comp.lang.ada
 help / color / mirror / Atom feed
From: Jeffrey Carter <spam.jrcarter.not@spam.not.acm.org>
Subject: Re: Problem implementing Iterators
Date: Tue, 27 Jan 2015 13:58:38 -0700
Date: 2015-01-27T13:58:38-07:00	[thread overview]
Message-ID: <ma8u4v$i7v$1@dont-email.me> (raw)
In-Reply-To: <ma8m6r$e0r$1@dont-email.me>

On 01/27/2015 11:43 AM, Björn Lundin wrote:
> 
> As I understand it, its only purpose is to tell the programmer that
> this variable may be changed via pointers - beware ...

My understanding is that there are certain optimizations that can be taken with
unaliased objects that cannot be done with aliased objects, so this provides
important information to the compiler.

Some time ago (late 1980s or early 1990s) there was an article in /Ada Letters/
about why the Tartan Ada compilers were able to produce faster code for Ada
versions of their C benchmarks than their C compilers. The fact that all C
variables are aliased, while all Ada-83 variables are not, was one of the factors.

-- 
Jeff Carter
"When Roman engineers built a bridge, they had to stand under it
while the first legion marched across. If programmers today
worked under similar ground rules, they might well find
themselves getting much more interested in Ada!"
Robert Dewar
62

  reply	other threads:[~2015-01-27 20:58 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-01-27 13:54 Problem implementing Iterators Laurent
2015-01-27 16:24 ` AdaMagica
2015-01-27 17:26   ` Laurent
2015-01-27 17:55     ` Björn Lundin
2015-01-27 18:06       ` Laurent
2015-01-27 18:43         ` Björn Lundin
2015-01-27 20:58           ` Jeffrey Carter [this message]
2015-01-28 12:20             ` Björn Lundin
2015-01-28 16:42     ` AdaMagica
2015-01-28 20:14       ` Laurent
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox