Hi,
see the following:
a: "Global"
use [a] [
a: "Local"
b: [print a]
c: "print a"
do b
do c
]
Result:
>> a: "Global"
== "Global"
>> use [a] [
[ a: "Local"
[ b: [print a]
[ c: "print a"
[ do b
[ do c
[ ]
Local
Global
I think, that this is what they meant.
Ladislav
> The new manual says:
>
> Generally, the use of do for evaluating strings is not a
good
> practice. It is much slower to process than evaluating
blocks,
> and the context of words in a string is unknown.
>
> The first drawback ("much slower") is certainly understandable,
but
> the second ("context ... is unknown") is not clear to me. Based
> on the experiment:
>
> >> printa: "print a"
> == "print a"
> >> a: "global"
> == "global"
> >> do printa
> global
> >> f: func [/local a] [a: "local" do printa]
> >> f
> global
> >> f2: func [/local a] [a: "local" do "print a"]
> >> f2
> global
> >> f3: func [a] [do printa]
> >> f3 "argument"
> global
> >> f4: func [a] [do "print a"]
> >> f4 "argument"
> global
>
> it would appear that the version I'm using (2.2.0.3.1)
consistently
> associates words done from a string with the global context.
>
> I'm familiar with language specifications that state that
something
> is "undefined", meaning something like the following:
>
> We really don't want you to do this, even though the formal
> syntax allows it -- it was just too much trouble to rewrite
> the syntax to make it illegal. So, if you do this, all bets
> are off. An implementation can do anything it pleases, and
> can change behavior between versions, or even successive
> executions of the same code. You have now been officially
> warned, so if you try it and get bitten, don't come whining
> to us.
>
> Consequently, I'm not sure whether the phrase, "the context of
> words in a string is unknown", should be understood as:
>
> a) formally "undefined", in the sense discussed above;
> b) no firm policy decision had been made on the subject, so
> the manual was intentionally left vague;
> c) the author of the manual just didn't know;
> d) the manual is simply wrong, and the statement should have
> read "words appearing in a string will be bound to the
> global context, which may not be what you expect. This
> may produce some very subtle bugs, which can frustrate you
> to the point of requiring professional help." (;-); or
> e) some other interpretation I need to have explained to me.
>
> Based on comments from Gabriele, Elan, and others regarding when
> contexts are created, I can certainly understand why
implementing
> option (d) might be easier than some of the alternatives. If
> that's the case, why not just tell us?
>
> Any opinions, anyone?
>
> -jn-
>
>
>
