Re: Ada as a first language

["Peter C. Chapin" <pcc482719@gmail.com>]

Brian Drummond <brian_drummond@btconnect.com> wrote in
news:r33fs4h5di2v0si3k8mi2vetqeglsdchfu@4ax.com: 

> which is why I suggested "fit and forget" for the C libraries. If they
> just work, you have saved time. If not... I'd rather use Ada for
> interfacing. Does AdaMagic have trouble translating low level code
> like interrupt handlers or representation clauses?

It has a configuration file where you describe the various 
characteristics of the underlying C compiler (sizes of the C basic 
types, endianness, etc). It uses this information to generate 
appropriate C code. When I set up the configuration file I had to run a 
few test programs using the C compiler to find out what it did in some 
cases. The documentation that came with the C compiler was not specific 
enough. In our project, which is fairly simple, we have not (yet) had 
the need to use representation clauses, but I assume they would work 
given the knowledge AdaMagic has about the C compiler's basic types, 
etc.

Also in our project we have only two simple interrupt handlers so we 
(the student) just wrote them in C. Essentially our device spends most 
of its time in a deep sleep mode and then is awakened by an interrupt. 
The Ada code calls a procedure, implemented in C, that puts the 
processor to sleep. That procedure won't return until it is awakened... 
something that is done in the interrupt handler. In fact that's 
basically all the interrupt handler does.


> AdaMagic doesn't look free, which may be a consideration for personal
> experiments.

That's right. It is a non-free product for sure. One nice think about 
working in academia is that I can often get commercial toys cheap. :-)


Peter