.getKind() == LengthModifier::AsUTF32)
+ return ArgType(ArgType::Char32Ty, "char32_t");
return ArgType(Ctx.WideCharTy, "wchar_t");
case ConversionSpecifier::pArg:
case ConversionSpecifier::PArg:
@@ -706,17 +731,21 @@
return true;
}
- // Handle strings next (char *, wchar_t *)
- if (QT->isPointerType() && (QT->getPointeeType()->isAnyCharacterType())) {
+ // Handle strings next (char *, wchar_t *, char16_t *, char32_t *)
+ if (QT->isPointerType() && (QT->getPointeeType()->isAnyCharacterType(LangOpt))) {
CS.setKind(ConversionSpecifier::sArg);
// Disable irrelevant flags
HasAlternativeForm = 0;
HasLeadingZeroes = 0;
- // Set the long length modifier for wide characters
+ // Set the length modifier for characters
if (QT->getPointeeType()->isWideCharType())
LM.setKind(LengthModifier::AsWideChar);
+ else if (QT->getPointeeType()->isChar16Type(LangOpt))
+ LM.setKind(LengthModifier::AsUTF16);
+ else if (QT->getPointeeType()->isChar32Type(LangOpt))
+ LM.setKind(LengthModifier::AsUTF32);
else
LM.setKind(LengthModifier::None);
@@ -736,19 +765,10 @@
VectorNumElts = OptionalAmount(VT->getNumElements());
}
}
-
- // We can only work with builtin types.
- if (!BT)
- return false;
-
+
// Set length modifier
switch (BT->getKind()) {
case BuiltinType::Bool:
- case BuiltinType::WChar_U:
- case BuiltinType::WChar_S:
- case BuiltinType::Char8: // FIXME: Treat like 'char'?
- case BuiltinType::Char16:
- case BuiltinType::Char32:
case BuiltinType::UInt128:
case BuiltinType::Int128:
case BuiltinType::Half:
@@ -819,6 +839,7 @@
case BuiltinType::UChar:
case BuiltinType::Char_S:
case BuiltinType::SChar:
+ case BuiltinType::Char8:
LM.setKind(LengthModifier::AsChar);
break;
@@ -840,6 +861,19 @@
case BuiltinType::LongDouble:
LM.setKind(LengthModifier::AsLongDouble);
break;
+
+ case BuiltinType::Char16:
+ LM.setKind(LengthModifier::AsUTF16);
+ break;
+
+ case BuiltinType::Char32:
+ LM.setKind(LengthModifier::AsUTF32);
+ break;
+
+ case BuiltinType::WChar_S:
+ case BuiltinType::WChar_U:
+ LM.setKind(LengthModifier::AsWide);
+ break;
}
// Handle size_t, ptrdiff_t, etc. that have dedicated length modifiers in C99.
@@ -873,7 +907,7 @@
// Set conversion specifier and disable any flags which do not apply to it.
// Let typedefs to char fall through to int, as %c is silly for uint8_t.
- if (!isa<TypedefType>(QT) && QT->isCharType()) {
+ if (!isa<TypedefType>(QT) && QT->isAnyCharacterType(LangOpt)) {
CS.setKind(ConversionSpecifier::cArg);
LM.setKind(LengthModifier::None);
Precision.setHowSpecified(OptionalAmount::NotSpecified);
Index: clang/lib/AST/FormatString.cpp
===================================================================
--- clang/lib/AST/FormatString.cpp
+++ clang/lib/AST/FormatString.cpp
@@ -232,7 +232,17 @@
break;
case 'l':
++I;
- if (I != E && *I == 'l') {
+ if (I + 1 != E && I[0] == '1' && I[1] == '6') {
+ ++I;
+ ++I;
+ lmKind = LengthModifier::AsUTF16;
+ break;
+ } else if (I + 1 != E && I[0] == '3' && I[1] == '2') {
+ ++I;
+ ++I;
+ lmKind = LengthModifier::AsUTF32;
+ break;
+ } else if (I + 1 != E && I[0] == 'l') {
++I;
lmKind = LengthModifier::AsLongLong;
} else {
@@ -459,6 +469,24 @@
return WInt == PromoArg ? Match : NoMatch;
}
+
+ case Char16Ty: {
+ const PointerType *PT = argTy->getAs<PointerType>();
+ if (!PT)
+ return NoMatch;
+ QualType pointeeTy =
+ C.getCanonicalType(PT->getPointeeType()).getUnqualifiedType();
+ return pointeeTy == C.getChar16Type() ? Match : NoMatch;
+ }
+
+ case Char32Ty: {
+ const PointerType *PT = argTy->getAs<PointerType>();
+ if (!PT)
+ return NoMatch;
+ QualType pointeeTy =
+ C.getCanonicalType(PT->getPointeeType()).getUnqualifiedType();
+ return pointeeTy == C.getChar32Type() ? Match : NoMatch;
+ }
case CPointerTy:
if (argTy->isVoidPointerType()) {
@@ -520,6 +548,12 @@
case WCStrTy:
Res = C.getPointerType(C.getWideCharType());
break;
+ case Char16Ty:
+ Res = C.getPointerType(C.getChar16Type());
+ break;
+ case Char32Ty:
+ Res = C.getPointerType(C.getChar32Type());
+ break;
case ObjCPointerTy:
Res = C.ObjCBuiltinIdTy;
break;
@@ -607,6 +641,10 @@
return "m";
case AsWide:
return "w";
+ case AsUTF16:
+ return "l16";
+ case AsUTF32:
+ return "l32";
case None:
return "";
}
@@ -860,6 +898,17 @@
default:
return false;
}
+ case LengthModifier::AsUTF16:
+ case LengthModifier::AsUTF32:
+ switch (CS.getKind()) {
+ case ConversionSpecifier::cArg:
+ case ConversionSpecifier::CArg:
+ case ConversionSpecifier::sArg:
+ case ConversionSpecifier::SArg:
+ return true;
+ default:
+ return false;
+ }
case LengthModifier::AsWide:
switch (CS.getKind()) {
case ConversionSpecifier::cArg:
@@ -886,6 +935,9 @@
case LengthModifier::AsSizeT:
case LengthModifier::AsPtrDiff:
case LengthModifier::AsLongDouble:
+ case LengthModifier::AsWide:
+ case LengthModifier::AsUTF16:
+ case LengthModifier::AsUTF32:
return true;
case LengthModifier::AsAllocate:
case LengthModifier::AsMAllocate:
@@ -893,7 +945,6 @@
case LengthModifier::AsInt32:
case LengthModifier::AsInt3264:
case LengthModifier::AsInt64:
- case LengthModifier::AsWide:
case LengthModifier::AsShortLong: // ???
return false;
}
@@ -997,6 +1048,12 @@
} else if (Identifier->getName() == "ptrdiff_t") {
LM.setKind(LengthModifier::AsPtrDiff);
return true;
+ } else if (Identifier->getName() == "char16_t") {
+ LM.setKind(LengthModifier::AsUTF16);
+ return true;
+ } else if (Identifier->getName() == "char32_t") {
+ LM.setKind(LengthModifier::AsUTF32);
+ return true;
}
QualType T = Typedef->getUnderlyingType();
Index: clang/include/clang/AST/FormatString.h
===================================================================
--- clang/include/clang/AST/FormatString.h
+++ clang/include/clang/AST/FormatString.h
@@ -80,6 +80,8 @@
AsLongDouble, // 'L'
AsAllocate, // for '%as', GNU extension to C90 scanf
AsMAllocate, // for '%ms', GNU extension to scanf
+ AsUTF16, // for '%l16(c|s)', Clang extension
+ AsUTF32, // for '%l32(c|s)', Clang extension
AsWide, // 'w' (MSVCRT, like l but only for c, C, s, S, or Z
AsWideChar = AsLong // for '%ls', only makes sense for printf
};
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
