R devel - Apr 2012 - A doubt about substitute() after delayedAssign()

Hello,

?delayedAssign presents substitute() as a way to look at the expression 
in the promise. However,

msg <- "old"
delayedAssign("x", msg)
msg <- "new!"
x #- new!
substitute(x) #- x (was 'msg' ?)

Here, we just got 'x'... shouldn't we got 'msg'?

Same result when the promise is not evaluated yet:

delayedAssign("x", msg)
substitute(x)

In a function, that works:

foo <- function (x = msg) substitute(x)
foo()

Did I misunderstood something? It seems to me that substitute() does not 
behaves as documented for promises created using delayedAssign().
Best,

Philippe
-- 
..............................................<?}))><........
  ) ) ) ) )
( ( ( ( (    Prof. Philippe Grosjean
  ) ) ) ) )
( ( ( ( (    Numerical Ecology of Aquatic Systems
  ) ) ) ) )   Mons University, Belgium
( ( ( ( (
..............................................................

On 12-04-29 3:30 AM, Philippe Grosjean wrote:
 > Hello,
 >
 > ?delayedAssign presents substitute() as a way to look at the expression
 > in the promise. However,
 >
 > msg<- "old"
 > delayedAssign("x", msg)
 > msg<- "new!"
 > x #- new!
 > substitute(x) #- x (was 'msg' ?)
 >
 > Here, we just got 'x'... shouldn't we got 'msg'?
 >
 > Same result when the promise is not evaluated yet:
 >
 > delayedAssign("x", msg)
 > substitute(x)
 >
 > In a function, that works:
 >
 > foo<- function (x = msg) substitute(x)
 > foo()
 >
 > Did I misunderstood something? It seems to me that substitute() does not
 > behaves as documented for promises created using delayedAssign().

I don't think this is well documented, but substitute() doesn't act the 
same when its "env" argument is the global environment.  So this works
the way you'd expect:

e <- new.env()
msg <- "old"
delayedAssign("x", msg, assign=e)
msg <- "new"
e$x
substitute(x, e)

I forget what the motivation was for special-casing globalenv().

Duncan Murdoch