Substitute / delayedAssign (was: Substitute unaware when promise objects are evaluated)

Duncan, Thank you for the clarification on how delayedAssign works. Should R-level interfaces to promise objects ever become available, I expect they would at time come in handy.

On the subject of substitute and delayedAssign, I do have a follow-up question for the list. I'm trying to convert a named list of expression objects into an environment of promise objects. After conversion, each expression in the list will be automatically evaluated when the variable with the same name is accessed in the environment. Effectively, I'm trying to create a hash table of promise objects.

Here's the code I wrote that works just fine.

x <- list(a=3, b=expression(a+2), sleep=expression(Sys.sleep(2)))
env <- new.env()
for (i in seq(x)) {
	key <- names(x)[i]
	.Internal(delayedAssign(key,
				   eval(substitute(x[[i]], list(x=x, i=i)))[[1]],
				   eval.env=env, assign.env=env))
}	
env$b     # 3+2
[1] 5
env$sleep # Sleeps for 2 seconds
NULL

The "problem" is that R CMD check complains that I shouldn't be using .Internal() to access the delayedAssign function. However, if I don't use .Internal(), then delayedAssign puts another substitute around my call that prevents the 'i' iterator variable from being evaluated at the correct time, which causes all variables to get the value x[[i]] for the very last value of 'i'.

Can I safely ignore this R CMD check warning about .Internal, or is there a better way to write this code?

Thanks, Robert



-----Original Message-----
From: Duncan Murdoch [mailto:murdoch.duncan at gmail.com]
Sent: Wednesday, May 15, 2013 6:04 PM
To: McGehee, Robert
Cc: R-Devel (r-devel at r-project.org)
Subject: Re: [Rd] Substitute unaware when promise objects are evaluated

On 13-05-15 11:54 AM, McGehee, Robert wrote:

R-devel,
I used the 'substitute' function to create labels for objects inside an environment, without actually evaluating the objects, as the objects might be promises.

However, I was surprised to see that 'substitute' returns the expression slot of the original promise even after the promise has been forcibly evaluated. (Doesn't the promise go away after evaluation?) This behavior probably falls under the "...no guarantee that the resulting expression makes any sense" clause of the ?substitute documentation, but in case there's something actually wrong here, I thought I'd send an example.

I think you misunderstand promises.

A promise has two (or three, depending how you count) parts:  an
expression with an associated environment, and a value.  The value isn't
filled in until the expression is evaluated, but the expression doesn't
go away then.  You can still see it until you change the variable that
holds the promise.

Here's an example showing how the evaluated expression returned by substitute does not match the actual variable value:

env <- new.env()
z <- 0
delayedAssign("var", z+2, assign.env=env)
substitute(var, env=env)

z + 2

The documentation for substitute may not be clear on this, but for a
promise, the env argument will be ignored.  It was the eval.env argument
to delayedAssign that set the promise's environment.

force(env$var)

[1] 2

z <- 10
substitute(var, env=env)

z + 2

eval(substitute(var, env=env))

[1] 12

force(env$var)

[1] 2

Is there any obvious way to code around this behavior, e.g. can I explicitly check if an object in an environment is an unevaluated promise?

Not at R level. In C code you could, but you probably shouldn't.  Think
of promises as values where you can look up the expression that gave the
value, and sometimes delay the calculation until you need it.

Duncan Murdoch

On Thu, May 16, 2013 at 6:06 AM, McGehee, Robert <
Robert.McGehee at geodecapital.com> wrote:

Duncan, Thank you for the clarification on how delayedAssign works. Should
R-level interfaces to promise objects ever become available, I expect they
would at time come in handy.

On the subject of substitute and delayedAssign, I do have a follow-up
question for the list. I'm trying to convert a named list of expression
objects into an environment of promise objects. After conversion, each
expression in the list will be automatically evaluated when the variable
with the same name is accessed in the environment. Effectively, I'm trying
to create a hash table of promise objects.

Populating a new environment with promises happens to be what "calling a
function" in R does anyway, so an elegant way to accomplish this goal is:

makePromiseEnv <- function(expressions, parent=parent.frame()) {
     f <- function() environment()
     formals(f) <- as.pairlist(expressions)
     environment(f) <- parent
     f()
}

e <- makePromiseEnv(alist(a = {print("hello"); 4}, b = {print("again");

6}))

e$a

[1] "hello"
[1] 4

e$a

[1] 4

e$b

[1] "again"
[1] 6

e$b

[1] 6

Duncan, Thank you for the clarification on how delayedAssign works. Should R-level interfaces to promise objects ever become available, I expect they would at time come in handy.

On the subject of substitute and delayedAssign, I do have a follow-up question for the list. I'm trying to convert a named list of expression objects into an environment of promise objects. After conversion, each expression in the list will be automatically evaluated when the variable with the same name is accessed in the environment. Effectively, I'm trying to create a hash table of promise objects.

Here's the code I wrote that works just fine.

x <- list(a=3, b=expression(a+2), sleep=expression(Sys.sleep(2)))
env <- new.env()
for (i in seq(x)) {
	key <- names(x)[i]
	.Internal(delayedAssign(key,
				   eval(substitute(x[[i]], list(x=x, i=i)))[[1]],
				   eval.env=env, assign.env=env))
}	
env$b     # 3+2
[1] 5
env$sleep # Sleeps for 2 seconds
NULL  

The "problem" is that R CMD check complains that I shouldn't be using .Internal() to access the delayedAssign function. However, if I don't use .Internal(), then delayedAssign puts another substitute around my call that prevents the 'i' iterator variable from being evaluated at the correct time, which causes all variables to get the value x[[i]] for the very last value of 'i'.

Can I safely ignore this R CMD check warning about .Internal, or is there a better way to write this code?

Thanks, Robert



-----Original Message-----
From: Duncan Murdoch [mailto:murdoch.duncan at gmail.com] 
Sent: Wednesday, May 15, 2013 6:04 PM
To: McGehee, Robert
Cc: R-Devel (r-devel at r-project.org)
Subject: Re: [Rd] Substitute unaware when promise objects are evaluated

On 13-05-15 11:54 AM, McGehee, Robert wrote:

R-devel,
I used the 'substitute' function to create labels for objects inside an environment, without actually evaluating the objects, as the objects might be promises.

However, I was surprised to see that 'substitute' returns the expression slot of the original promise even after the promise has been forcibly evaluated. (Doesn't the promise go away after evaluation?) This behavior probably falls under the "...no guarantee that the resulting expression makes any sense" clause of the ?substitute documentation, but in case there's something actually wrong here, I thought I'd send an example.

I think you misunderstand promises.

A promise has two (or three, depending how you count) parts:  an 
expression with an associated environment, and a value.  The value isn't 
filled in until the expression is evaluated, but the expression doesn't 
go away then.  You can still see it until you change the variable that 
holds the promise.

Here's an example showing how the evaluated expression returned by substitute does not match the actual variable value:

env <- new.env()
z <- 0
delayedAssign("var", z+2, assign.env=env)
substitute(var, env=env)

z + 2

The documentation for substitute may not be clear on this, but for a 
promise, the env argument will be ignored.  It was the eval.env argument 
to delayedAssign that set the promise's environment.

force(env$var)

[1] 2

z <- 10
substitute(var, env=env)

z + 2

eval(substitute(var, env=env))

[1] 12

force(env$var)

[1] 2

Is there any obvious way to code around this behavior, e.g. can I explicitly check if an object in an environment is an unevaluated promise?

Not at R level. In C code you could, but you probably shouldn't.  Think 
of promises as values where you can look up the expression that gave the 
value, and sometimes delay the calculation until you need it.

Duncan Murdoch