rcompilertools
The goal of rcompilertools is access work with R’s bytecode. Not a lot to see at the moment.
Installation
remotes::install_github("dirkschumacher/rcompilertools")Example
This is a basic example which shows you how to solve a common problem:
library(rcompilertools)
some_code <- quote({
x <- 1
for (i in 1:10) x <- x + 1
x
})
tidy_bytecode(some_code)
#> $code
#> # A tibble: 28 x 3
#> op_code op_code_int type
#> <chr> <int> <chr>
#> 1 12 12 OPERAND
#> 2 LDCONST.OP 16 OPERATOR
#> 3 1 1 OPERAND
#> 4 SETVAR.OP 22 OPERATOR
#> 5 3 3 OPERAND
#> 6 POP.OP 4 OPERATOR
#> 7 LDCONST.OP 16 OPERATOR
#> 8 5 5 OPERAND
#> 9 STARTFOR.OP 11 OPERATOR
#> 10 9 9 OPERAND
#> # … with 18 more rows
#>
#> $constant_pool
#> $constant_pool[[1]]
#> {
#> x <- 1
#> for (i in 1:10) x <- x + 1
#> x
#> }
#>
#> $constant_pool[[2]]
#> [1] 1
#>
#> $constant_pool[[3]]
#> x <- 1
#>
#> $constant_pool[[4]]
#> x
#>
#> $constant_pool[[5]]
#> x <- 1
#>
#> $constant_pool[[6]]
#> [1] 1 2 3 4 5 6 7 8 9 10
#>
#> $constant_pool[[7]]
#> 1:10
#>
#> $constant_pool[[8]]
#> for (i in 1:10) x <- x + 1
#>
#> $constant_pool[[9]]
#> i
#>
#> $constant_pool[[10]]
#> for (i in 1:10) x <- x + 1
#>
#> $constant_pool[[11]]
#> x + 1
#>
#> $constant_pool[[12]]
#> x <- x + 1
#>
#> $constant_pool[[13]]
#> x
#>
#> $constant_pool[[14]]
#> [1] NA 1 1 4 4 4 6 6 9 9 9 9 3 3 1 1 10 10 11 11 9 9 9 9 9
#> [26] 3 3 3
#> attr(,"class")
#> [1] "expressionsIndex"
#>
#> $constant_pool[[15]]
#> [1] NA 2 2 2 2 2 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7
#> [26] 12 12 12
#> attr(,"class")
#> [1] "srcrefsIndex"
#>
#>
#> attr(,"class")
#> [1] "tidy_bytecode"You can also get all OP codes
opcodes()
#> # A tibble: 129 x 3
#> name args code
#> <chr> <int> <int>
#> 1 BCMISMATCH.OP 0 0
#> 2 RETURN.OP 0 1
#> 3 GOTO.OP 1 2
#> 4 BRIFNOT.OP 2 3
#> 5 POP.OP 0 4
#> 6 DUP.OP 0 5
#> 7 PRINTVALUE.OP 0 6
#> 8 STARTLOOPCNTXT.OP 2 7
#> 9 ENDLOOPCNTXT.OP 1 8
#> 10 DOLOOPNEXT.OP 0 9
#> # … with 119 more rowsYou can also evaluate the byte code
eval_code(tidy_bytecode(some_code))
#> [1] 11