Re: [Vala] Few questions from complete vala begginer

[pancake <pancake youterm com>]

 On 10/08/10 15:47, vml arsis net wrote:
> i'm currently testing vala if i could move some of my work done on Mono/c#
> to it. but, i couldn't find any definitive information on few topics even
> though i studied all documentation available before writing this mail.
> even "NO WAY IN HELL" is acceptable answer as i'm just trying to find best
> porting methods for some services i wrote. i'd simply adapt to what i can
> do, mostly i already have some really ugly workarounds.
>
> are any of these in future plans?
> - extension methods
vala is a static language, you cannot add methods to objects at runtime. 
it takes
no sense, refactoring the code will make this feature unnecessary, can you
provide a example to argue this feature?
> - indexer properties
This feature would hide the underlying C code, but nothing limits you to 
implement
a collection class overwriting the hashmap methods or so that would 
result in the same.
> - custom attributes (any kind, even prepending some xml like comment or
> anything would do the trick, just so one can specify custom descriptions
> in code)
As far as I know this is possible, as long as valac shows no errors, but 
there's no
way to access them from the language, because is something used by the 
compiler.
you cannot introspect this information in runtime. What is the purpose 
of using it
in your code? why do you need this?
> - partial classes
It is already possible to extend the definition of a class in more than 
one file. It is ugly,
but supported, at least for vapis. I never find a reason to use this on 
.vala files. In this
case you can create an abstract class or a base class and extend the 
functionality
in other .vala files.
> partial classes are probably my most needed feature as i auto build a lot
> of source parts and auto building them in the same source as code you
> write is really hard and unpredictable.
you can use 'cat(1)'. it have a very cool feature:  file concatenation.
Nevertheless any preprocessor like cpp, spp or so will do this task. ti 
dont know
how you build those sources, but this is usually the way to do it.
> as far as attributes go, even making the most basic
> [CustomAttribute("Name", value)] parseable in runtime would go well beyond
> the scope of really useful feature if only one can prepend it to
> class/property
You cannot parse code attribute in runtime, those are just compiler hints.
Vala is not dynamic, so you can't do this unless you do it as class 
properties
or so, code attributes are not translated into C code, you have to 
refactor the
code .. and doing all this stuff dinamic is probably 'cool' , but I 
didnt find useful
any of the proposed features. They just make the code more slow and hard to
read.

--pancake