| Issue |
135050
|
| Summary |
On x86-64 trunk, `-Wconditional-uninitialized` false positive in a "simple" case
|
| Labels |
false-positive
|
| Assignees |
|
| Reporter |
blanefrye
|
`-Wconditional-uninitialized` claims `c` may be uninitialized when used in `if (c)`, but if it's used `b` must be true, meaning `b` must not be false, meaning `c` must have been assigned `false`.
https://godbolt.org/z/PzdeMq84M
```
void test(bool a)
{
bool b = true;
bool c;
if (a)
{
c = false;
}
else
{
b = false;
}
if (b)
{
if (c)
{
}
}
}
```
I see as of 2018, "-Wconditional-uninitialized is expected to have false positives: it warns in cases where a simple control flow analysis cannot prove that a variable was written to prior to each use." but I'm not sure what qualifies as "simple". https://bugs.llvm.org/show_bug.cgi?id=38856
![]()
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
