Subject: | Re: date grief
| Date: | Wed, 6 Dec 2017 07:09:12 +1000
| From: | Bernie van't Hof <berniev@bje.com.au>
| Newsgroups: | pnews.paradox-dos
|
One little example for you to ponder...
A "spec" implies a thought-out structure.
pdox qbe date column:
23-jun-2016 + 1 (ok)
1 + 23-jun-2016 ("expression in this field has the wrong type")
If there was a "spec" I would have thought such a simple date expression
would be evaluated before deciding if the resultant type was correct.
So in some ways pdox qbe is astoundingly capable, yet in others it is
simplistically flawed. Finding these quirks is proving interesting.
- Bernie
On 5/12/17 5:56 pm, Leslie wrote:
> Not on hand, I would have to go back to my Borland white papers from the
> 80's. I did scan a whole heap of stuff and archive it so I may have an
> old copy from about 1991-ish.
>
> But are you saying that the QBE definition is not online somewhere?
>
> Come to think of it has anyone done any searches for an existing qbe
> parser. I haven't but perhaps someone has written one already.
>
> When I get some time I will see what I can find out.
> Leslie.
>
> On 5/12/2017 5:43 PM, Bernie van't Hof wrote:
>> If you have a copy I'd be more than happy to have a look..
>>
>> On 5/12/17 4:53 pm, Leslie wrote:
>>> QBE parser as per the standard
>
> ---
> This email has been checked for viruses by AVG.
> http://www.avg.com
>
|