Add unicode option to as.yaml (fixes #4)

vubiostat · Jan 17, 2013 · 3774e67 · 3774e67
1 parent 29e2022
commit 3774e67
Show file tree

Hide file tree

Showing 4 changed files with 22 additions and 7 deletions.
diff --git a/pkg/R/as.yaml.R b/pkg/R/as.yaml.R
@@ -1,5 +1,5 @@
 `as.yaml` <-
-function(x, line.sep = c('\n', '\r\n', '\r'), indent = 2, omap = FALSE, column.major = TRUE) {
+function(x, line.sep = c('\n', '\r\n', '\r'), indent = 2, omap = FALSE, column.major = TRUE, unicode = TRUE) {
   line.sep <- match.arg(line.sep)
-  .Call("as.yaml", x, line.sep, indent, omap, column.major, PACKAGE="yaml")
+  .Call("as.yaml", x, line.sep, indent, omap, column.major, unicode, PACKAGE="yaml")
 }
diff --git a/pkg/inst/tests/test_as_yaml.R b/pkg/inst/tests/test_as_yaml.R
@@ -132,3 +132,9 @@ test_that("test should escape strings", {
   result <- as.yaml("12345")
   expect_equal("'12345'\n", result)
 })
+
+test_that("unicode strings are not escaped", {
+  x <- list('име' = 'Александар', 'презиме' = 'Благотић')
+  result <- as.yaml(x, unicode = TRUE)
+  expect_equal("име: Александар\nпрезиме: Благотић\n", result, label = result)
+})
diff --git a/pkg/man/as.yaml.Rd b/pkg/man/as.yaml.Rd
@@ -5,14 +5,15 @@
   Convert an R object into a YAML string
 }
 \usage{
-  as.yaml(x, line.sep = c("\n", "\r\n", "\r"), indent = 2, omap = FALSE, column.major = TRUE)
+  as.yaml(x, line.sep = c("\n", "\r\n", "\r"), indent = 2, omap = FALSE, column.major = TRUE, unicode = TRUE)
 }
 \arguments{
   \item{x}{ the object to be converted }
   \item{line.sep}{ the line separator character(s) to use }
   \item{indent}{ the number of spaces to use for indenting }
   \item{omap}{ determines whether or not to convert a list to a YAML omap; see Details }
   \item{column.major}{ determines how to convert a data.frame; see Details }
+  \item{unicode}{ determines whether or not to allow unescaped unicode characters in output }
 }
 \details{
   If you set the \code{omap} option to TRUE, as.yaml will create ordered maps

diff --git a/pkg/src/r-ext.c b/pkg/src/r-ext.c
@@ -1729,18 +1729,19 @@ emit_object(emitter, event, obj, tag, omap, column_major)
 }
 
 SEXP
-as_yaml(s_obj, s_line_sep, s_indent, s_omap, s_column_major)
+as_yaml(s_obj, s_line_sep, s_indent, s_omap, s_column_major, s_unicode)
   SEXP s_obj;
   SEXP s_line_sep;
   SEXP s_indent;
   SEXP s_omap;
   SEXP s_column_major;
+  SEXP s_unicode;
 {
   SEXP retval;
   yaml_emitter_t emitter;
   yaml_event_t event;
   s_emitter_output output;
-  int status, line_sep, indent, omap, column_major;
+  int status, line_sep, indent, omap, column_major, unicode;
   const char *c_line_sep;
 
   c_line_sep = CHAR(STRING_ELT(s_line_sep, 0));
@@ -1789,7 +1790,14 @@ as_yaml(s_obj, s_line_sep, s_indent, s_omap, s_column_major)
   }
   column_major = LOGICAL(s_column_major)[0];
 
+  if (!isLogical(s_unicode) || length(s_unicode) != 1) {
+    error("argument `unicode` must be either TRUE or FALSE");
+    return R_NilValue;
+  }
+  unicode = LOGICAL(s_unicode)[0];
+
   yaml_emitter_initialize(&emitter);
+  yaml_emitter_set_unicode(&emitter, unicode);
   yaml_emitter_set_break(&emitter, line_sep);
   yaml_emitter_set_indent(&emitter, indent);
 
@@ -1798,7 +1806,7 @@ as_yaml(s_obj, s_line_sep, s_indent, s_omap, s_column_major)
   yaml_emitter_set_output(&emitter, as_yaml_write_handler, &output);
 
   /* FIXME: get the encoding from R */
-  yaml_stream_start_event_initialize(&event, YAML_UTF8_ENCODING);
+  yaml_stream_start_event_initialize(&event, YAML_ANY_ENCODING);
   if (!(status = yaml_emitter_emit(&emitter, &event)))
     goto done;
 
@@ -1842,7 +1850,7 @@ as_yaml(s_obj, s_line_sep, s_indent, s_omap, s_column_major)
 
 R_CallMethodDef callMethods[] = {
   {"yaml.load", (DL_FUNC)&load_yaml_str, 3},
-  {"as.yaml",   (DL_FUNC)&as_yaml,       5},
+  {"as.yaml",   (DL_FUNC)&as_yaml,       6},
   {NULL, NULL, 0}
 };