Path: csiph.com!news.swapon.de!eternal-september.org!feeder.eternal-september.org!reader01.eternal-september.org!.POSTED!not-for-mail
From: Keith Thompson <kst-u@mib.org>
Newsgroups: comp.lang.c
Subject: Re: "...the value becomes,indeterminate each time the declaration is reached"
Date: Tue, 29 Jan 2019 16:36:41 -0800
Organization: None to speak of
Lines: 46
Message-ID: <lno97zvv52.fsf@kst-u.example.com>
References: <q2q79p$1j7$1@dont-email.me> <q2q9l2$h95$1@dont-email.me> <q2ql6g$tk7$1@dont-email.me>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
Injection-Info: reader02.eternal-september.org; posting-host="b08549a0708368ce183a470765040d39"; logging-data="24969"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX19jWEQSQiY4Ln91UEKPJq5T"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.1 (gnu/linux)
Cancel-Lock: sha1:FfqVNb6NHlBDmk9eWNoi/fqaLO4= sha1:RSCAe8GA/QoeljIPhlVXTSpM4+E=
Xref: csiph.com comp.lang.c:140402

Andrey Tarasevich <andreytarasevich@hotmail.com> writes:
> On 1/29/2019 11:30 AM, David Brown wrote:
>> Why would you want i to keep its value?  If you choose to write 
>> spaghetti code and jump around like this, the compiler is free to do 
>> whatever it likes with the value of "i" the second time around 
>> (including keeping the value).  
>
> The question is "why?". The code "jumps around" well within the lifetime 
> of the object in question. Why is the object not guaranteed to keep its 
> value at that specific spot of its lifetime? What is the reason this 
> potential possibility for loss of value was introduced in C99? It wasn't 
> there in C89/90 (unless it was a defect, of course).

Here's one possible rationale.

An uninitialized local object has an indeterminate value until a value
has been assigned to it.

    {
        int x;
        int y = 42;
    }

A compiler could treat `int x;` as `int x = SOME_WEIRD_VALUE;` to aid in
debugging.  It might store some consistent bit pattern that's unlikely
to appear accidentally.  (I think I've seen a repeating 0xe5 pattern used
this way).

    {
        int x /* = (int)0xe5e5e5e5 */;
        int y = 42;
    }

If you write code that uses a goto to transfer control to a point
within the lifetime of x but before its declaration (yes, my wording
is a little sloppy), it would not be reasonable to require such a
compiler to treat this as a special case and omit the initialization
when the declaration is reached a second time.

I don't know whether the authors of the standard had this kind of
scenario in mind.

-- 
Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
Will write code for food.
void Void(void) { Void(); } /* The recursive call of the void */