1. RE: Euphoria Language Changes
- Posted by bensler at mail.com
Jan 19, 2002
I tend to agree with this. If you want to pass a value through a
routine, use a function.
HOWEVER, referenced variables for type defines could be quite useful..
We would be able to do parameter formatting, and keep the syntax
checking out of our routines..
If I could get assignment on declaration, the slicing shortcut I
mentioned in my previous post, trailing comma support for variable
declarations(this one I could do without), and referenced variables for
type defines, I would gladly pay another $40 or more to have it. That
would be enough modifications to warrant commercial distribution I
think..
Oh yeah, add support for viewing sliced sequences in the trace facility.
Is that part of the source? Or would that comprimise the PD limitations?
Chris
Shawn Pringle wrote:
> Pass by reference in Euphoria? One of the things that has made Euphoria
>
> different and I think easier is it's pass by value policy.
> Before implementors remind me "it is not really pass by value"... I know
>
> it is copy on write I am refering only to how it looks to the language
> definition only not if it actually does memcpy all over the place.
>
> Basically I like EUPHORIA's behavior the way it is NOW. Nothing is
> modified on the right hand side of the assignment operator. Pass
> valueable information into any function or procedure and not worry that
> data is accidently modified when it shouldn't. It also makes EUPHORIA
> an ideal learning language. EUPHOIA objects returned can be of
> unlimited size and structure why pass by reference? only to modify
> things in place.
>
>
> Regards,
> Shawn Pringle
>
>
2. RE: Euphoria Language Changes
-------Phoenix-Boundary-07081998-
Chris Bensler wrote on 1/19/02 7:15:16 PM:
>If I could get assignment on declaration, the slicing shortcut I
>mentioned in my previous post, trailing comma support for variable
>declarations(this one I could do without), and referenced variables for
>type defines, I would gladly pay another $40 or more to have it. That
>would be enough modifications to warrant commercial distribution I
>think..
I have modified my Euphoria source to add:
1) Variable initializations
2) Slicing shorthands
3) String-aware '?'
4) Pass by reference
Example:
sequence test1 = "abc", test2 = "zot"
procedure swap (sequence s, sequence t)
object x
x = s[2]
s[2] = t[2]
t[2] = x
end procedure
swap (!!test1, !!test2) -- '!!' causes pass-by-reference
?test1 -- prints "aoc"
?test2 -- prints "zbt"
?test2[2..] -- prints "bt"
What do you mean by 'referenced variables for type defines'?
Karl Bochert
-------Phoenix-Boundary-07081998---
3. RE: Euphoria Language Changes
-------Phoenix-Boundary-07081998-
You wrote on 1/19/02 7:15:16 PM:
>
>If I could get assignment on declaration, the slicing shortcut I
>mentioned in my previous post, trailing comma support for variable
>declarations(this one I could do without), and referenced variables for
>type defines, I would gladly pay another $40 or more to have it. That
>would be enough modifications to warrant commercial distribution I
>think..
>Oh yeah, add support for viewing sliced sequences in the trace facility.
>Is that part of the source? Or would that comprimise the PD
limitations?
>
>Chris
I have modified the source to provide:
1) Variable initialization.
2) String-aware '?' operator
3) Slicing shorthands.
4) Pass by reference.
Next up: structures (sequences with named elements)
and dot-notation.
sequence test1 = "abc", test2 = "zot"
procedure swp (sequence s, sequence t)
object x
x = s[2]
s[2] = t[2]
t[2] = x
end procedure
foo (!!test1, !!test2) -- '!!' invokes pass-by-reference
?test1 -- prints "aoc"
?test2 -- prints "zbt"
?test2[2..] -- prints "bt"
Karl Bochert
-------Phoenix-Boundary-07081998---
