https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94626
--- Comment #3 from Allison Karlitskaya <allison.karlitskaya at redhat dot com> --- (In reply to Martin Sebor from comment #2) > Thanks for the report (and for getting Glibc to update its headers). The > problem with strncpy is that its correct and intended uses (filling a buffer > without necessarily nul-terminating it) are indistinguishable from the > misuses (attempting to create a possibly truncated string as a copy of > another). > > Since there are more misuses than the correct uses of the function, the > warning relies on authors of the correct code to annotate it with attribute > nonstring. This is mentioned in the documentation of the warning. The > warning itself doesn't mention this because the right solution is likely to > fix the copy. > > I realize this approach makes the minority of authors of correct/safe code > pay the price for the mistakes of the majority of others and so must be > frustrating if you're in the first group. Sorry. The opposite approach > obviously wouldn't be effective. > > If it helps, here's an article I wrote on this topic when the warning was > first added: > https://developers.redhat.com/blog/2018/05/24/detecting-string-truncation- > with-gcc-8 hi Martin, Thanks for the reply. I agree with everything you've written, and I wouldn't want to change the behaviour here. The only thing I'm asking is that a brief mention about "consider the nonstring attribute" be added to the warning text.
