Merge pull request #8748 from rstudio/bugfix/repl-python-unicode-windows

use Python-compatible unicode escapes
rstudio · Jan 19, 2021 · 3f69f20 · 3f69f20
2 parents bb4cb50 + 84b76df
commit 3f69f20
Showing 1 changed file with 10 additions and 1 deletion.
diff --git a/src/cpp/core/StringUtils.cpp b/src/cpp/core/StringUtils.cpp
@@ -287,9 +287,18 @@ std::string utf8ToSystem(const std::string& str,
       if (n == -1)
       {
          if (escapeInvalidChars)
-            output << "\\u{" << std::hex << wide[i] << "}";
+         {
+            // NOTE: in R, both '\u{1234}' and '\u1234' are valid
+            // ways of specifying a unicode literal, but only the
+            // latter is accepted by Python, and since the reticulate
+            // REPL uses the same conversion routines we prefer the
+            // format compatible with both parsers
+            output << "\\u" << std::hex << wide[i];
+         }
          else
+         {
             output << "?"; // TODO: Use GetCPInfo()
+         }
       }
       else
       {