[llvm-bugs] [Bug 25325] New: Inaccurate warning message with -Wabsolute-value

[via llvm-bugs]

https://llvm.org/bugs/show_bug.cgi?id=25325

            Bug ID: 25325
           Summary: Inaccurate warning message with -Wabsolute-value
           Product: clang
           Version: trunk
          Hardware: All
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Frontend
          Assignee: unassignedclangbugs at nondot.org
          Reporter: Yunzhong_Gao at playstation.sony.com
                CC: llvm-bugs at lists.llvm.org, rtrieu at google.com
    Classification: Unclassified

Hi,
We encounter this issue during internal documentation review.

// test.cpp
#include <cstdlib>
#include <cstdio>
int foo(unsigned u) { return abs(u); }
int main() {
  printf("%d\n", foo(0xFFFFFFFFu)); // abs(0xffffffff) => abs(-1) => 1
  return 0;
}
// warning: taking the absolute value of unsigned type 'unsigned int' has no
// effect [-Wabsolute-value]


This is a useful warning message, but in this case, it is not accurate to say
that the abs() call has no effect.

On the other hand, in many cases the abs() does indeed have no effect, like
below, so I find it tricky to come up with a warning message that covers all
cases. I am hoping that someone else is able to suggest a better warning
message.

e.g. in the following two cases, abs() and labs() have no effect:
  int f1(unsigned short x) { return abs(x); }
  long f2(unsigned x) { return labs(x); }



Some alternative texts that we have internally discussed:
(a)
warning: taking the absolute value of unsigned type 'unsigned int' may have an
unintended effect.
[But, it sounds vague if we do not describe what is the unintended effect.]

(b)
warning: taking the absolute value of unsigned type 'unsigned int' has an
effect such that an unsigned number gets implicitly converted to its
corresponding signed integral type before its absolute value is taken. This
conversion could turn a large unsigned number into a negative number.
[But, it sounds verbose and does not mention the "no effect" case.]

-- 
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/20151026/35f51a74/attachment-0001.html>