[LLVMbugs] [Bug 21632] New: Assertion failure in CXXNameMangler::mangleFunctionParam with typeid() of VLA
bugzilla-daemon at llvm.org
bugzilla-daemon at llvm.org
Fri Nov 21 10:52:35 PST 2014
http://llvm.org/bugs/show_bug.cgi?id=21632
Bug ID: 21632
Summary: Assertion failure in
CXXNameMangler::mangleFunctionParam with typeid() of
VLA
Product: clang
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: normal
Priority: P
Component: C++
Assignee: unassignedclangbugs at nondot.org
Reporter: warren_ristow at playstation.sony.com
CC: dgregor at apple.com, llvmbugs at cs.uiuc.edu
Classification: Unclassified
This is a low-priority issue.
When compiling the following test-case:
#include <typeinfo>
#include <stdio.h>
void foo(int i) {
char bar[i];
printf("%s\n", typeid(bar).name());
return;
}
with a Release+Asserts build of Clang, an assertion-failure triggers. A
Release (without Asserts) version compiles it successfully (and without any
warnings). FTR, in the Release version, 'typeid(bar).name()' returns "Afp__c",
whereas if (for example) the literal 17 had been used for the array (as in
'char bar[17];'), then it returns "A17_c". I've tested this with a modern
compiler (r222492), but the behavior appears to have existed for a long time.
The assertion-failure is:
clang: llvm/tools/clang/lib/AST/ItaniumMangle.cpp:3292:
void (anonymous namespace)::CXXNameMangler::mangleFunctionParam(const
clang::ParmVarDecl *):
Assertion `parmDepth < FunctionTypeDepth.getDepth()' failed.
Note that the test-case contains ill-formed C++ code. Specifically, the
typeid() of a Variable-Length array type is ill-formed. For reference, the
link:
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2013/n3639.html
contains the following example, which explicitly has a typeid() of a VLA
indicated as being ill-formed:
void f(std::size_t n)
{
int a[n];
unsigned int x = sizeof(a); // ill-formed
const std::type_info& ti = typeid(a); // ill-formed
typedef int t[n]; // ill-formed
}
g++ considers this to be a user-error (tested with g++ version 4.8.2, but
probably is an error with all relevant versions of g++). Also tested with
other compilers, and it's considered a user-error. Likely, the Clang fix will
be to consider it a user-error, and then the ICE will be avoided.
--
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20141121/d9550fbd/attachment.html>
More information about the llvm-bugs
mailing list