Re: Ada, the best language with the not-so-best tool chain

[Niklas Holsti <niklas.holsti@tidorum.invalid>]

On 12-10-10 10:06 , Ludovic Brenta wrote:
> Niklas Holsti writes:
>> On 12-10-10 08:56 , Stephen Leake wrote:
>>> kalvin.news@gmail.com writes:
>>>
>>>> I agree with Niklas. Some bare metal system require that the (C
>>>> language) startup code does some initialisation for the runtime
>>>> system.
>>>
>>> I'm confused. If you are using someone else's "runtime system", then
>>> in what sense is this "baremetal"?
>>
>> It was Ludovic who said that baremetal programming is hard in any
>> language, especially the bootstrap stage.
>>
>> The bootstrap can indeed be complex. Some of the bigger, more complex
>> microcontrollers start up after reset in a kind of "safe mode" with
>> perhaps no usable SW stack, perhaps a very slow clock, and perhaps
>> unable to access off-chip memory. To enable normal operation, the
>> clock generator has to be configured, the off-chip memory controller
>> has to be configured, perhaps some parts of the app SW have to be
>> copied from the flash to RAM, etc.
> 
> Yes.  When I mentioned bootstrap I meant it in two senses: the one you
> just described i.e. bootstrap of the processor and other devices, and
> also the bootstrap of the toolchain.

Ah, I didn't understand that you meant also toolchain bootstrap, sorry.
I agree that is difficult, both for C and for Ada.

> If your target doesn't have a
> compiler available, you must create one before you can do anything.

Yep. And here is the main difference between C and Ada. C
cross-compilers for most targets exist; for Ada not.

> Practically speaking that means binutils and GNAT configured as
> cross-tools for that particular board.  Also the connection between the
> host and target computers (USB? JTAG? RS-232? etc.)  This doubles
> complexity of the *initial* investment one has to do before even
> starting to program their application.

Yes. And this leads to the interesting and (I think) crucial question:
assuming that GNU cross-tools exist for C for some target, how difficult
is it to add GNAT to those tools? With and/or without tasking? And how
difficult is it to maintain it when GNU and GNAT evolve? Apparently
quite difficult, since so few have been done.

> I agree that it would be very nice if there were some preconfigured
> cross-toolchain for Ada for some kind of "standard" bare-board target.
> AVR-Ada and GNAT for Lego Mindstorms are currently the closest things to
> such a "standard" but beyond that, programmers are on their own to
> create a cross-toolchain.

Agreed.

> It is feasible (after all the ones who
> created AVR-Ada and GNAT for Lego Mindstorms were "on ther own" too) but
> not for the faint of heart and certainly not for beginners.

I think the OP's concerns was that there seems to be little activity to
drive the creation of such Ada tool-chains, and what can we do about
that? Can we interest the chip-makers (e.g. Atmel) or cross-tool
providers (e.g. Crossworks) to do or support such activities? Can we
cooperatively provide how-to guides for adding cross-GNATs to
cross-GNU-C tools?

> That's why I stand by my recommendation that hobbyists who have a
> specific application in mind (rather than a specific target board they
> want to play with) should avoid these two bootstrap stages altogether
> and start programming their application on a Raspberry Pi or Lego
> Mindstorms NXT.

I strongly agree.

-- 
Niklas Holsti
Tidorum Ltd
niklas holsti tidorum fi
      .      @       .