SET, REVISITED
A more precise statement is that evaluating a block containing a
set-word (remember that a balanced entry into the REBOL console is
treated as if contained by an implicit block) associates the word
with the value of the expression following it. After evaluating:
a: "1234"
x: [a: "987"]
we have
a -> {{string! 1 *}}
|
V
<<#"1" #"2" #"3" #"4">>
x -> {{block! 1 *}}
|
V
<<a: *>>
|
V
{{string! 1 *}}
|
V
<<#"9" #"8" #"7">>
(There is undoubtedly more "under the hood" for a set-word, but this
essay is about series, so the details of set-word are ignored in the
above diagram.)
After evaluating
do x
the situation is different (and fairly distinctive to REBOL):
a -------------------------+
|
|
x -> {{block! 1 *}} |
| /
V /
<<a: *>> /
\ /
V
{{string! 1 *}}
|
V
<<#"9" #"8" #"7">>
Note that a now refers to the same string as referred to by the
second element of x's value. This is different from setting a to
refer to a COPY of that "string literal", which does NOT happen in
REBOL. To achieve this last effect, x would have to be set with:
x: [a: copy "987"]
Therefore, if a is passed to an operation (such as 'append or
'insert) which actually modifies the value of its character
sequence, the modification will show up via x as well. The console
transcript bears out these predictions:
>> a: "1234"
== "1234"
>> x: [a: "987"]
== [a: "987"]
>> do x
== "987"
>> a
== "987"
>> append a "!!!"
== "987!!!"
>> x
== [a: "987!!!"]
Since a and second x refer to the same string, this trick works in
the other direction as well.
>> insert second x "???"
== "987!!!"
>> x
== [a: "???987!!!"]
>> a
== "???987!!!"
(continued in essay/6)
