Revisiting old tickets

[CHR]

So I've been looking at some old tickets marked as false positives. All examples were run with --enable=all --inconclusive. At least five tickets no longer reproduce with the current version.

Ticket#	Repro 1.87	Repro 2.3	Comment
6081	no	no
6259	yes	yes
6342	yes	no
6477	yes	yes	can be avoided by setting ptr = 0 after delete
6552	yes	yes	wacky overloaded >>= with side effects
6572	yes	yes
6809	no	no
7013	yes	no
7334	yes	yes
7366	yes	yes
7459	yes	yes
7622	yes	yes	arguably not a FP
7739	yes	yes
7754	yes	yes
7768	yes	yes
7948	yes	yes
7954	yes	yes
7959	yes	yes	arguably not a FP
8000	yes	no

[Daniel Marjamäki]

6342: the "analysis failed" is not reported anymore however the debug output is not good.

[Daniel Marjamäki]

7013: thanks! I closed that.

[Daniel Marjamäki]

8000: thanks! I closed that

[Daniel Marjamäki]

6809: hmm.. amai said he could reproduce on windows but not on linux. did you try it on windows also?

[CHR]

6809: Here is what I get on Windows 64bit, which seems correct. The original complaint was about requires 'ssize_t' but the argument type is 'ssize_t'.
Code:

void main(void) {
    ssize_t n = -1;
    printf("%zd", n); // should be allowed, like zu for size_t, gcc accepts it but cppcheck doesn't
    printf("%ld", n); // acceptable to both cppcheck and gcc
}

#include <stdio.h>
int main2(void) {
    ssize_t n = -1;
    size_t x = 2;
    printf("%ld \n", n);
    printf("%zd \n", n);

    printf("%lu \n", x);
    printf("%zu \n", x);
    return 0;
}

Output:

Checking 6809.cpp ...
[6809.cpp:4]: (portability) %ld in format string (no. 1) requires 'long' but the argument type is 'ssize_t {aka signed long long}'.
[6809.cpp:11]: (portability) %ld in format string (no. 1) requires 'long' but the argument type is 'ssize_t {aka signed long long}'.
[6809.cpp:14]: (portability) %lu in format string (no. 1) requires 'unsigned long' but the argument type is 'size_t {aka unsigned long long}'.
[6809.cpp:8]: (style) The function 'main2' is never used.
(information) Cppcheck cannot find all the include files (use --check-config for details)

[Daniel Marjamäki]

thanks! I have closed 6809

[CHR]

Tickets 8215, 8805 and 8909 also don't reproduce for me with either v1.87 or v2.3.

[Daniel Marjamäki]

8215: I can reproduce if file extension is .c . I have the feeling that the evaluation order is defined since c++11.

[Daniel Marjamäki]

thanks! I can reproduce 8909 when file extension is .c. Not sure why that makes a difference.

[CHR]

Hmm, I must have overlooked the file extension on some of the code snippets.

[CHR]

8805 has been closed by @amai.

[CHR]

Tickets 8094, 8227, 8526, 8942, 8961, 9026 and 9223 reproduce with v1.87, but not with v2.3.
Edited for typo: 8904 -> 8094

[Daniel Marjamäki]

thanks! I closed 9223

[Daniel Marjamäki]

thanks! I closed 8961

[CHR]

It seems I was wrong about 8227, it still reproduces after all.

[Daniel Marjamäki]

I closed 9026

[CHR]

9937 no longer reproduces with 2.3.

[CHR]

Just a reminder: 9937, 8094, 8526, and 6081 are still open.

[Daniel Marjamäki]

I closed 8094

[Daniel Marjamäki]

I closed 8526

[Daniel Marjamäki]

6081 was closed by amai

[Daniel Marjamäki]

Thanks I have closed 9937

[CHR]

Thanks for your efforts :-)

[CHR]

I just noticed that 8742 no longer reproduces with head (at least the original example, the modified one still does).