[Gambas-user] tags

Jean-Yves F. Barbier 12ukwn at ...626...
Sat Nov 14 20:42:04 CET 2009


Doriano Blengino a écrit :
...
> You can do the same in gambas, using Object[] or Collection, and 
> creating a class for every kind of record.

I found coll["key"]=value and I think that could do the trick,
I'm gonna test this tonite

> Unfortunately I don't know them enough to explain to you. But I can 
> spend two words about python and ruby, and in general these "agile" 
> languages. They are very handy, but also two problems arise. The first 
> is that, in your example, if you write "o.a", where "o" and "a" are two 
> identifiers (because they are not surrounded by apices or quotes), they 
> don't resolve to two fixed address in memory - so they are no more 
> identifiers. This wastes a lot of CPU cycles, because the interpreter 
> must scan all pool of objects to find them. In other words, python is 
> *slow*. Good to explain concepts, but in practical, heavy applications 

Yeah, I agree; it is easy to see that on my old C2.4GHz

> it is a pain. We can compare two programming IDEs: gambas and some other 
> written in python (Boa? or others I can't remember...). On my machine, 
> they simply suck - when gambas is quick and responsive. And think that 
> an IDE is not a particulary heavy application.
> 
> The second problem is that, I think, they don't check enough at compile 
> time (because they can't). For me, coming from pascal, this is a big 
> issue. Most of the time the compiler (pascal, or C) catches all my 
> typing errors, and the rest is ok. But if the compiler does not catch 

well, that doesn't prevent mallocs missing...

> errors, you are never sure that your code is ok. I am already critic 
> with some constructs that gambas does not check enough (for me) - so I 
> really can't stand with less rigid languages.

there should be 2 modes: regular (work as of now, useful when you just have
one proc to test, even if syntax doesn't match for similar things) and strict
(check every detail).

> Just a simple opinion. I think that if you investigate well your needs, 
> you will find a clean and effective way to solve with gambas.

this is the "problem": I learn it while making my pgm.

-- 
Good news.  Ten weeks from Friday will be a pretty good day.




More information about the User mailing list