On 2/3/26 3:42 PM, Jakub Jelinek wrote:
Hi!My P2246R1 patch caused diagnostics reported by running ubsan instrumented compiler on cpp26/static_assert1.C - if len is 0, we don't bother to allocate msg, so it stays NULL, and when I've added cpp_translate_string call, that can invoke memcpy (something, NULL, 0); in that case. While that is no longer UB in C2Y since N3322, libsanitizer doesn't know that yet and reports it anyway. While we could just do if (len) { ... } else msg = ""; there is really no point in trying to translate "" and allocate memory for that, so the following patch instead by passes that translation for len == 0. Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
OK.
2026-02-03 Jakub Jelinek <[email protected]> PR c++/123918 * semantics.cc (cexpr_str::extract): Bypass cpp_translate_string for len == 0. --- gcc/cp/semantics.cc.jj 2026-01-28 19:17:49.532306466 +0100 +++ gcc/cp/semantics.cc 2026-02-02 10:52:20.787120028 +0100 @@ -12860,7 +12860,10 @@ cexpr_str::extract (location_t location, cpp_string istr, ostr; istr.len = len; istr.text = (const unsigned char *) msg; - if (!cpp_translate_string (parse_in, &istr, &ostr, CPP_STRING, true)) + if (len == 0) + ; + else if (!cpp_translate_string (parse_in, &istr, &ostr, CPP_STRING, + true)) { error_at (location, "could not convert constexpr string from " "ordinary literal encoding to source character " Jakub
