bpo-37050: Remove expr_text from FormattedValue ast node, use Constant node instead (GH-13597)

When using the "=" debug functionality of f-strings, use another Constant node (or a merged constant node) instead of adding expr_text to the FormattedValue node.

4 files changed, 52 insertions, 89 deletions

diff --git a/Python/Python-ast.c b/Python/Python-ast.c
index 39a40ee..7c8e438 100644
--- a/Python/Python-ast.c
+++ b/Python/Python-ast.c
@@ -314,12 +314,10 @@ static char *Call_fields[]={
 static PyTypeObject *FormattedValue_type;
 _Py_IDENTIFIER(conversion);
 _Py_IDENTIFIER(format_spec);
-_Py_IDENTIFIER(expr_text);
 static char *FormattedValue_fields[]={
     "value",
     "conversion",
     "format_spec",
-    "expr_text",
 };
 static PyTypeObject *JoinedStr_type;
 static char *JoinedStr_fields[]={
@@ -954,7 +952,7 @@ static int init_types(void)
     Call_type = make_type("Call", expr_type, Call_fields, 3);
     if (!Call_type) return 0;
     FormattedValue_type = make_type("FormattedValue", expr_type,
-                                    FormattedValue_fields, 4);
+                                    FormattedValue_fields, 3);
     if (!FormattedValue_type) return 0;
     JoinedStr_type = make_type("JoinedStr", expr_type, JoinedStr_fields, 1);
     if (!JoinedStr_type) return 0;
@@ -2253,9 +2251,9 @@ Call(expr_ty func, asdl_seq * args, asdl_seq * keywords, int lineno, int
 }
 
 expr_ty
-FormattedValue(expr_ty value, int conversion, expr_ty format_spec, string
-               expr_text, int lineno, int col_offset, int end_lineno, int
-               end_col_offset, PyArena *arena)
+FormattedValue(expr_ty value, int conversion, expr_ty format_spec, int lineno,
+               int col_offset, int end_lineno, int end_col_offset, PyArena
+               *arena)
 {
     expr_ty p;
     if (!value) {
@@ -2270,7 +2268,6 @@ FormattedValue(expr_ty value, int conversion, expr_ty format_spec, string
     p->v.FormattedValue.value = value;
     p->v.FormattedValue.conversion = conversion;
     p->v.FormattedValue.format_spec = format_spec;
-    p->v.FormattedValue.expr_text = expr_text;
     p->lineno = lineno;
     p->col_offset = col_offset;
     p->end_lineno = end_lineno;
@@ -3507,11 +3504,6 @@ ast2obj_expr(void* _o)
         if (_PyObject_SetAttrId(result, &PyId_format_spec, value) == -1)
             goto failed;
         Py_DECREF(value);
-        value = ast2obj_string(o->v.FormattedValue.expr_text);
-        if (!value) goto failed;
-        if (_PyObject_SetAttrId(result, &PyId_expr_text, value) == -1)
-            goto failed;
-        Py_DECREF(value);
         break;
     case JoinedStr_kind:
         result = PyType_GenericNew(JoinedStr_type, NULL, NULL);
@@ -7169,7 +7161,6 @@ obj2ast_expr(PyObject* obj, expr_ty* out, PyArena* arena)
         expr_ty value;
         int conversion;
         expr_ty format_spec;
-        string expr_text;
 
         if (_PyObject_LookupAttrId(obj, &PyId_value, &tmp) < 0) {
             return 1;
@@ -7210,22 +7201,8 @@ obj2ast_expr(PyObject* obj, expr_ty* out, PyArena* arena)
             if (res != 0) goto failed;
             Py_CLEAR(tmp);
         }
-        if (_PyObject_LookupAttrId(obj, &PyId_expr_text, &tmp) < 0) {
-            return 1;
-        }
-        if (tmp == NULL || tmp == Py_None) {
-            Py_CLEAR(tmp);
-            expr_text = NULL;
-        }
-        else {
-            int res;
-            res = obj2ast_string(tmp, &expr_text, arena);
-            if (res != 0) goto failed;
-            Py_CLEAR(tmp);
-        }
-        *out = FormattedValue(value, conversion, format_spec, expr_text,
-                              lineno, col_offset, end_lineno, end_col_offset,
-                              arena);
+        *out = FormattedValue(value, conversion, format_spec, lineno,
+                              col_offset, end_lineno, end_col_offset, arena);
         if (*out == NULL) goto failed;
         return 0;
     }
diff --git a/Python/ast.c b/Python/ast.c
index 6259827..7ffdf4a 100644
--- a/Python/ast.c
+++ b/Python/ast.c
@@ -5006,10 +5006,16 @@ fstring_parse(const char **str, const char *end, int raw, int recurse_lvl,
    closing brace doesn't match an opening paren, for example. It
    doesn't need to error on all invalid expressions, just correctly
    find the end of all valid ones. Any errors inside the expression
-   will be caught when we parse it later. */
+   will be caught when we parse it later.
+
+   *expression is set to the expression.  For an '=' "debug" expression,
+   *expr_text is set to the debug text (the original text of the expression,
+   *including the '=' and any whitespace around it, as a string object).  If
+   *not a debug expression, *expr_text set to NULL. */
 static int
 fstring_find_expr(const char **str, const char *end, int raw, int recurse_lvl,
-                  expr_ty *expression, struct compiling *c, const node *n)
+                  PyObject **expr_text, expr_ty *expression,
+                  struct compiling *c, const node *n)
 {
     /* Return -1 on error, else 0. */
 
@@ -5020,9 +5026,6 @@ fstring_find_expr(const char **str, const char *end, int raw, int recurse_lvl,
     int conversion = -1; /* The conversion char.  Use default if not
                             specified, or !r if using = and no format
                             spec. */
-    int equal_flag = 0; /* Are we using the = feature? */
-    PyObject *expr_text = NULL; /* The text of the expression, used for =. */
-    const char *expr_text_end;
 
     /* 0 if we're not in a string, else the quote char we're trying to
        match (single or double quote). */
@@ -5198,7 +5201,6 @@ fstring_find_expr(const char **str, const char *end, int raw, int recurse_lvl,
        expr_text. */
     if (**str == '=') {
         *str += 1;
-        equal_flag = 1;
 
         /* Skip over ASCII whitespace.  No need to test for end of string
            here, since we know there's at least a trailing quote somewhere
@@ -5206,7 +5208,14 @@ fstring_find_expr(const char **str, const char *end, int raw, int recurse_lvl,
         while (Py_ISSPACE(**str)) {
             *str += 1;
         }
-        expr_text_end = *str;
+
+        /* Set *expr_text to the text of the expression. */
+        *expr_text = PyUnicode_FromStringAndSize(expr_start, *str-expr_start);
+        if (!*expr_text) {
+            goto error;
+        }
+    } else {
+        *expr_text = NULL;
     }
 
     /* Check for a conversion char, if present. */
@@ -5227,17 +5236,6 @@ fstring_find_expr(const char **str, const char *end, int raw, int recurse_lvl,
         }
 
     }
-    if (equal_flag) {
-        Py_ssize_t len = expr_text_end - expr_start;
-        expr_text = PyUnicode_FromStringAndSize(expr_start, len);
-        if (!expr_text) {
-            goto error;
-        }
-        if (PyArena_AddPyObject(c->c_arena, expr_text) < 0) {
-            Py_DECREF(expr_text);
-            goto error;
-        }
-    }
 
     /* Check for the format spec, if present. */
     if (*str >= end)
@@ -5261,16 +5259,16 @@ fstring_find_expr(const char **str, const char *end, int raw, int recurse_lvl,
     assert(**str == '}');
     *str += 1;
 
-    /* If we're in = mode, and have no format spec and no explict conversion,
-       set the conversion to 'r'. */
-    if (equal_flag && format_spec == NULL && conversion == -1) {
+    /* If we're in = mode (detected by non-NULL expr_text), and have no format
+       spec and no explict conversion, set the conversion to 'r'. */
+    if (*expr_text && format_spec == NULL && conversion == -1) {
         conversion = 'r';
     }
 
     /* And now create the FormattedValue node that represents this
        entire expression with the conversion and format spec. */
     *expression = FormattedValue(simple_expression, conversion,
-                                 format_spec, expr_text, LINENO(n),
+                                 format_spec, LINENO(n),
                                  n->n_col_offset, n->n_end_lineno,
                                  n->n_end_col_offset, c->c_arena);
     if (!*expression)
@@ -5313,7 +5311,7 @@ error:
 static int
 fstring_find_literal_and_expr(const char **str, const char *end, int raw,
                               int recurse_lvl, PyObject **literal,
-                              expr_ty *expression,
+                              PyObject **expr_text, expr_ty *expression,
                               struct compiling *c, const node *n)
 {
     int result;
@@ -5341,7 +5339,8 @@ fstring_find_literal_and_expr(const char **str, const char *end, int raw,
     /* We must now be the start of an expression, on a '{'. */
     assert(**str == '{');
 
-    if (fstring_find_expr(str, end, raw, recurse_lvl, expression, c, n) < 0)
+    if (fstring_find_expr(str, end, raw, recurse_lvl, expr_text,
+                          expression, c, n) < 0)
         goto error;
 
     return 0;
@@ -5604,7 +5603,7 @@ FstringParser_ConcatFstring(FstringParser *state, const char **str,
 
     /* Parse the f-string. */
     while (1) {
-        PyObject *literal = NULL;
+        PyObject *literal[2] = {NULL, NULL};
         expr_ty expression = NULL;
 
         /* If there's a zero length literal in front of the
@@ -5612,31 +5611,34 @@ FstringParser_ConcatFstring(FstringParser *state, const char **str,
            the f-string, expression will be NULL (unless result == 1,
            see below). */
         int result = fstring_find_literal_and_expr(str, end, raw, recurse_lvl,
-                                                   &literal, &expression,
-                                                   c, n);
+                                                   &literal[0], &literal[1],
+                                                   &expression, c, n);
         if (result < 0)
             return -1;
 
-        /* Add the literal, if any. */
-        if (!literal) {
-            /* Do nothing. Just leave last_str alone (and possibly
-               NULL). */
-        } else if (!state->last_str) {
-            /*  Note that the literal can be zero length, if the
-                input string is "\\\n" or "\\\r", among others. */
-            state->last_str = literal;
-            literal = NULL;
-        } else {
-            /* We have a literal, concatenate it. */
-            assert(PyUnicode_GET_LENGTH(literal) != 0);
-            if (FstringParser_ConcatAndDel(state, literal) < 0)
-                return -1;
-            literal = NULL;
+        /* Add the literals, if any. */
+        for (int i = 0; i < 2; i++) {
+            if (!literal[i]) {
+                /* Do nothing. Just leave last_str alone (and possibly
+                   NULL). */
+            } else if (!state->last_str) {
+                /*  Note that the literal can be zero length, if the
+                    input string is "\\\n" or "\\\r", among others. */
+                state->last_str = literal[i];
+                literal[i] = NULL;
+            } else {
+                /* We have a literal, concatenate it. */
+                assert(PyUnicode_GET_LENGTH(literal[i]) != 0);
+                if (FstringParser_ConcatAndDel(state, literal[i]) < 0)
+                    return -1;
+                literal[i] = NULL;
+            }
         }
 
-        /* We've dealt with the literal now. It can't be leaked on further
+        /* We've dealt with the literals now. They can't be leaked on further
            errors. */
-        assert(literal == NULL);
+        assert(literal[0] == NULL);
+        assert(literal[1] == NULL);
 
         /* See if we should just loop around to get the next literal
            and expression, while ignoring the expression this
diff --git a/Python/ast_unparse.c b/Python/ast_unparse.c
index f1b991a..f376e86 100644
--- a/Python/ast_unparse.c
+++ b/Python/ast_unparse.c
@@ -665,11 +665,6 @@ append_formattedvalue(_PyUnicodeWriter *writer, expr_ty e, bool is_format_spec)
     }
     Py_DECREF(temp_fv_str);
 
-    if (e->v.FormattedValue.expr_text) {
-        /* Use the = for debug text expansion. */
-        APPEND_STR("=");
-    }
-
     if (e->v.FormattedValue.conversion > 0) {
         switch (e->v.FormattedValue.conversion) {
         case 'a':
diff --git a/Python/compile.c b/Python/compile.c
index 425d0d6..f1c97bd 100644
--- a/Python/compile.c
+++ b/Python/compile.c
@@ -3963,12 +3963,6 @@ compiler_formatted_value(struct compiler *c, expr_ty e)
     int conversion = e->v.FormattedValue.conversion;
     int oparg;
 
-    if (e->v.FormattedValue.expr_text) {
-        /* Push the text of the expression (which already has the '=' in
-           it. */
-        ADDOP_LOAD_CONST(c, e->v.FormattedValue.expr_text);
-    }
-
     /* The expression to be formatted. */
     VISIT(c, expr, e->v.FormattedValue.value);
 
@@ -3991,11 +3985,6 @@ compiler_formatted_value(struct compiler *c, expr_ty e)
     /* And push our opcode and oparg */
     ADDOP_I(c, FORMAT_VALUE, oparg);
 
-    /* If we have expr_text, join the 2 strings on the stack. */
-    if (e->v.FormattedValue.expr_text) {
-        ADDOP_I(c, BUILD_STRING, 2);
-    }
-
     return 1;
 }