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!.POSTED!not-for-mail
From: Simon Wright <simon@pushface.org>
Newsgroups: comp.lang.ada
Subject: Re: Ada to Ada Translator ?
Date: Thu, 23 May 2019 08:56:29 +0100
Organization: A noiseless patient Spider
Message-ID: <lyv9y1y576.fsf@pushface.org>
References: <100ad407-090e-4316-9746-a4469568b53e@googlegroups.com>
	<64883feb-3e49-4c6a-855c-6673068e970c@googlegroups.com>
	<qbvcmq$98f$1@franka.jacob-sparre.dk>
	<a8d0f6ca-d5cf-4831-8803-9c1bb6198116@googlegroups.com>
	<qc1tap$6pk$1@franka.jacob-sparre.dk>
	<1bb467ae-6e2c-4fef-aaab-556a8df71747@googlegroups.com>
	<qc4an4$bjv$1@franka.jacob-sparre.dk>
	<bb12e299-8620-4c81-9c48-873302e5dfc0@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain
Injection-Info: reader02.eternal-september.org;
 posting-host="050c443c50f1dfb0b3a49102ad6088c2";
	logging-data="30266"; mail-complaints-to="abuse@eternal-september.org";
	posting-account="U2FsdGVkX19q3X74c5DxUyrOd/eiIaPuV4HyLaWmzmM="
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (darwin)
Cancel-Lock: sha1:NciH8KOKiSk/UR9uoNBhNq5FOME=
	sha1:Jzr1Kpiqxsy3p+qguJfYCEzv6l0=
Xref: reader01.eternal-september.org comp.lang.ada:56376
Date: 2019-05-23T08:56:29+01:00
List-Id: <comp.lang.ada>

foo wong <crap@spellingbeewinnars.org> writes:

> So I guess where I was going with the pragmas, is that I would like to
> be able to write Ada code that had tasking/exceptions etc but by the
> time it reached gnat, these pragmas could be used to avoid linking to
> any runtime from Adacore.

You might be interested in a note I wrote about creating an RTS that can
be used by the existing GNAT compiler; see [1], scroll down to 'RTS:
second stage'.

In my case, I was OK with a GPL/RLE runtime, so quite happy to modify
existing FSF GCC files where they were useful; but I don't think there'd
be any problem with reimplementing them. You might need to consider a
'clean room' approach (or not; IANAL).

The sort of assumptions made by the compiler are that - for example, to
create a Ravenscar task - there will exist a package
System.Tasking.Restricted.Stages, containing (at GCC 8/9, GNAT CE 2018)

   procedure Create_Restricted_Task
     (Priority             :        Integer;
      Stack_Address        :        System.Address;
      Size                 :        System.Parameters.Size_Type;
      Sec_Stack_Address    :        System.Secondary_Stack.SS_Stack_Ptr;
      Secondary_Stack_Size :        System.Parameters.Size_Type;
      Task_Info            :        System.Task_Info.Task_Info_Type;
      CPU                  :        Integer;
      State                :        Task_Procedure_Access;
      Discriminants        :        System.Address;
      Elaborated           :        Access_Boolean;
      Chain                : in out Activation_Chain;
      Task_Image           :        String;
      Created_Task         :        Task_Id);

I don't support exceptions, they're fearsomely complicated, and I don't
believe that embedded software should rely on exceptions to provide
continued execution past errors, and SPARK wouldn't allow them
anyway. See notes at [2] - especially Robert Dewar's remarks linked to
at the end.

[1]
https://forward-in-code.blogspot.com/2015/06/building-runtime-system-for-arm-eabi.html
[2] https://github.com/simonjwright/cortex-gnat-rts/wiki/ExceptionHandling