Re: Ravenscar References

["Randy Brukardt" <randy@rrsoftware.com>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1405 bytes --]

"Jeffrey R. Carter" <spam.jrcarter.not@spam.not.acm.org> wrote in message 
news:q22a5c$d8o$1@dont-email.me...
> On 1/18/19 8:18 PM, lyttlec wrote:
>>
>> I need to port lots of existing more or less standard
>> components to meet Ravenscar. This is to satisfy some regulatory
>> authorities.
>
> I don't know that "port" is a good word for this activity. I once looked 
> at implementing Sandén's FMS problem using Ravenscar. Starting from the 
> requirements, I first had to find a Ravenscar-suitable design. The 
> standard design has a dynamic task per job, and is clearly not possible 
> using Ravenscar. An alternative design using a task per workstation had to 
> be used.
>
> From that choice, Ravenscar drove a proliferation of protected objects and 
> helper tasks. Things that were simple in full Ada became much more complex 
> to meet the restrictions of the profile.
>
> Presumably you would need to apply a similar process to each of the 
> components you need to convert.

Note that the less strict profile Jorvik, defined in Ada 2020 (and already 
implemented in GNAT) would simplify this process.

I don't think it is possible to "convert" regular Ada code into Ravenscar 
(unless, of course, it doesn't use any tasks ;-). You pretty much have to 
completely rewrite it with Ravenscar in mind. (In this way, it is very much 
like using SPARK.)

                                    Randy.