From mboxrd@z Thu Jan  1 00:00:00 1970
From: Aldy Hernandez <aldyh@redhat.com>
To: mike stump <mrs@windriver.com>
Cc: geoffk@redhat.com, ira@apple.com, gcc@gcc.gnu.org
Subject: Re: Target-specific Front-Ends? (Was: front end changes for  altivec)
Date: Tue, 27 Nov 2001 18:28:00 -0000
Message-ID: <m3adx7vd7i.fsf@litecycle.cc.andrews.edu>
References: <200111280212.SAA10114@kankakee.wrs.com>
X-SW-Source: 2001-11/msg01455.html
Message-ID: <20011127182800.4QybJJ0Bpc5Kw2lsVrCOAGacO-WSnXRqCmEa47V1P7c@z>

>>>>> "mike" == mike stump <mrs@windriver.com> writes:

 > gnu::gcc::vector<int[4]> a, b;
 > vector<int> cxx_vec;

 > void foo() {
 >      a += b;
 >      cxx_vec.begin();
 > }

 > A completely trivial problem.  :-)


for the record, i'm not proposing a vector keyword-- but a __vector
keyword, to be implemented in a target independent manner.

I don't think there's any plausible way of implementing "vector" and
having it coexist with STL and a myraid of other things.  That much is
clear.