Re: Faulty languages and Liability

["Pat Rogers" <progers@classwide.com>]

"Hyman Rosen" <hyrosen@mail.com> wrote in message
news:PNvR8.12669$cE5.8689@nwrddc02.gnilink.net...
> Pat Rogers wrote:
> > No, that allows the compiler to prevent dangling references.
>
> How's that? That is, what is the difference between the compiler
> noticing which objects you take the address of, or you telling
> the compiler which objects you intend to take the address of?

Yes, from that angle there is no difference. (Note, however, that there is
certainly a difference from taking the address of something, which the compiler
cannot help you with, and using the checked approach we're discussing via
'Access and aliased objects.  Taking the address of something for the purpose of
converting it to an access value is generally not recommended!)

However, marking objects as aliased does have an important effect that the
compiler cannot achieve by just noticing all the references (because the aliased
object will likely be compiled separately from the references).  Specifically,
it prevents the object from being stored in a way that precludes the later
access. For example, it prevents the object from being stored in a register.
See RM 13.1{24}.


--
Patrick Rogers                       Consulting and Training in:
http://www.classwide.com          Real-Time/OO Languages
progers@classwide.com               Hard Deadline Schedulability Analysis
(281)648-3165                                 Software Fault Tolerance