1 files changed, 56 insertions, 8 deletions

diff --git a/Parser/parsetok.c b/Parser/parsetok.c
index bb9afc4..7beb735 100644
--- a/Parser/parsetok.c
+++ b/Parser/parsetok.c
@@ -13,7 +13,7 @@
 
 /* Forward */
 static node *parsetok(struct tok_state *, grammar *, int, perrdetail *, int *);
-static void initerr(perrdetail *err_ret, const char* filename);
+static int initerr(perrdetail *err_ret, const char* filename);
 
 /* Parse input coming from a string.  Return error code, print some errors. */
 node *
@@ -48,7 +48,8 @@ PyParser_ParseStringFlagsFilenameEx(const char *s, const char *filename,
     struct tok_state *tok;
     int exec_input = start == file_input;
 
-    initerr(err_ret, filename);
+    if (initerr(err_ret, filename) < 0)
+        return NULL;
 
     if (*flags & PyPARSE_IGNORE_COOKIE)
         tok = PyTokenizer_FromUTF8(s, exec_input);
@@ -59,7 +60,10 @@ PyParser_ParseStringFlagsFilenameEx(const char *s, const char *filename,
         return NULL;
     }
 
-    tok->filename = filename ? filename : "<string>";
+#ifndef PGEN
+    Py_INCREF(err_ret->filename);
+    tok->filename = err_ret->filename;
+#endif
     return parsetok(tok, g, start, err_ret, flags);
 }
 
@@ -90,13 +94,17 @@ PyParser_ParseFileFlagsEx(FILE *fp, const char *filename,
 {
     struct tok_state *tok;
 
-    initerr(err_ret, filename);
+    if (initerr(err_ret, filename) < 0)
+        return NULL;
 
     if ((tok = PyTokenizer_FromFile(fp, (char *)enc, ps1, ps2)) == NULL) {
         err_ret->error = E_NOMEM;
         return NULL;
     }
-    tok->filename = filename;
+#ifndef PGEN
+    Py_INCREF(err_ret->filename);
+    tok->filename = err_ret->filename;
+#endif
     return parsetok(tok, g, start, err_ret, flags);
 }
 
@@ -216,6 +224,36 @@ parsetok(struct tok_state *tok, grammar *g, int start, perrdetail *err_ret,
     if (err_ret->error == E_DONE) {
         n = ps->p_tree;
         ps->p_tree = NULL;
+
+#ifndef PGEN
+        /* Check that the source for a single input statement really
+           is a single statement by looking at what is left in the
+           buffer after parsing.  Trailing whitespace and comments
+           are OK.  */
+        if (start == single_input) {
+            char *cur = tok->cur;
+            char c = *tok->cur;
+
+            for (;;) {
+                while (c == ' ' || c == '\t' || c == '\n' || c == '\014')
+                    c = *++cur;
+
+                if (!c)
+                    break;
+
+                if (c != '#') {
+                    err_ret->error = E_BADSINGLE;
+                    PyNode_Free(n);
+                    n = NULL;
+                    break;
+                }
+
+                /* Suck up comment. */
+                while (c && c != '\n')
+                    c = *++cur;
+            }
+        }
+#endif
     }
     else
         n = NULL;
@@ -226,7 +264,7 @@ parsetok(struct tok_state *tok, grammar *g, int start, perrdetail *err_ret,
     PyParser_Delete(ps);
 
     if (n == NULL) {
-        if (tok->lineno <= 1 && tok->done == E_EOF)
+        if (tok->done == E_EOF)
             err_ret->error = E_EOF;
         err_ret->lineno = tok->lineno;
         if (tok->buf != NULL) {
@@ -269,14 +307,24 @@ done:
     return n;
 }
 
-static void
+static int
 initerr(perrdetail *err_ret, const char *filename)
 {
     err_ret->error = E_OK;
-    err_ret->filename = filename;
     err_ret->lineno = 0;
     err_ret->offset = 0;
     err_ret->text = NULL;
     err_ret->token = -1;
     err_ret->expected = -1;
+#ifndef PGEN
+    if (filename)
+        err_ret->filename = PyUnicode_DecodeFSDefault(filename);
+    else
+        err_ret->filename = PyUnicode_FromString("<string>");
+    if (err_ret->filename == NULL) {
+        err_ret->error = E_ERROR;
+        return -1;
+    }
+#endif
+    return 0;
 }