From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00
	autolearn=unavailable autolearn_force=no version=3.4.4
Path: 
 eternal-september.org!reader01.eternal-september.org!reader02.eternal-september.org!news.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Simon Wright <simon@pushface.org>
Newsgroups: comp.lang.ada
Subject: Re: avoiding builtin memset
Date: Wed, 26 Apr 2017 16:22:34 +0100
Organization: A noiseless patient Spider
Message-ID: <ly1ssf6vnp.fsf@pushface.org>
References: <c298b2db-ecfe-4597-8eec-7b69650dcc85@googlegroups.com>
	<934912c8-c189-4d36-8a96-b2aa97e38685@googlegroups.com>
	<c76b2227-105a-4ceb-93e1-5f7a14788574@googlegroups.com>
	<987550450.514781782.903623.laguest-archeia.com@nntp.aioe.org>
	<8d61281f-e908-4965-a7f4-1e0d4b57c967@googlegroups.com>
	<20287957.514851298.264687.laguest-archeia.com@nntp.aioe.org>
	<lya8737hau.fsf@pushface.org>
	<6bff1b83-68c0-4faa-a2db-03df1dd2a108@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain
Injection-Info: mx02.eternal-september.org;
 posting-host="5b4a0556432eb07c6673eeb3967d7d34";
	logging-data="26571"; mail-complaints-to="abuse@eternal-september.org";
	posting-account="U2FsdGVkX1+YhK9tnESIjVJgLcbO6M2E8hxheN+e7Mg="
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/25.1 (darwin)
Cancel-Lock: sha1:NJ/18NFj/GCgLChaP2MOVyjcfnY=
	sha1:ER1ePdW3QPZzXw86eJ2xlUCRZbk=
Xref: news.eternal-september.org comp.lang.ada:46634
Date: 2017-04-26T16:22:34+01:00
List-Id: <comp.lang.ada>

Lucretia <laguest9000@googlemail.com> writes:

> On Wednesday, 26 April 2017 08:35:05 UTC+1, Simon Wright  wrote:
>> Luke A. Guest <me@me.com> writes:
>> 
>> > Shark8 <onewingedshark@> wrote:
>> >  
>> >>> You'll need to export that as the C memset function.
>> >> 
>> >> Even if you're only using it in Ada functions?
>> >
>> >
>> > Yup, GNAT expects certain functions to be present and generates calls
>> > to them. Check system.ads for mention of allowing assignment, that
>> > basically calls memcpy on assignment for objects that require a memory
>> > copy rather than a register to register copy.
>> 
>> Not quite sure this is set in system.ads: in the 6.1 sources, the only
>> reference to memset in the compiler itself is in exp_aggr.adb,
>
> See Support_Composite_Assign and the comment inside targparm.ads:
>
>    --  The assignment of composite objects other than small records and
>    --  arrays whose size is 64-bits or less and is set by an explicit
>    --  size clause may generate calls to memcpy, memmove, and bcopy.
>    --  If versions of all these routines are available, then this flag
>    --  is set to True. If any of these routines is not available, then
>    --  the flag is set False, and composite assignments are not allowed.

Oh, I was only looking in targparm.adb.