[cfe-commits] r68975 - in /cfe/trunk: lib/CodeGen/CodeGenModule.cpp test/CodeGen/illegal-UTF8.m
Steve Naroff
snaroff at apple.com
Mon Apr 13 12:08:09 PDT 2009
Author: snaroff
Date: Mon Apr 13 14:08:08 2009
New Revision: 68975
URL: http://llvm.org/viewvc/llvm-project?rev=68975&view=rev
Log:
Fixed crasher in <rdar://problem/6780904> [irgen] Assertion failed: (Result == conversionOK && "UTF-8 to UTF-16 conversion failed"), function GetAddrOfConstantCFString, file CodeGenModule.cpp, line 1063.
Still a diagnostic related FIXME (will discuss with Daniel/Fariborz offline).
Added:
cfe/trunk/test/CodeGen/illegal-UTF8.m
Modified:
cfe/trunk/lib/CodeGen/CodeGenModule.cpp
Modified: cfe/trunk/lib/CodeGen/CodeGenModule.cpp
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CodeGenModule.cpp?rev=68975&r1=68974&r2=68975&view=diff
==============================================================================
--- cfe/trunk/lib/CodeGen/CodeGenModule.cpp (original)
+++ cfe/trunk/lib/CodeGen/CodeGenModule.cpp Mon Apr 13 14:08:08 2009
@@ -1060,16 +1060,25 @@
Result = ConvertUTF8toUTF16(&FromPtr, FromPtr+Literal->getByteLength(),
&ToPtr, ToPtr+Literal->getByteLength(),
strictConversion);
- assert(Result == conversionOK && "UTF-8 to UTF-16 conversion failed");
+ if (Result == conversionOK) {
+ // FIXME: Storing UTF-16 in a C string is a hack to test Unicode strings
+ // without doing more surgery to this routine. Since we aren't explicitly
+ // checking for endianness here, it's also a bug (when generating code for
+ // a target that doesn't match the host endianness). Modeling this as an
+ // i16 array is likely the cleanest solution.
+ StringLength = ToPtr-&ToBuf[0];
+ str.assign((char *)&ToBuf[0], StringLength*2);// Twice as many UTF8 chars.
+ isUTF16 = true;
+ } else if (Result == sourceIllegal) {
+ // FIXME: GCC currently emits the following warning (in the backend):
+ // "warning: input conversion stopped due to an input byte that does not
+ // belong to the input codeset UTF-8"
+ // The clang backend doesn't currently emit any warnings.
+ str.assign(Literal->getStrData(), Literal->getByteLength());
+ StringLength = str.length();
+ } else
+ assert(Result == conversionOK && "UTF-8 to UTF-16 conversion failed");
- // FIXME: Storing UTF-16 in a C string is a hack to test Unicode strings
- // without doing more surgery to this routine. Since we aren't explicitly
- // checking for endianness here, it's also a bug (when generating code for
- // a target that doesn't match the host endianness). Modeling this as an i16
- // array is likely the cleanest solution.
- StringLength = ToPtr-&ToBuf[0];
- str.assign((char *)&ToBuf[0], StringLength*2); // Twice as many UTF8 chars.
- isUTF16 = true;
} else {
str.assign(Literal->getStrData(), Literal->getByteLength());
StringLength = str.length();
Added: cfe/trunk/test/CodeGen/illegal-UTF8.m
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/test/CodeGen/illegal-UTF8.m?rev=68975&view=auto
==============================================================================
--- cfe/trunk/test/CodeGen/illegal-UTF8.m (added)
+++ cfe/trunk/test/CodeGen/illegal-UTF8.m Mon Apr 13 14:08:08 2009
@@ -0,0 +1,8 @@
+// RUN: clang %s -S -m64
+
+ at class NSString;
+
+// FIXME: GCC emits the following warning:
+// CodeGen/illegal-UTF8.m:4: warning: input conversion stopped due to an input byte that does not belong to the input codeset UTF-8
+
+NSString *S = @"\xff\xff___WAIT___";
More information about the cfe-commits
mailing list