From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id 957EA3858D32; Thu,  2 Nov 2023 08:47:24 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 957EA3858D32
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1698914844;
	bh=Qqt5wJNwlxzFlDeMTcarHveGCp34OEhS2K5JLLEiDOI=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=lhDuOA9dm2xOrOMlC/T398qD/WEspIw2/NNGMfRk18xsUbjqIRiIN7rlYoblms6s7
	 Ygi38cxIhgttx/WFqz/BPB05Uf+lU4Io2L2QEorbg8EVaFkHAvazyrJ9sLHyLFvEvl
	 KxLZQSGlL6FCiKWAebj0on200oLAtZWcUOGxHiPo=
From: "jdapena at igalia dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug libstdc++/112314] Missing index assertions in basic_string_view
Date: Thu, 02 Nov 2023 08:47:23 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: libstdc++
X-Bugzilla-Version: 12.3.0
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: enhancement
X-Bugzilla-Who: jdapena at igalia dot com
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-112314-4-WVga7KXJqs@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-112314-4@http.gcc.gnu.org/bugzilla/>
References: <bug-112314-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D112314

--- Comment #2 from Jose Dapena Paz <jdapena at igalia dot com> ---
(In reply to Jonathan Wakely from comment #1)
> For the second one, I did start work on a patch that attempts to verify t=
hat
> the provided length is not more than __builtin_object_size. I can't remem=
ber
> why I stopped working on that. There's no way to make that check 100%
> reliable, because __builtin_object_size depends on optimizations and whet=
her
> the compiler can see enough data.

I guess in this case a "best-effort" approach could be good enough? Providi=
ng
there are no false positives asserting for actually valid calls. In the cas=
e of
Chromium unit test it is passing a predefined const char[] so, at least mak=
ing
sure those cases are validated would add an extra layer of protection.

In any case, the failing test is actually passing -1, my understanding is t=
hat
that one should always assert no matter what we are passing as const char*.=