Improve REPL compount statement detection.

3 files changed, 48 insertions, 0 deletions

diff --git a/py/lexer.c b/py/lexer.c
index 88bc0a1aae..cd2e05ece0 100644
--- a/py/lexer.c
+++ b/py/lexer.c
@@ -10,6 +10,9 @@
 
 #define TAB_SIZE (8)
 
+// TODO seems that CPython allows NULL byte in the input stream
+// don't know if that's intentional or not, but we don't allow it
+
 struct _py_lexer_t {
     const char *name;           // name of source
     void *stream_data;          // data for stream
diff --git a/py/repl.c b/py/repl.c
new file mode 100644
index 0000000000..f295aff23d
--- /dev/null
+++ b/py/repl.c
@@ -0,0 +1,44 @@
+#include "misc.h"
+#include "repl.h"
+
+bool str_startswith_word(const char *str, const char *head) {
+    int i;
+    for (i = 0; str[i] && head[i]; i++) {
+        if (str[i] != head[i]) {
+            return false;
+        }
+    }
+    return head[i] == '\0' && (str[i] == '\0' || !g_unichar_isalpha(str[i]));
+}
+
+bool py_repl_is_compound_stmt(const char *line) {
+    // compound if line starts with a certain keyword
+    if (
+           str_startswith_word(line, "if")
+        || str_startswith_word(line, "while")
+        || str_startswith_word(line, "for")
+        || str_startswith_word(line, "true")
+        || str_startswith_word(line, "with")
+        || str_startswith_word(line, "def")
+        || str_startswith_word(line, "class")
+        || str_startswith_word(line, "@")
+       ) {
+        return true;
+    }
+
+    // also "compound" if unmatched open bracket
+    int n_paren = 0;
+    int n_brack = 0;
+    int n_brace = 0;
+    for (const char *l = line; *l; l++) {
+        switch (*l) {
+            case '(': n_paren += 1; break;
+            case ')': n_paren -= 1; break;
+            case '[': n_brack += 1; break;
+            case ']': n_brack -= 1; break;
+            case '{': n_brace += 1; break;
+            case '}': n_brace -= 1; break;
+        }
+    }
+    return n_paren > 0 || n_brack > 0 || n_brace > 0;
+}
diff --git a/py/repl.h b/py/repl.h
new file mode 100644
index 0000000000..014e8609b4
--- /dev/null
+++ b/py/repl.h
@@ -0,0 +1 @@
+bool py_repl_is_compound_stmt(const char *line);