Re: std::string add nullptr attribute

[Jonny Grant <jg@jguk.org>]

On 20/02/2023 19:35, Jonathan Wakely wrote:
> 
> 
> On Mon, 20 Feb 2023, 19:21 Jonny Grant, <jg@jguk.org <mailto:jg@jguk.org>> wrote:
> 
> 
> 
>     On 20/02/2023 13:44, Jonathan Wakely wrote:
>     > On Mon, 20 Feb 2023 at 12:59, Xi Ruoyao <xry111@xry111.site> wrote:
>     >>
>     >> On Mon, 2023-02-20 at 11:30 +0000, Jonny Grant wrote:
>     >>
>     >>> Thank you for the suggestion, I gave that nonnull attribute a try, but
>     >>> it doesn't appear to warn for this example.
>     >>>
>     >>> https://godbolt.org/z/boqTj6oWE <https://godbolt.org/z/boqTj6oWE>
>     >>
>     >> Ouch... The optimizer inlined make_std_string so both -Wnonnull and -
>     >> fanalyzer fails to catch the issue here.
>     >>
>     >> Adding noipa attribute for make_std_string will work, but will also
>     >> cause the generated code stupidly slow.  Maybe:
>     >>
>     >> #ifdef WANT_DIAGNOSTIC
>     >> #define MAKE_STD_STRING_ATTR __attribute__ ((noipa, nonnull))
>     >> #else
>     >> #define MAKE_STD_STRING_ATTR
>     >> #endif
>     >>
>     >> std::string make_std_string(const char * const str) MAKE_STD_STRING_ATTR;
>     >>
>     >> It still looks very stupid though.
>     >>
>     >>> Feels useful to get build warnings if compiler knows nullptr is going
>     >>> to be dereferenced, as clang does.
>     >>
>     >> The problem is in this case nullptr is not dereferenced, at all.  So if
>     >> we want a warning here we'll have to invent some new __builtin or
>     >> __attribute__ to give the compiler a hint.  AFAIK there is no such
>     >> facility now.
>     >
>     > I think __attribute__((access(read_only, 1))) should do it.
> 
>     I tried here, but no luck, is this example the way you mean? gcc trunk.
> 
> 
> I don't mean it does do it, I mean it should do it.

Ok I see. I filed an issue with a simpler test case.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108893

Kind regards
Jonny