Re: (LibAnet)(Patch) This Patch made possible the use in High Integrity Systems.

Hi!

2018-05-28 6:48 GMT-03:00 Reto Buerki <reet@codelabs.ch>:

Hi,

On 05/25/2018 07:31 PM, Dani wrote:
> Hi!
>
> What is the timeframe for including the patch?
>
> Note1: Distributed Anex programmers are also benefited.
> Note2: Does the patch make LibAnet and programs that use LibAnet more
> efficient? Yes!
> Why? Because the Ada compiler can make better decisions when
> compiling programs! :-)

I'm sorry but I'm not going to include a patch set of that size just for
the sake of some detect blocking, pure and preelaboration pragmas. In my
opinion the changes are just to invasive for a small benefit.

At this point I disagree with you because the benefit is quite large. I will enumerate some:

Only on anet.ad{s,b} was eliminate 5 predefinide packages dependences of predefinide library.

Issues with unbounded string was confined in unix socket. other packages, beyond

other eliminate dependences, dont use unbounded string anymore.

Only this speed up LibAnet e Software using LibAnet considerably, But are more:

Types declared with pragma Preelaborable_Initialization are load more speed up both in run-time and compile-time,

both the library itself and software using LibAnet.

Functions declared in pure packages, and functions declared pure_functions, too, beyond

considerable real speed up, both library and software, the compiler can reuse a result

these functions without a new call if values are equal and speed up yet more.

The ultimate first example of these pragmas/attributes is the Ada Pre-defined Library. When it's time

to upgrade Ada ISO Standard, there are whole and heated debates about whether 'package such will be pure or preelaborate' ,

because this affects a huge amount of programmers and software houses, especially those linked to contracts

with governments, banks, and embedded applications.

Without these pragmas/atributes programmers don't can use Distributed Anex neither the categorização pragmas,

because gnat will proibit this use with a Great 'Compilation Error ...'

As for the names of tools you asked this is a secret of the software manufacturers and

the governments for which they work. Spark also has these features but I've never heard a single name for Spark.

Consider at least creating a Git Branch and a package for this patch, I'll be glad if that happens.

I hope you understand my point,

Best Whishes,

Dani.

I hope you understand my point.

Regards,
- reto

> 2018-05-14 15:28 GMT-03:00 Dani <danielcheagle@gmail.com>:
>
>> Hi! :-)
>>
>>
>> 2018-05-14 6:16 GMT-03:00 Reto Buerki <reet@codelabs.ch>:
>>
>>> Hi,
>>>
>>> Thanks for your contribution to the anet project.
>>>
>>> On 05/12/2018 09:45 PM, Dani wrote:
>>>> Because I Known the source code of both,
>>>> I believe Libanet is faster and safer than gnat-sockets,
>>>> But gnat-sockets et.all claim categorically to be secure in the domains
>>> of
>>>> High Integrity, High Security and Real Time Systems. Whats "The Magic"
>>> for
>>>> this ? the Magic for this
>>>> is have categorization pragmas ( or attributes) and some initial
>>> re-work;
>>>
>>> I doubt that adding some pragmas makes a binding to the OS provided
>>> socket implementation high secure all of a sudden, but on the other hand
>>> the term 'high secure' is used for SSL websites nowadays, so what do I
>>> know.
>>>
>>
>> I agree with you; adding some pragmas does not make it any safer than
>> it already is. these pragmas only allow categorization. these
>> categorizations in
>> Ada help the compiler to separate what can and can not be used in a
>> program.
>>
>> Libanet already is secure in my vision.
>>
>>
>>>> With this categorization added tools from this fields can be used
>>>> flawlesse. :-D
>>>
>>> Can you please elaborate which tools you are referring to?
>>>
>>
>>
>> I Don't known theses tools beyond gnat itself, mas the _ use and
>> existence_ these tools
>> are touted by Software Manufacturers and required by Government Agencies
>> and Buyers of such systems.
>>
>> My prime motivation is a really huge system that I'm creating in Ada
>> with a friend. this system
>> involves banks and money and at some point we will be required by some
>> regulatory agency to
>> pass through the sieve of those tools required by these.
>>
>> I can't create we system already with categorization pragmas and or
>> proibited
>> by compiler the use o Libanet because Libanet don't have these
>> categorization pragmas. :-D
>>
>> Really does not stop being more bureaucracy. : -D But the existence of
>> these
>> categorizations in the Ada Language is well publicized and much required
>> in some niches.
>>
>> Unfortunately a large part of our system will have to pass the sieve of
>> one of these
>> niches at some point in the future.
>>
>> and Thanks for understanding my point.
>>
>>
>>> Kind regards,
>>> - reto
>>>
>>
>> Best Wishes, Dani.
>>
>> --
>> "There are many plans in the Human heart, But
>> is the Lord's Purpose that prevails"
>>
>> "Existem Muitos planos e desejos no coração Humano, MAS
>> são os Propósitos do Senhor que prevalecem"
>>
>> []'s Dani:-)
>>
>
>
>

"There are many plans in the Human heart, But
is the Lord's Purpose that prevails"

"Existem Muitos planos e desejos no coração Humano, MAS
são os Propósitos do Senhor que prevalecem"

[]'s Dani:-)