Cppcheck-2.5

[Georgiy Komarov]

I also checked the official MISRA forum and found two related threads:
https://www.misra.org.uk/forum/viewtopic.php?f=223&t=1288
https://www.misra.org.uk/forum/viewtopic.php?f=223&t=1669
These cases are incompatible. But they don't explain why.

[Richard Smith]

Yeah.. I don't understand why those are not ok. Seems reasonable to me that the checker could not consider those violations. The end result is the same.

[Richard Smith]

[CHR]

Using release-windows #431, I'm getting apparently random/hard to reproduce crashes. I realize this is not very specific, but maybe someone else is also seeing this?

[Daniel Marjamäki]

hmm.. the daca@home show that there are some crashes but it has not escalated a lot. but if you can provide a example code that is very interesting..

[Daniel Marjamäki]

For information.. I have started the release process and would like to release tomorrow..

feel free to complain if you think there are still blockers that means we should delay the release further..

[Absol]

Someone can add FLT_MIN DBL_MIN detection ?

https://github.com/danmar/cppcheck/pull/3393#issuecomment-899035687

[Daniel Marjamäki]

Can you please post all the details here in the forum.. I created this discussion topic:
https://sourceforge.net/p/cppcheck/discussion/development/thread/7d0052a398/

A small example code that we should warn about is wanted.

[Absol]

include <iostream></iostream>

void f(int * main)
{
int test[1];
test[0] = main[5];
std::cout << test[0];
}

int main()
{
int * test2 = new int[1];
test2[0] = 5;
f(test2);
}

No error about main[5] is uninitialized memory.
And this is not cause error in 90% cases. http://cpp.sh/2xq7t

[Daniel Marjamäki]

It is not uninitialized memory. It is buffer overrun. This is a regression. We used to detect this. cppcheck-1.87 says:

[1.cpp:12] -> [1.cpp:4]: (error) Array '[1]' accessed at index 5, which is out of bounds.

not sure where it was lost but it was before cppcheck-2.2

[Daniel Marjamäki]

Thanks for reporting this! I have created https://trac.cppcheck.net/ticket/10415

[Absol]

Thanks!

[Absol]

clear

[Absol]

Someone can add detection for this error ?

[Absol]

I founs new problem or this is regression too?

#include <iostream>

int main()
{
    char *p = new char[256];
    p = &p[255];
    p[0] = 'x';
    p[10] = 'x';
    std::cout << p;
}

[Absol]

this is compiled and executed without error, but cause error?)))

[Absol]

impossible to fix it and add new detection ?

[Daniel Marjamäki]

sorry .. I have the feeling there could be a related ticket already and wanted to look but then didn't have time to look. I create a new ticket anyway: https://trac.cppcheck.net/ticket/10422

[Absol]

I know it exists https://trac.cppcheck.net/ticket/10415. I writing about new problem

#include <iostream>

int main()
{
    char *p = new char[256];
    p = &p[255];
    p[0] = 'x';
    p[10] = 'x';
    std::cout << p;
}

It not detected as code mistake

[Daniel Marjamäki]

could you open a new forum. I don't see this discussion as related to the cppcheck-2.5 release. but I took that code and simplified it and created https://trac.cppcheck.net/ticket/10422 so there is a ticket already.

[Absol]

yes I write about this problem, thanks