rlang 0.2.0

This release of rlang is mostly an effort at polishing the tidy evaluation framework. All tidy eval functions and operators have been rewritten in C in order to improve performance. Capture of expression, quasiquotation, and evaluation of quosures are now vastly faster. On the UI side, many of the inconveniences that affected the first release of rlang have been solved:

See the first section below for a complete list of changes to the tidy evaluation framework.

This release also polishes the rlang API. Many functions have been renamed as we get a better feel for the consistency and clarity of the API. Note that rlang as a whole is still maturing and some functions are even experimental. In order to make things clearer for users of rlang, we have started to develop a set of conventions to document the current stability of each function. You will now find “lifecycle” sections in documentation topics. In addition we have gathered all lifecycle information in the ?rlang::lifecycle help page. Please only use functions marked as stable in your projects unless you are prepared to deal with occasional backward incompatible updates.

Tidy evaluation

Technically the !! operator has the same precedence as unary - and +. This means that !!a:b and !!a + b are equivalent to (!!a):b and (!!a) + b. On the other hand !!a^b and !!a$b are equivalent to!!(a^b) and !!(a$b).

In addition, enexprs() is like enquos() but like exprs() it returns bare expressions. And ensyms() expects strings or symbols.

select_one <- function(df, var) { df %>% dplyr::select(!!enquo(var)) }

Technically, this is because enquo() now also captures arguments in parents of the current environment rather than just in the current environment. The flip side of this increased flexibility is that if you made a typo in the name of the variable you want to capture, and if an object of that name exists anywhere in the parent contexts, you will capture that object rather than getting an error.

expr(`!!`(var)) quo(call(`!!!`(var)))

This is consistent with the way native R operators parses to function calls. These new functional forms are to be preferred to UQ() and UQS(). We are now questioning the latter and might deprecate them in a future release.



Various features


API changes

The rlang API is maturing and still in flux. However we have made an effort to better communicate what parts are stable. We will not introduce breaking changes for stable functions unless the payoff for the change is worth the trouble. See ?rlang::lifecycle for the lifecycle status of exported functions.

In rlang 0.1 calls were called “language” objects in order to follow the R type nomenclature as returned by base::typeof(). We wanted to avoid adding to the confusion between S modes and R types. With hindsight we find it is better to use more meaningful type names.

Following this new terminology, the new functions as_data_mask() and new_data_mask() replace as_overscope() and new_overscope(). as_data_mask() has also a more consistent interface. These functions are only meant for developers of tidy evaluation interfaces.

Breaking changes

Upcoming breaking changes

In addition, lang_head() and lang_tail() are soft-deprecated without replacement because these are low level accessors that are rarely needed.

rlang 0.1.6

rlang 0.1.4

rlang 0.1.2

This hotfix release makes rlang compatible with the R 3.1 branch.

rlang 0.1.1

This release includes two important fixes for tidy evaluation:

New functions:

UI improvements:


rlang 0.1.0

Initial release.