Add support for UCS-2 strings

src/python_data_access.rs

@@ -33,10 +33,11 @@ pub fn copy_string<T: StringObject, P: ProcessMemory>(
             Ok(chars.iter().collect())
         }
         (2, _) => {
-            // UCS2 strings aren't used internally after v3.3: https://www.python.org/dev/peps/pep-0393/
-            // TODO: however with python 2.7 they could be added with --enable-unicode=ucs2 configure flag.
-            //            or with python 3.2 --with-wide-unicode=ucs2
-            Err(format_err!("ucs2 strings aren't supported yet!"))
+            #[allow(clippy::cast_ptr_alignment)]
+            let chars = unsafe {
+                std::slice::from_raw_parts(bytes.as_ptr() as *const u16, bytes.len() / 2)
+            };
+            Ok(String::from_utf16(chars)?)
         }
         (1, true) => Ok(String::from_utf8(bytes)?),
         (1, false) => Ok(bytes.iter().map(|&b| b as char).collect()),

tests/integration_test.rs

@@ -229,6 +229,34 @@ fn test_unicode() {
     assert!(!traces[0].owns_gil);
 }
 
+#[test]
+fn test_cyrillic() {
+    #[cfg(target_os = "macos")]
+    {
+        if unsafe { libc::geteuid() } != 0 {
+            return;
+        }
+    }
+
+    // Identifiers with characters outside the ASCII range are supported from Python 3
+    let runner = TestRunner::new(Config::default(), "./tests/scripts/longsleep.py");
+    if runner.spy.version.major == 2 {
+        return;
+    }
+
+    let mut runner = TestRunner::new(Config::default(), "./tests/scripts/cyrillic.py");
+
+    let traces = runner.spy.get_stack_traces().unwrap();
+    assert_eq!(traces.len(), 1);
+    let trace = &traces[0];
+
+    assert_eq!(trace.frames[0].name, "кириллица");
+    assert_eq!(trace.frames[0].line, 4);
+
+    assert_eq!(trace.frames[1].name, "<module>");
+    assert_eq!(trace.frames[1].line, 7);
+}
+
 #[test]
 fn test_local_vars() {
     #[cfg(target_os = "macos")]

tests/scripts/cyrillic.py

@@ -0,0 +1,7 @@
+import time
+
+def кириллица(seconds):
+    time.sleep(seconds)
+
+if __name__ == "__main__":
+    кириллица(100)