Re: Ada GC

[Courtesy cc: sent to author.]


Kevin Weise <kweise@pluto.colsa.com> wrote:
> I admit I haven't read the specifications and definitions of all new
> languages for the past twenty years. But I have *never* seen a language
> specification that did this, except for certain assemblers that told you
> how many machine cycles each instruction required to execute. I would
> certainly appreciate some reference to a real-time programming language
> that provides as part of its requirements/specifications bounds on the
> various primitive operations it provides.


http://www-ccs.cs.umass.edu/spring/compiler.html and various technical
reports from the real-time group at UMASS on Spring C. I don't recall
TR numbers offhand, but can find them if you can't.


You can specify hard deadlines for sections of code. There are
limitations (one is that the proof-of-concept implementation relies on
the architecture manual's specification of instruction execution times
(which are often wrong -- you need to design test to get better
readings -- one's that assume worst-case scenarious (flushing the
pipeline, etc..)).


anyway, that's one such example.


I think this is getting off of comp.lang.ada Follow-ups directed to
comp.compilers and comp.realtime


-- Jason
--

Return to the comp.compilers page.
Search the comp.compilers archives again.