[3.14] gh-130077: Properly match full soft keywords in the parser (GH-135317) (#135348)

miss-islington · pablogsal · web-flow · commit 77c4b5d413aa · 2025-06-11T02:01:37.000+01:00
gh-130077: Properly match full soft keywords in the parser (GH-135317) (cherry picked from commit ff2b5f4) Co-authored-by: Pablo Galindo Salgado <Pablogsal@gmail.com>
diff --git a/Lib/test/test_syntax.py b/Lib/test/test_syntax.py
@@ -382,6 +382,13 @@
 Traceback (most recent call last):
 SyntaxError: invalid syntax
 
+# But prefixes of soft keywords should
+# still raise specialized errors
+
+>>> (mat x)
+Traceback (most recent call last):
+SyntaxError: invalid syntax. Perhaps you forgot a comma?
+
 From compiler_complex_args():
 
 >>> def f(None=1):
diff --git a/Misc/NEWS.d/next/Core_and_Builtins/2025-06-09-23-57-37.gh-issue-130077.MHknDB.rst b/Misc/NEWS.d/next/Core_and_Builtins/2025-06-09-23-57-37.gh-issue-130077.MHknDB.rst
@@ -0,0 +1,2 @@
+Properly raise custom syntax errors when incorrect syntax containing names
+that are prefixes of soft keywords is encountered.  Patch by Pablo Galindo.
diff --git a/Parser/pegen.c b/Parser/pegen.c
@@ -610,7 +610,8 @@ expr_ty _PyPegen_soft_keyword_token(Parser *p) {
     Py_ssize_t size;
     PyBytes_AsStringAndSize(t->bytes, &the_token, &size);
     for (char **keyword = p->soft_keywords; *keyword != NULL; keyword++) {
-        if (strncmp(*keyword, the_token, (size_t)size) == 0) {
+        if (strlen(*keyword) == (size_t)size &&
+            strncmp(*keyword, the_token, (size_t)size) == 0) {
             return _PyPegen_name_from_token(p, t);
         }
     }

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,2 @@`
	`1`	`+Properly raise custom syntax errors when incorrect syntax containing names`
	`2`	`+that are prefixes of soft keywords is encountered. Patch by Pablo Galindo.`
Original file line number	Diff line number	Diff line change
`@@ -610,7 +610,8 @@ expr_ty _PyPegen_soft_keyword_token(Parser *p) {`
`610`	`610`	`Py_ssize_t size;`
`611`	`611`	`PyBytes_AsStringAndSize(t->bytes, &the_token, &size);`
`612`	`612`	`for (char *keyword = p->soft_keywords; keyword != NULL; keyword++) {`
`613`		`- if (strncmp(*keyword, the_token, (size_t)size) == 0) {`
	`613`	`+ if (strlen(*keyword) == (size_t)size &&`
	`614`	`+ strncmp(*keyword, the_token, (size_t)size) == 0) {`
`614`	`615`	`return _PyPegen_name_from_token(p, t);`
`615`	`616`	`}`
`616`	`617`	`}`