Re: compile time asserts [preconditions]

From: William Kempf (sirwillard_at_[hidden])
Date: 2000-07-31 16:48:18

--- In boost_at_[hidden], "David Abrahams" <abrahams_at_m...> wrote:
> 
> ----- Original Message -----
> From: "William Kempf" <sirwillard_at_m...>
> To: <boost_at_[hidden]>
> Sent: Monday, July 31, 2000 10:08 AM
> Subject: [boost] Re: compile time asserts [preconditions]
> 
> 
> > I strongly suggest you pick a different name.  Preconditions 
already
> > have a very well known meaning, for which a compile time assert 
does
> > not work.  Specifically, a precondition is one of the three
> > components that build up the foundation of DBC (design by 
contract):
> > precondition, postcondition and invariant.  All of these are 
runtime
> > assertions.  A precondition is that that must be upheld before a
> > function can be called (aka it's what you must insure when you 
call a
> > function).
> 
> They're not completely unrelated: the compile-time assert is a very 
special
> case of precondition, where the precondition can be checked by the 
compiler.

First, this is like saying that it's not unrelated to assert, which 
is true, but obviously we won't call it assert.  Second, it's not 
really a precondition at all, compile time or otherwise.  A 
precondition is very specific in what it's asserting, or more 
accurately, where it's asserting.  A precondition is what needs to be 
fullfilled before a function call can be made.  These compile time 
asserts are often used outside of this context.

> That said, I think the name "precondition" suffers the same problem 
as many
> of the other names that we've seen: namely, that it looks like just 
another
> way to write "assert". I am getting tired of listening to myself 
nag y'all
> about this, but I insist that the name needs to contain some 
indication that
> there is static checking going on:
> 
> BOOST_COMPILER_ASSERT
> BOOST_STATIC_PRECONDITION
> BOOST_STATIC_ASSERT
> BOOST_STATIC_CHECK
> 
> all work, with BOOST_STATIC_ASSERT being my favorite so far.

I agree with this totally.  The word "static" is overused, but most 
of us can easily understand its meaning in this context.  It's 
simpler (shorter) than BOOST_COMPILE_TIME_ASSERT, and though I 
usually don't recommend a name based on it's brevity, in this case I 
like the fact that it's shorter.  I don't like the BOOST prefix, but 
understand the necessity for this little evil.

As for the other suggestions:

BOOST_COMPILER_ASSERT - I don't think conveys the fact that this is a 
compile time assert.

BOOST_STATIC_PRECONDITION - Still has the same problems as 
precondition, IMHO.

BOOST_STATIC_CHECK - Is OK, but doesn't convey to me what the action 
will be.  The assert version does, since it shares a name with the 
assert macro.

William E. Kempf