[llvm-bugs] [Bug 30550] LLVM's Global Value Numbering pass bug
via llvm-bugs
llvm-bugs at lists.llvm.org
Sat Dec 31 03:26:25 PST 2016
https://llvm.org/bugs/show_bug.cgi?id=30550
Daniel Berlin <dberlin at dberlin.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution|--- |INVALID
--- Comment #2 from Daniel Berlin <dberlin at dberlin.org> ---
First, note:
/Users/dannyb/Downloads/test (1).cpp:38:23: warning: when type is in
parentheses, array cannot have dynamic
size
p1 = new (20) (char[i]);
~ ^ ~
/Users/dannyb/Downloads/test (1).cpp:39:23: warning: when type is in
parentheses, array cannot have dynamic
size
p2 = new (20) (char[i]);
~ ^ ~
2 warnings generated.
Clang's warning is correct here:
"With the parentheses, the type to be newed comes from a type-id, in this case
char[i]. This is a variable length array which is not standard behavior.
Without the parentheses, the type to be newed is a new-type-id, an array of X."
g++ will give you the same warning:
/Users/dannyb/Downloads/test (1).cpp: In function 'int tst()':
/Users/dannyb/Downloads/test (1).cpp:38:25: warning: non-constant array new
length must be specified without parentheses around the type-id [-Wvla]
p1 = new (20) (char[i]);
^
/Users/dannyb/Downloads/test (1).cpp:39:25: warning: non-constant array new
length must be specified without parentheses around the type-id [-Wvla]
p2 = new (20) (char[i]);
^
Second, these functions overwrite the heap (and thus, i'm not sure why you
believe they have to return 0).
Both g++ and clang with -fsanitize=address will show you this.
This is because you have:
int i = 0;
p1 = new (20) char[i];
Regardless of what your allocation function does, you've told it you've got a
dynamic array of size 0 :)
If you change the code to
p1 = new (20) char[20];
p2 = new (20) char[20];
or
int i = 20
p1 = new (20) char[i];
p2 = new (20) char[i];
It will succeed in all cases.
Marking as invalid for now.
--
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/20161231/18ee9b33/attachment-0001.html>
More information about the llvm-bugs
mailing list