Require stable rlang

DESCRIPTION

@@ -15,7 +15,7 @@ BugReports: https://github.com/egnha/gestalt/issues
 Depends:
     R (>= 3.3.0)
 Imports:
-    rlang (>= 0.99.0.9000),
+    rlang (>= 1.0.0),
     utils
 Suggests:
     magrittr (>= 1.5),

NEWS.md

@@ -2,9 +2,9 @@
 
 ## 0.1.9
 
-This maintenance release fixes test failures caused by changes in version 1.0.0
-of the rlang package (namely, `rlang::is_expression()`). There are no
-user-facing changes.
+Gestalt now depends on a stable release of rlang, version 1.0.0 and above.
+(A minor internal fix was made to address a change in the behavior of
+`rlang::is_expression()`.) There are no user-facing changes.
 
 ## 0.1.8
 

inst/NEWS

@@ -1,5 +1,43 @@
 # gestalt
 
+## 0.1.9
+
+Gestalt now depends on a stable release of rlang, version 1.0.0 and above.
+(A minor internal fix was made to address a change in the behavior of
+`rlang::is_expression()`.) There are no user-facing changes.
+
+## 0.1.8
+
+  * The environment of a partial function expression in a `%>>>%` chain
+    (e.g., the base-3 logarithm in `abs %>>>% log(base = 3)`) is now properly
+    captured. Previously, it was erroneously matched to an rlang data mask, due
+    to an internal call to `rlang::eval_tidy()` using positional arguments.
+
+## 0.1.7
+
+  * In a `%>>>%` chain, a point (`.`) is now only matched as an argument value
+    when it is a symbol, not a character (`"."`) (#27).
+
+## 0.1.6
+
+This is a maintenance release to fix test failures caused by changes in the
+rlang package.
+
+ * The (minimum) required rlang version has been increased to 0.3.1. This
+   version fixed a bug which prevented certain operator "sections" from being
+   expressed. A chain like `` `/`(2) %>>>% sin`` (halve and apply sine) now
+   works as expected.
+
+ * Formals of primitive functions now agree with those of `base::args()` (#18, #24).
+   This means you can use `args()` to determine the names of arguments when
+   using `partial()`. Thus, `` partial(`/`, e2 = 3) `` is the same as
+   `` partial(`/`, , 3) `` is the same as division-by-3. Moreover, `%>>>%`
+   chains are verified against the argument names given by `args()`. Thus,
+   `` `/`(e2 = 2) %>>>% sin `` is valid, but `` `/`(y = 2) %>>>% sin `` is
+   invalid—`` `/`() `` viewed as a closure has no argument called `y`.
+
+ * Support for R 3.1 has been dropped.
+
 ## 0.1.5
 
   * Fixed a segfault caused by leakage of rlang internals (thanks @lionel-).