Merge pull request #478 from lionel-/release-0-2-0

Prepare release 0.2.0

DESCRIPTION

@@ -1,11 +1,19 @@
 Package: vctrs
 Title: Vector Helpers
-Version: 0.1.99.9000
+Version: 0.2.0
 Authors@R: 
     c(person(given = "Hadley",
              family = "Wickham",
              role = c("aut", "cre"),
              email = "[email protected]"),
+      person(given = "Lionel",
+             family = "Henry",
+             role = "aut",
+             email = "[email protected]"),
+      person(given = "Davis",
+             family = "Vaughan",
+             role = "aut",
+             email = "[email protected]"),
       person(given = "RStudio",
              role = "cph"))
 Description: Defines new notions of prototype and size that are

NEWS.md

@@ -1,5 +1,5 @@
 
-# vctrs 0.1.99.9000
+# vctrs 0.2.0
 
 With the 0.2.0 release, many vctrs functions have been rewritten with
 native C code to improve performance. Functions like `vec_c()` and
@@ -19,13 +19,11 @@ or possible cast.
 
 ## Breaking changes
 
-* We have renamed the `type` particle to `ptype`:
-
-  - `vec_type()` => `vec_ptype()`
-  - `vec_type2()` => `vec_ptype2()`
-  - `vec_type_common()` => `vec_ptype_common()`
-
-  Consequently, `vec_ptype()` was renamed to `vec_ptype_show()`.
+* Lossy casts now throw errors of type `vctrs_error_cast_lossy`.
+  Previously these were warnings. You can suppress these errors
+  selectively with `allow_lossy_cast()` to get the partial cast
+  results. To implement your own lossy cast operation, call the new
+  exported function `maybe_lossy_cast()`.
 
 * `vec_c()` now fails when an input is supplied with a name but has
   internal names or is length > 1:
@@ -52,21 +50,24 @@ or possible cast.
   vec_c(foo = c(a = 1), .name_spec = ~ paste(.x, .y, sep = "_"))
   ```
 
+* `vec_empty()` has been renamed to `vec_is_empty()`.
+
+* `vec_dim()` and `vec_dims()` are no longer exported.
+
 * `vec_na()` has been renamed to `vec_init()`, as the primary use case
   is to initialize an output container.
 
 * `vec_slice<-` is now type stable (#140). It always returns the same
   type as the LHS. If needed, the RHS is cast to the correct type, but
   only if both inputs are coercible. See examples in `?vec_slice`.
 
-* Lossy casts now throw errors of type `vctrs_error_cast_lossy`.  You
-  can suppress these errors selectively with `allow_lossy_cast()` to
-  get the partial cast results. To implement your own lossy cast
-  operation, call the new exported function `maybe_lossy_cast()`.
+* We have renamed the `type` particle to `ptype`:
 
-* `vec_empty()` has been renamed to `vec_is_empty()`.
+  - `vec_type()` => `vec_ptype()`
+  - `vec_type2()` => `vec_ptype2()`
+  - `vec_type_common()` => `vec_ptype_common()`
 
-* `vec_dim()` and `vec_dims()` are no longer exported.
+  Consequently, `vec_ptype()` was renamed to `vec_ptype_show()`.
 
 
 ## New features
@@ -96,17 +97,28 @@ or possible cast.
   If you inherit from `vctrs_vctr` or `vctrs_rcrd` you don't need to
   implement `vec_proxy()`.
 
-* `vec_slice()` now supports row names.
+* `vec_c()`, `vec_rbind()`, and `vec_cbind()` gain a `.name_repair`
+  argument (#227, #229).
 
-* `list_of()` now has an `as.character()` method. It uses
-  `vec_ptype_abbr()` to collapse complex objects into their type
-  representation (tidyverse/tidyr#654).
+* `vec_c()`, `vec_rbind()`, `vec_cbind()`, and all functions relying
+  on `vec_ptype_common()` now have more informative error messages
+  when some of the inputs have nested data frames that are not
+  convergent:
+
+  ```
+  df1 <- tibble(foo = tibble(bar = tibble(x = 1:3, y = letters[1:3])))
+  df2 <- tibble(foo = tibble(bar = tibble(x = 1:3, y = 4:6)))
+
+  vec_rbind(df1, df2)
+  #> Error: No common type for `..1$foo$bar$y` <character> and `..2$foo$bar$y` <integer>.
+  ```
 
 * `vec_cbind()` now turns named data frames to packed columns.
 
   ```r
   data <- tibble::tibble(x = 1:3, y = letters[1:3])
   data <- vec_cbind(data, packed = data)
+  data
   # A tibble: 3 x 3
         x y     packed$x $y
     <int> <chr>    <int> <chr>
@@ -131,86 +143,87 @@ or possible cast.
   We are planning to use this syntax more widely in the tidyverse.
 
 * New `vec_is()` function to check whether a vector conforms to a
-  protopty and/or a size. Unlike `vec_assert()`, it doesn't throw
+  prototype and/or a size. Unlike `vec_assert()`, it doesn't throw
   errors but returns `TRUE` or `FALSE` (#79).
 
-  Called without a specific type or size, it tests whether an object
-  is a data vector or a scalar. S3 lists are treated as scalars by
-  default. Implement a `vec_is_vector()` for your class to override
-  this property (or derive from `vctrs_vctr`).
+  Called without a specific type or size, `vec_assert()` tests whether
+  an object is a data vector or a scalar. S3 lists are treated as
+  scalars by default. Implement a `vec_is_vector()` for your class to
+  override this property (or derive from `vctrs_vctr`).
 
-* New `vec_split(x, by)` is a generalisation of `split()` that can divide
-  a vector into groups formed by the unique values of another vector. Returns
-  a two-column data frame containing unique values of `by` aligned with
-  matching `x` values (#196).
-
-* `vec_init()` checks input is a vector.
+* New `vec_order()` and `vec_sort()` for ordering and sorting
+  generalised vectors.
 
 * New `.names_to` parameter for `vec_rbind()`. If supplied, this
   should be the name of a column where the names of the inputs are
   copied. This is similar to the `.id` parameter of
   `dplyr::bind_rows()`.
 
-* `vec_c()`, `vec_rbind()`, and `vec_cbind()` gain a `.name_repair`
-  argument (#227, #229).
+* New `vec_seq_along()` and `vec_init_along()` create useful sequences (#189).
+
+* `vec_slice()` now preserves character row names, if present.
+
+* New `vec_split(x, by)` is a generalisation of `split()` that can divide
+  a vector into groups formed by the unique values of another vector. Returns
+  a two-column data frame containing unique values of `by` aligned with
+  matching `x` values (#196).
 
 
 ## Other features and bug fixes
 
+* Using classed errors of class `"vctrs_error_assert"` for failed
+  assertions, and of class `"vctrs_error_incompatible"` (with
+  subclasses `_type`, `_cast` and `_op`) for errors on incompatible
+  types (#184).
+
+* Character indexing is now only supported for named objects, an error
+  is raised for unnamed objects (#171).
+
 * Predicate generics now consistently return logical vectors when
   passed a `vctrs_vctr` class. They used to restore the output to
   their input type (#251).
 
+* `list_of()` now has an `as.character()` method. It uses
+  `vec_ptype_abbr()` to collapse complex objects into their type
+  representation (tidyverse/tidyr#654).
+
 * New `stop_incompatible_size()` to signal a failure due to mismatched sizes.
 
-* `vec_slice<-` now has a `vec_assign()` alias. Use `vec_assign()`
-  when you don't want to modify the original input.
+* New `validate_list_of()` (#193).
 
-* `vec_slice()` now calls `vec_restore()` automatically. Unlike the
-  default `[` method from base R, attributes are preserved by default.
+* `vec_arith()` is consistent with base R when combining `difftime`
+  and `date`, with a warning if casts are lossy (#192).
 
-* `vec_size()` now works with positive short row names. This fixes
-  issues with data frames created with jsonlite (#220).
+* `vec_c()` and `vec_rbind()` now handle data.frame columns properly
+  (@yutannihilation, #182).
 
-* New `validate_list_of()` (#193).
+* `vec_cast(x, data.frame())` preserves the number of rows in `x`.
 
 * `vec_equal()` now handles missing values symmetrically (#204).
 
+* `vec_equal_na()` now returns `TRUE` for data frames and records when
+  every component is missing, not when _any_ component is missing
+  (#201).
+
+* `vec_init()` checks input is a vector.
+
 * `vec_proxy_compare()` gains an experimental `relax` argument, which
   allows data frames to be orderable even if all their columns are not
   (#210).
 
-* New `vec_order()` and `vec_sort()` for ordering and sorting
-  generalised vectors.
+* `vec_size()` now works with positive short row names. This fixes
+  issues with data frames created with jsonlite (#220).
 
-* New `vec_repeat()` for repeating each element of a vector the same number
-  of times.
+* `vec_slice<-` now has a `vec_assign()` alias. Use `vec_assign()`
+  when you don't want to modify the original input.
 
-* `vec_slice()` can correct slice 0-row data frames (#179).
+* `vec_slice()` now calls `vec_restore()` automatically. Unlike the
+  default `[` method from base R, attributes are preserved by default.
 
-* New `vec_seq_along()` and `vec_init_along()` create useful sequences (#189).
+* `vec_slice()` can correct slice 0-row data frames (#179).
 
-* `vec_equal_na()` now returns `TRUE` for data frames and records when
-  every component is mising, not when _any_ component is missing
-  (#201).
+* New `vec_repeat()` for repeating each element of a vector the same number
+  of times.
 
 * `vec_type2(x, data.frame())` ensures that the returned object has
   names that are a length-0 character vector.
-
-* `vec_cast(x, data.frame())` preserves the number of rows in the `x`.
-
-* Added a `NEWS.md` file to track changes to the package.
-
-* Using classed errors of class `"vctrs_error_assert"` for failed
-  assertions, and of class `"vctrs_error_incompatible"` (with
-  subclasses `_type`, `_cast` and `_op`) for errors on incompatible
-  types (#184).
-
-* Character indexing is now only supported for named objects, an error
-  is raised for unnamed objects (#171).
-
-* `vec_c()` and `vec_rbind()` now handle data.frame columns properly
-  (@yutannihilation, #182).
-
-* `vec_arith()` is consistent with base R when combining `difftime`
-  and `date`, with a warning if casts are lossy (#192).

R/type-unspecified.R

@@ -1,6 +1,6 @@
 #' A 1d vector of unspecified type
 #'
-#' This is a [partial type](new_partial) used to represent logical vectors
+#' This is a [partial type][new_partial] used to represent logical vectors
 #' that only contain `NA`. These require special handling because we want to
 #' allow `NA` to specify missingness without requiring a type.
 #'

R/type.R

@@ -9,7 +9,7 @@
 #' successively calls [vec_ptype2()] to find a common type.
 #'
 #' @section Prototype:
-#' A prototype is [size](vec_size) 0 vector containing attributes, but no
+#' A prototype is [size][vec_size] 0 vector containing attributes, but no
 #' data. Generally, this is just `vec_slice(x, 0L)`, but some inputs
 #' require special handling.
 #'

_pkgdown.yml

@@ -6,18 +6,43 @@ reference:
   - vec_data
   - vec_ptype
   - vec_size
+  - vec_is
+  - vec_assert
 
 - title: Casting and combining
   contents:
   - vec_c
   - vec_cbind
 
-- title: Matching
+- title: Slicing and recycling
+  contents:
+  - vec_init
+  - vec_slice
+  - vec_assign
+  - vec_recycle
+
+- title: Equality and ordering
+  contents:
+  - vec_equal
+  - vec_compare
+
+- title: Sorting
+  contents:
+  - vec_sort
+  - vec_order
+
+- title: Matching and splitting
   contents:
   - vec_count
   - vec_duplicate
   - vec_unique
   - vec_in
+  - vec_split
+
+- title: Sequences and repetitions
+  contents:
+  - vec_repeat
+  - vec_seq_along
 
 - title: New classes
   contents:
@@ -28,12 +53,20 @@ reference:
   - vctr
   - vec_cast
   - vec_list_cast
-  - vec_type2
-  - vec_grp_compare
+  - vec_ptype2
+  - vec_proxy_compare
   - vec_proxy_equal
   - vec_ptype_full
+  - vec_ptype_finalise
   - "vctrs-conditions"
   - rcrd
   - fields
   - s3_register
   - vec_resize
+  - vec_arith
+  - vec_math
+  - unspecified
+  - vec_default_cast
+  - vec_default_ptype2
+  - vec_as_names
+  - vec_as_index

cran-comments.md

@@ -1,15 +1,17 @@
-This is a resubmission, fixing an obvious problem in the description.
-
----
 
 ## Test environments
-* local OS X install, R 3.5.1
-* ubuntu 14.04 (on travis-ci), R 3.5.1
-* win-builder (devel and release)
+
+* local macOS: release
+* Travis Ubuntu: 3.2, 3.3, oldrel, release, devel
+* win-builder: release, devel
+* R-Hub: UBSAN, rchk, and valgrind builds
 
 ## R CMD check results
 
-0 errors | 0 warnings | 1 note
+0 errors | 0 warnings | 0 notes
+
+There are 3 reverse dependencies, two of which are broken by this
+vctrs release (evaluator and probably).
 
-* This is a new release.
- 
+- We will send an update of {probably} shortly.
+- We have notified the maintainer of {evaluator} of the failure.