Path: csiph.com!eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: Keith Thompson <Keith.S.Thompson+u@gmail.com>
Newsgroups: comp.lang.c++
Subject: Re: Has "stack overflow" specified behavior?
Date: Mon, 13 Dec 2021 15:58:41 -0800
Organization: None to speak of
Lines: 52
Message-ID: <8735mw9hvy.fsf@nosuchdomain.example.com>
References: <f5624610-cbe0-4e27-9f52-f4a900172c89n@googlegroups.com> <877dc89i1n.fsf@nosuchdomain.example.com>
Mime-Version: 1.0
Content-Type: text/plain
Injection-Info: reader02.eternal-september.org; posting-host="abb7162027a59d3701de80734b2d80ee"; logging-data="7453"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX18n8VFwgXqsKp78l7RIKNob"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/27.2 (gnu/linux)
Cancel-Lock: sha1:Q9yjR15YXgdZnz858nug6c42Fpg= sha1:TKhWs4NaHf8nutgIy2XwA0vN0hc=
Xref: csiph.com comp.lang.c++:82617

Keith Thompson <Keith.S.Thompson+u@gmail.com> writes:
> wij <wyniijj@gmail.com> writes:
>> void t() {
>>   int a;
>>   ++a;
>>   t();
>> };
>>
>> int main()
>> {
>>  t();
>> }
>>
>> ---
>> Has "stack overflow" specified behavior?
>
> The standard does not specify what happens when a resource limit is
> exceeded.  In my opinion that matches the standard's definition of
> "undefined behavior" (behavior that is not defined).
>
> Is the variable `a` intended to track the depth of the recursion?  It
> doesn't.  A new instance of `a` is created every time t is called.  You
> didn't initialize `a`, but if you initialized it to 0 then `++a` would
> simply set it to 1.
>
> If you made `a` static, it would count the depth of the recursion -- and
> you'd have undefined behavior after the value of `a` reaches INT_MAX.
>
> With `a` defined as you did here, there's a distinct instance of `a` for
> each call to t, and each instance has its own distinct address, a value
> of type int*.  There can be at most 2**(CHAR_BIT * sizeof (int*))
> distinct int* values.  If you never refer to the value or address of `a`,
> an optimizing compiler is likely to eliminate it and change the
> recursive call to a loop, but that doesn't apply in the abstract
> machine.  See N1570 5.1.2.3 "Program execution":
>
>     The semantic descriptions in this International Standard describe
>     the behavior of an abstract machine in which issues of optimization
>     are irrelevant.
>
> In the abstract machine, each instance of `a` occupies sizeof (int) bytes
> and has a unique address of type int*.  After optimization, `a` might not
> exist.

My apologiess, I didn't notice which newsgroup I was in and gave a C
answer.  But the C++ standard also discusses the "abstract machine" in
section 4.6 [intro.execution], with a very similar meaning.

-- 
Keith Thompson (The_Other_Keith) Keith.S.Thompson+u@gmail.com
Working, but not speaking, for Philips
void Void(void) { Void(); } /* The recursive call of the void */