From 8950133019234075584
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: f78e5,17221ee87fa64d29
X-Google-Attributes: gidf78e5,public
X-Google-ArrivalTime: 1993-03-09 11:20:28 PST
Path: sparky!uunet!gatech!howland.reston.ans.net!agate!darkstar.UCSC.EDU!daniel
From: daniel@cse.ucsc.edu (Daniel R. Edelson)
Newsgroups: comp.std.c++
Subject: Re: useless declarations
Date: 9 Mar 1993 19:13:19 GMT
Organization: University of California, Santa Cruz (CE/CIS Boards)
Lines: 19
Message-ID: <1niq8fINNa45@darkstar.UCSC.EDU>
References: <1ng5q8INNeu0@darkstar.UCSC.EDU> <1993Mar9.135445.23786@lpi.liant.com>
NNTP-Posting-Host: oak.ucsc.edu

In article <1993Mar9.135445.23786@lpi.liant.com> pkt@lpi.liant.com (Scott Turner) writes:

>> 	struct outer {
>> 		struct {
>> 		    static int y;
>> 		};
>> 	};

>This example is not legal because it's impossible to satisfy the 
>requirement of defining the static member y.

Ok. However, syntactically this is legal (enough)
for this translation unit to compile to an object module?
Since the static member could conceivably be defined
in another translation unit, the error might not be
detected here?