Re: Janus/Ada 3.2.1 Released!

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

"Shark8" <onewingedshark@gmail.com> wrote in message 
news:db462b34-69f3-4641-9e6b-016e9f48bd54@googlegroups.com...
> On Thursday, June 27, 2019 at 11:48:49 AM UTC-6, Randy Brukardt wrote:
>> 5th (68020 for SunOS), built for a specific customer and never released 
>> (or
>> quite finished);
> Hm, interesting.
> How much would be required to do a 68020 vxWorks [5.2]?
> Or would bare-metal on the 68020 be easier?

Hard to know. That project was abandoned around 1990, so it was Ada 83. I'd 
guess that the backend would not need a lot of changes (the basic design 
hasn't changed much). Probably the target-specific stuff wouldn't need a lot 
of changes, either (but there were a lot of options added for the U2200 
project, and those would have to be redone). The Ada parts of the runtime 
would probably not need much updating, either. However, tasking was never 
finished for that compiler, and new tasking like protected objects would 
have to be created. Similarly, finalization and exception information are 
newer than that compiler. And target I/O would have to created (for whatever 
sort of target). There's enough unknowns that it would be hard to estimate. 
Of course, if there was actually a customer for it, one could spend some 
time poring over old code to see how much would need to be changed (not 
worth it for random cla chat).

>> 6th (SPARK), also built for a specific cutomer and never really tested 
>> after
>> funding disappeared;
> How much would be required to bring this up to SPARK 2014?

SPARC, the processor, not SPARK the whatever it is :-). Sorry about any 
confusion.

> Also, regarding DSA, would it be possible to use Janus/Ada to split a 
> program across, say:
> (1) a Motorola 68k (running vxWorks, or bare-metal),
> (2) a Sun Ultra 45 (UltraSPARC IIIi, running SunOS),
> (3) several x86-64 (Intel and/or AMd, running Windows)
> Or is that beyond the capabilities of the tooling? (And if so, how much 
> would you estimate this DSA tooling to cost?)

We never did any DSA tooling (or support). I tried to keep the design so 
that it would be possible if demand for it appeared, but it never did. The 
few cases where one needed that sort of support were easier to manage with 
cooperating programs using TCP/IP streams (rather than building lots of 
complex compiler support).

I'm sure it's possible, but I don't have a clear idea of how much effort it 
would take. (Besides,anything always takes twice as much effort as expected, 
even something simple like balancing one's credit card amounts.)

                                                Randy.