[3.10] bpo-44885: Correct the ast locations of f-strings with format specs and repeated expressions (GH-27729) (GH-27743)
(cherry picked from commit 8e832fb2a2cb54d7262148b6ec15563dffb48d63)
Co-authored-by: Pablo Galindo Salgado <Pablogsal@gmail.com>
diff --git a/Parser/string_parser.c b/Parser/string_parser.c
index 66405b2..fb37d37 100644
--- a/Parser/string_parser.c
+++ b/Parser/string_parser.c
@@ -279,49 +279,48 @@ _PyPegen_parsestr(Parser *p, int *bytesmode, int *rawmode, PyObject **result,
/* Fix locations for the given node and its children.
`parent` is the enclosing node.
+ `expr_start` is the starting position of the expression (pointing to the open brace).
`n` is the node which locations are going to be fixed relative to parent.
`expr_str` is the child node's string representation, including braces.
*/
static bool
-fstring_find_expr_location(Token *parent, char *expr_str, int *p_lines, int *p_cols)
+fstring_find_expr_location(Token *parent, const char* expr_start, char *expr_str, int *p_lines, int *p_cols)
{
*p_lines = 0;
*p_cols = 0;
+ assert(expr_start != NULL && *expr_start == '{');
if (parent && parent->bytes) {
const char *parent_str = PyBytes_AsString(parent->bytes);
if (!parent_str) {
return false;
}
- const char *substr = strstr(parent_str, expr_str);
- if (substr) {
- // The following is needed, in order to correctly shift the column
- // offset, in the case that (disregarding any whitespace) a newline
- // immediately follows the opening curly brace of the fstring expression.
- bool newline_after_brace = 1;
- const char *start = substr + 1;
- while (start && *start != '}' && *start != '\n') {
- if (*start != ' ' && *start != '\t' && *start != '\f') {
- newline_after_brace = 0;
- break;
- }
- start++;
+ // The following is needed, in order to correctly shift the column
+ // offset, in the case that (disregarding any whitespace) a newline
+ // immediately follows the opening curly brace of the fstring expression.
+ bool newline_after_brace = 1;
+ const char *start = expr_start + 1;
+ while (start && *start != '}' && *start != '\n') {
+ if (*start != ' ' && *start != '\t' && *start != '\f') {
+ newline_after_brace = 0;
+ break;
}
+ start++;
+ }
- // Account for the characters from the last newline character to our
- // left until the beginning of substr.
- if (!newline_after_brace) {
- start = substr;
- while (start > parent_str && *start != '\n') {
- start--;
- }
- *p_cols += (int)(substr - start);
+ // Account for the characters from the last newline character to our
+ // left until the beginning of expr_start.
+ if (!newline_after_brace) {
+ start = expr_start;
+ while (start > parent_str && *start != '\n') {
+ start--;
}
- /* adjust the start based on the number of newlines encountered
- before the f-string expression */
- for (const char *p = parent_str; p < substr; p++) {
- if (*p == '\n') {
- (*p_lines)++;
- }
+ *p_cols += (int)(expr_start - start);
+ }
+ /* adjust the start based on the number of newlines encountered
+ before the f-string expression */
+ for (const char *p = parent_str; p < expr_start; p++) {
+ if (*p == '\n') {
+ (*p_lines)++;
}
}
}
@@ -365,7 +364,7 @@ fstring_compile_expr(Parser *p, const char *expr_start, const char *expr_end,
len = expr_end - expr_start;
/* Allocate 3 extra bytes: open paren, close paren, null byte. */
- str = PyMem_Malloc(len + 3);
+ str = PyMem_Calloc(len + 3, sizeof(char));
if (str == NULL) {
PyErr_NoMemory();
return NULL;
@@ -373,17 +372,10 @@ fstring_compile_expr(Parser *p, const char *expr_start, const char *expr_end,
// The call to fstring_find_expr_location is responsible for finding the column offset
// the generated AST nodes need to be shifted to the right, which is equal to the number
- // of the f-string characters before the expression starts. In order to correctly compute
- // this offset, strstr gets called in fstring_find_expr_location which only succeeds
- // if curly braces appear before and after the f-string expression (exactly like they do
- // in the f-string itself), hence the following lines.
- str[0] = '{';
+ // of the f-string characters before the expression starts.
memcpy(str+1, expr_start, len);
- str[len+1] = '}';
- str[len+2] = 0;
-
int lines, cols;
- if (!fstring_find_expr_location(t, str, &lines, &cols)) {
+ if (!fstring_find_expr_location(t, expr_start-1, str+1, &lines, &cols)) {
PyMem_Free(str);
return NULL;
}