[V4REV] Constants

[ron barber]

Ruslan, Lynn,

I will expose my ignorance with this post but I think I may represent 
at least some in the Rev community that do not care to use dots at all. 
Perhaps the rest of the of the world uses dot notation, but those in 
the xtalk world don't have the.slightest.idea what you are talking 
about. The docs at this point are not very helpful because they have 
largely been adapted from the RB docs, in fact multiple references 
remain to RB itself and RB notation - eg classes, subclasses. If you 
end up using dot notation, you must put some kind of explanation in the 
v4rev docs because it is foreign to xtalk users.

I understand your desire to be industry standard and you must cater to 
the large RB community (that is laughing at me right now!) that Val 
serves, but if you want to also serve the Rev community, please make 
some concessions to our way of doing things.

I realize we are still in beta and this is what beta testing is all 
about.  Lynn, your xtalk background should at least understand what 
Robert and I (and others?) are saying.

I know you want to produce a good product and I'm trying to give some 
input in that direction. I am a little frustrated after spending a 
considerable time testing and crashing and it is due at least partially 
to my misuse of dot notated parameters.

Thanks
Ron

ps and don't get me started about the + sign...

On Jun 20, 2005, at 2:35 AM, Ruslan Zasukhin wrote:

> On 6/19/05 7:01 PM, "Robert Brenstein" <rjb at robelko.com> wrote:
>
>>> Currently, there is not variable name but just a string which is
>>> parsed on plugin side.
>>>
>>>
>>> --
>>> Best regards,
>>> Ivan Smahin
>>
>> It is a string if it is quoted, ie "kUnique".
>>
>> If it is not quoted, ie kUnique, Rev checks first if it is an
>> existing variable. If it is not, it creates a variable on the fly and
>> sets its value to its name. At least that is what I understand as
>> general operating procedure.
>
> Yes, this is the only way.
>
>> For your function, the value passed is the same in both cases, but
>> the process is somewhat different.
>>
>> Quoting will be required for "kIndexed + kUnique"; otherwise, users
>> will get runtime error unless they define kIndexed and kUnique as
>> variables or constants themselves.
>
> right
>
> This should be pointed in docs by big letters!
>
>> I have no clue whether adding OO features and thus a special meaning
>> to a period inside an unquoted string will break this or not. This is
>> why I suggested to consult with RunRev team.
>
> Robert,
>
> Your point is very correct....
> If we use DOT, then we still can future use quotes...
>
> With underscores we get:
>
>         EVOnDelete_kCascade
>
>         "EVFlag_fIndexed + EVFlag_fUnique"
>
>         "fIndexed + fUnique"
>
> Well, also acceptable...
>
> -------
> Ivan, you have implement  "+"  for flags ?
> Not coma ?