Fix #8545 FN Condition '...' is always true ' uint16_t i; ( i <= 0xFFFF)'

[chrchr-github]

No description provided.

[chrchr-github]

Without the suppression, CMake test fails.
With the added suppression, runtests.sh complains about an unmatched suppression.
Are these checks using different platforms or something?

[firewave]

Interesting - I thought the format was cppcheck-suppress id1;id2.

[chrchr-github]

According to the manual: // cppcheck-suppress [aaaa, bbbb]

[firewave]

Without the suppression, CMake test fails.
With the added suppression, runtests.sh complains about an unmatched suppression.

In the same job or different ones? Maybe a FILESDIR issue?

Are these checks using different platforms or something?

The only case I am aware of is with a Windows binary - see https://trac.cppcheck.net/ticket/10850#comment:2.

[pfultz2]

I feel like this would be better in fixed in ValueFlow as this could enable inferring more values and improving non-style checks. In valueFlowImpossibleValues, we set out of range values to impossible, for example booleans are set to !<=-1 and !>=-2 here:

https://github.com/danmar/cppcheck/blob/main/lib/valueflow.cpp#L1650

We could add the impossible values for integers smaller than 32 bits, but several unit tests for valueflow may need to be updated to use removeImpossible.

[chrchr-github]

There also is some overlap with unsignedLessThanZero and unsignedPositive.