underscores as placeholders

This commit is contained in:
Jana Dönszelmann 2026-01-05 18:27:49 +01:00
parent ab2849da6f
commit 1c9882a2cb
No known key found for this signature in database

View file

@ -1,4 +1,4 @@
From d947b1d3160d8a6990f9b7bed476184271fedd9a Mon Sep 17 00:00:00 2001
From 79b30daff39dc53c4d822250946e1667cae1798e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jana=20D=C3=B6nszelmann?= <jana@donsz.nl>
Date: Mon, 5 Jan 2026 16:14:52 +0100
Subject: [PATCH] add to parser
@ -6,9 +6,9 @@ Subject: [PATCH] add to parser
---
Grammar/Tokens | 1 +
Grammar/python.gram | 77 ++++++++++++++++++++++++++++++-----------
Parser/action_helpers.c | 26 ++++++++++++++
Parser/action_helpers.c | 62 +++++++++++++++++++++++++++++++++
Parser/pegen.h | 1 +
4 files changed, 84 insertions(+), 21 deletions(-)
4 files changed, 120 insertions(+), 21 deletions(-)
diff --git a/Grammar/Tokens b/Grammar/Tokens
index 0547e6ed08f..fd0ce412f6c 100644
@ -23,7 +23,7 @@ index 0547e6ed08f..fd0ce412f6c 100644
OP
TYPE_IGNORE
diff --git a/Grammar/python.gram b/Grammar/python.gram
index 110136af81b..7afaa1c140e 100644
index 110136af81b..30f7950d1a9 100644
--- a/Grammar/python.gram
+++ b/Grammar/python.gram
@@ -726,13 +726,13 @@ star_expressions[expr_ty]:
@ -94,9 +94,9 @@ index 110136af81b..7afaa1c140e 100644
+ | lhs=pipe '|>' rhs=primary_nocall b=genexp {
+ _PyAST_Call(rhs,
+ _PyPegen_desugar_pipe(
+ p,
+ CHECK(asdl_expr_seq*, (asdl_expr_seq*)_PyPegen_singleton_seq(p, b)),
+ lhs,
+ p->arena
+ lhs
+ )
+ , NULL,
+ EXTRA
@ -106,9 +106,9 @@ index 110136af81b..7afaa1c140e 100644
+ _PyAST_Call(
+ rhs,
+ _PyPegen_desugar_pipe(
+ p,
+ (arg) ? ((expr_ty) arg)->v.Call.args : NULL,
+ lhs,
+ p->arena
+ lhs
+ ),
+ (arg) ? ((expr_ty) arg)->v.Call.keywords : NULL,
+ EXTRA
@ -195,16 +195,20 @@ index 110136af81b..7afaa1c140e 100644
invalid_starred_expression_unpacking:
| a='*' expression '=' b=expression { RAISE_SYNTAX_ERROR_KNOWN_RANGE(a, b, "cannot assign to iterable argument unpacking") }
diff --git a/Parser/action_helpers.c b/Parser/action_helpers.c
index 50856686335..44d617f48cc 100644
index 50856686335..3241c007aa8 100644
--- a/Parser/action_helpers.c
+++ b/Parser/action_helpers.c
@@ -1152,6 +1152,32 @@ _PyPegen_get_last_comprehension_item(comprehension_ty comprehension) {
@@ -1152,6 +1152,68 @@ _PyPegen_get_last_comprehension_item(comprehension_ty comprehension) {
return PyPegen_last_item(comprehension->ifs, expr_ty);
}
+asdl_expr_seq * _PyPegen_desugar_pipe(asdl_expr_seq* args, expr_ty piped_lhs, PyArena* arena) {
+asdl_expr_seq * _PyPegen_desugar_pipe(Parser * p, asdl_expr_seq* args, expr_ty piped_lhs) {
+ // loop index
+ Py_ssize_t i = 0;
+
+ // if the list was somehow null, replace it with a list only containing the piped argument
+ if (args == NULL) {
+ asdl_expr_seq *new_args = _Py_asdl_expr_seq_new(1, arena);
+ asdl_expr_seq *new_args = _Py_asdl_expr_seq_new(1, p->arena);
+ if (new_args == NULL) {
+ return NULL;
+ }
@ -213,14 +217,46 @@ index 50856686335..44d617f48cc 100644
+ return new_args;
+ }
+
+ // calculate the initial length
+ Py_ssize_t orig_args_len = asdl_seq_LEN(args);
+ asdl_expr_seq *new_args = _Py_asdl_expr_seq_new(orig_args_len + 1, arena);
+
+ // look for a `_` to replace
+ Py_ssize_t underscore_index = -1;
+ for (i = 0; i < orig_args_len; i++) {
+ expr_ty arg = asdl_seq_GET(args, i);
+
+ // if we see an underscore, count it
+ if (
+ arg->kind == Name_kind
+ && PyUnicode_CompareWithASCIIString(arg->v.Name.id, "_") == 0
+ ) {
+ // maybe this is the 2nd underscore, raise a syntax error
+ if (underscore_index != -1) {
+ return RAISE_SYNTAX_ERROR_KNOWN_RANGE(arg, arg, "only one `_` is allowed when piping");
+ }
+
+ underscore_index = i;
+ }
+
+ assert(current_elem->kind == Constant_kind);
+ }
+
+ // overwrite the `_` element if found
+ if (underscore_index != -1) {
+ asdl_seq_SET(args, underscore_index, piped_lhs);
+ return args;
+ }
+
+ // otherwise, allocate a new expr seq of one item longer than the original
+ asdl_expr_seq *new_args = _Py_asdl_expr_seq_new(orig_args_len + 1, p->arena);
+ if (new_args == NULL) {
+ return NULL;
+ }
+
+ // stick the piped element at the end (so at orig_args_len)
+ // which is usually 1 past the end but now we allocated one more element
+ asdl_seq_SET(new_args, orig_args_len, piped_lhs);
+ Py_ssize_t i = 0;
+ // and copy the rest of the elements.
+ for (i = 0; i < orig_args_len; i++) {
+ asdl_seq_SET(new_args, i, asdl_seq_GET(args, i));
+ }
@ -232,14 +268,14 @@ index 50856686335..44d617f48cc 100644
int lineno, int col_offset, int end_lineno,
int end_col_offset, PyArena *arena) {
diff --git a/Parser/pegen.h b/Parser/pegen.h
index be5333eb268..8e109999b54 100644
index be5333eb268..d9557494e6c 100644
--- a/Parser/pegen.h
+++ b/Parser/pegen.h
@@ -344,6 +344,7 @@ stmt_ty _PyPegen_class_def_decorators(Parser *, asdl_expr_seq *, stmt_ty);
KeywordOrStarred *_PyPegen_keyword_or_starred(Parser *, void *, int);
asdl_expr_seq *_PyPegen_seq_extract_starred_exprs(Parser *, asdl_seq *);
asdl_keyword_seq *_PyPegen_seq_delete_starred_exprs(Parser *, asdl_seq *);
+asdl_expr_seq * _PyPegen_desugar_pipe(asdl_expr_seq* args, expr_ty piped_lhs, PyArena* arena);
+asdl_expr_seq * _PyPegen_desugar_pipe(Parser*, asdl_expr_seq*, expr_ty);
expr_ty _PyPegen_collect_call_seqs(Parser *, asdl_expr_seq *, asdl_seq *,
int lineno, int col_offset, int end_lineno,
int end_col_offset, PyArena *arena);