Attempt at sha1.formula Method

[billdenney]

I just tried to write an sha1.formula method (as mentioned in #23), but I couldn't get the test to work. My attempts were:

Function:

sha1.formula <- function(x, digits = 14L, zapsmall = 7L, ..., algo = "sha1"){
    dots <- list(...)
    if (is.null(dots$environment)) {
        dots$environment <- TRUE
    }
    if (isTRUE(dots$environment)) {
        y <- append(
            as.list(x),
            list(environment = digest(environment(x), algo = algo))
        )
    } else {
        y <- as.list(x)
    }
    y <- vapply(
        y,
        sha1,
        digits = digits,
        zapsmall = zapsmall,
        environment = dots$environment,
        ... = dots,
        algo = algo,
        FUN.VALUE = NA_character_
    )
    attr(y, "digest::sha1") <- attr_sha1(
        x = y, digits = digits, zapsmall = zapsmall, algo = algo, dots
    )
    digest(y, algo = algo)
}

Test:

x.formula <- a~b+c|d
expect_true(
    identical(
        sha1(x.formula, environment=FALSE),
        {
            y <- sapply(
                X=as.list(x.formula),
                FUN=sha1
            )
            attr(y, "digest::sha1") <- list(
                digits=14L,
                zapsmall=7L,
                algo="sha1"
            )
            sha1(y)
        }
    )
)

Edited to add the formula being tested.

[billdenney]

I've traced my issue to the fact that the right hand side of the formula which becomes a call evaluates to a different result when inside or outside of the function. That appears to be an issue with sha1.call() (and within that, digest) giving a different result. I will try to make a reproducible example and open a new issue for that.

Nevermind, this was an issue where I wasn't setting the attributes correctly.

[billdenney]

I have found my issue, and I'm finalizing the PR now.