0.96: build fails about ledmon HOT 11 CLOSED

Sure, that is a good point- it is a common standard now.
To fix that we need to:

• add verification to our CI.
• resolve existing issues.

It is not critical, so I will create appropriative internal task, hopefully we can resolve that until next release.
Reopening then.

from ledmon.

Could you provide me versions of gcc and make packages?

from ledmon.

Latest gcc 12.1.1.

from ledmon.

Thanks. In attached log I see that these errors are expected and coms from additional flag "-Werror=format-truncation". We use buffer with PATH_MAX size, which is the same as components size. In case that all buffer will be written, output will be truncated, which is expected.
I suggest you to turn off this flag to avoid these errors.

from ledmon.

Please do not swipe that under the carpet.

from ledmon.

Linux supports name with a size up to 4096, which is PATH_MAX. That size shall be our max length.
If we are concatenating many paths like in those cases, we accept the risk of buffer truncation. I believe that issues cannot occur real live, especially that we are scanning sysfs, interface created by kernel.
I don't see any advantage of creating dedicated and precisely calculated memory for each buffer. For me it is more bug prone and makes code less understandable.

For me it is not an issue but if it is important to you we will try to meet your requirements. Please elaborate why this flag is important for you.

from ledmon.

No feedback, closing.
Fell free to reopen the issue if necessary.

from ledmon.

Please do not swipe that under the carpet.

I've provided feedback #102 (comment)

from ledmon.

Simple I don't see any reason why only ledmon shoud be only exception about not pasing compile with
-Werror=format-truncation. I have now cleaned +4.2k packages. All are correctly build with -Werror=format-truncation.

from ledmon.

Thank you 👍

from ledmon.

Closing, the flag is in configure:
https://github.com/intel/ledmon/blob/master/configure.ac#L64

from ledmon.