Better assert's status? (Was: Proposal for std.path replacement)

[spir]

On 03/07/2011 07:20 AM, Jonathan M Davis wrote:
> On Sunday 06 March 2011 21:57:30 Nick Sabalausky wrote:
>> "Jonathan M Davis"<jmdavisProg at gmx.com>  wrote in message
>> news:mailman.2293.1299467610.4748.digitalmars-d at puremagic.com...
>>
>>> On Sunday 06 March 2011 18:08:49 Andrei Alexandrescu wrote:
>>>> Yah, thing is people work on stuff they care about, not the most urgent
>>>> stuff - surprise! :o) As such we don't have a ton of proposals for
>>>> networking and xml, but we do have one (and I won't argue it's a bad
>>>> one) for rehashing a module that basically worked.
>>
>> I'm not sure I'd say the current std.path "basically works", but I get what
>> you mean.
>>
>>> I _was_ thinking of putting forward a new proposal which includes the
>>> unit testing functionality that assertPred had which won't end up in an
>>> improved
>>> assert,
>>
>> Speaking of which: Now that assertPred has been rejected on the grounds of
>> an improved assert that doesn't yet exist, what is the current status of
>> the improved assert?
>
> There's an enhancement request for it:
>
> http://d.puremagic.com/issues/show_bug.cgi?id=5547
>
> I have no idea of any work is actually being done on it or not. It hasn't
> actually been assigned to anyone yet, for whatever that's worth. Honestly, it
> wouldn't surprise me if it doesn't happen for a while. I'm not sure that anyone
> who is capable of doing it is particularly motivated to do it (though I'm not
> sure that they're _not_ either). It was clear that a number of people wanted
> assert to be smarter rather than having assertPred, but it isn't clear that
> assert is going to be made smarter any time soon. I suspect that it will be a
> while before it's done. We'll have to wait and see though.

IIUC:
The problem is this feature belongs to the category of things that cannot be 
implemented by any D programmer, in D, as a lib feature, even by an expert in 
the domain.
It needs to get a representation of the unevaluated expression beeing asserted, 
meaning compiler support, meaning hard low-level C/++ and a great knowledge of 
the compiler architecture, esp the construction of the AST. If there was a way 
to "quote" D expressions, and get their representation at runtime, then we 
could do it ourselves (would imply some perf penalty, but I consider this worth 
compared to the terrible expressive power gained, and in fact totally 
neglectible for an assert statement).
Please tell me where I'm wrong.

With the same power, I would implement at once 'varWrite':
	int x = 3; s = square(x);
	varWrite("value: 'x' --> square: 's'");
	// --> "value: 3 --> square: 9"
or even maybe:
	int x = 3;
	varWrite("value: 'x' --> square: 'x*x'");
	// --> "value: 3 --> square: 9"


Denis
-- 
_________________
vita es estrany
spir.wikidot.com