Path: csiph.com!eternal-september.org!feeder.eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: Keith Thompson <kst-u@mib.org>
Newsgroups: comp.lang.c
Subject: Re: Global variables
Date: Mon, 07 May 2018 08:49:02 -0700
Organization: None to speak of
Lines: 40
Message-ID: <ln1senbh5d.fsf@kst-u.example.com>
References: <pc5tpv$1o5f$1@gioia.aioe.org> <fks3f3Ffpt3U1@mid.individual.net> <tu5GC.34633$ga6.7485@fx24.am4> <qDjGC.43146$o43.39682@fx33.iad> <SslGC.160031$Cd3.94316@fx33.am4> <fkumo1Fl1fsU1@mid.individual.net> <khqGC.223876$5v7.65642@fx02.am4> <fkuqilFl1fsU2@mid.individual.net> <RSqGC.80267$6A2.3553@fx21.am4> <slrnpem0sv.jmn.dan@xps-linux.djph.net> <ffEGC.34377$U32.22742@fx04.am4> <pch9jg$66e$1@dont-email.me> <M6DHC.72415$Ze.56654@fx39.am4> <pcng2e$i93$1@dont-email.me> <pcnkjb$i85$1@dont-email.me> <pcnsmp$94e$1@dont-email.me> <lMKHC.29474$RI2.6333@fx26.iad> <pcntoh$fbd$1@dont-email.me> <wCMHC.149313$RM2.140840@fx05.iad> <87o9hsqk7d.fsf@bsb.me.uk> <pcou2j$k4p$1@dont-email.me>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Injection-Info: reader02.eternal-september.org; posting-host="9b5f9547cb5f5c2e6ac0a9ad03204119"; logging-data="30491"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1/uSgz/o1dsC7qn6nBp9w/3"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.1 (gnu/linux)
Cancel-Lock: sha1:hdkJMuFZXmWOe1nLlbFkrYqqM+w= sha1:lLUIwGSB+ZF/nvSNNlaksfHl6iA=
Xref: csiph.com comp.lang.c:130320

David Brown <david.brown@hesbynett.no> writes:
> On 07/05/18 04:21, Ben Bacarisse wrote:
>> Richard Damon <Richard@Damon-Family.org> writes:
>>> On 5/6/18 5:57 PM, David Brown wrote:
> <snip>
>>>
>>> Yes, the normative description is implicitly of a single definition, so
>>> the lack of any other mention makes it fall as undefined by omission.
>> 
>> I'm not seeing this, though I may have lost track of this discussion!
>> The behaviour appears to be explicitly undefined due to the violation of
>> a "shall":
>> 
>>   6.9
>>   ...
>>   5 [...]  If an identifier declared with external linkage is used in an
>>     expression [...], somewhere in the entire program there shall be
>>     exactly one external definition for the identifier; otherwise, there
>>     shall be no more than one."
>> 
>
> It seems 6.9p5 is the bit I have been missing, and that it does say
> explicitly that you may not have more than one definition of the same
> object (or function) in the program.  So despite being commonly
> supported by most C implementations (unless you use something like
> "-fno-common"), having multiple "int x;" definitions in different
> compilation units in the same program is not allowed.
>
> Thanks to Richard and Ben (and others) for educating me here.

It's not exactly "not allowed".  Having multiple definitions violates
a "shall" outside a constraint, so it causes undefined behavior.
An implementation is free to define the behavior in any way it likes,
or to behave in any way it likes without documenting the behavior.

-- 
Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
Working, but not speaking, for JetHead Development, Inc.
"We must do something.  This is something.  Therefore, we must do this."
    -- Antony Jay and Jonathan Lynn, "Yes Minister"