Skip to content

Commit c41eed1

Browse files
vstinnervstinner
authored
bpo-40521: Make bytes singletons per interpreter (GH-21074)
Each interpreter now has its own empty bytes string and single byte character singletons. Replace STRINGLIB_EMPTY macro with STRINGLIB_GET_EMPTY() macro.
1 parent 32f2eda commit c41eed1

File tree

13 files changed

+96
-53
lines changed

13 files changed

+96
-53
lines changed

‎Include/internal/pycore_interp.h

Lines changed: 6 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -65,6 +65,11 @@ struct _Py_unicode_fs_codec {
6565
_Py_error_handler error_handler;
6666
};
6767

68+
struct _Py_bytes_state {
69+
PyBytesObject *characters[256];
70+
PyBytesObject *empty_string;
71+
};
72+
6873
struct _Py_unicode_state {
6974
struct _Py_unicode_fs_codec fs_codec;
7075
};
@@ -233,6 +238,7 @@ struct _is {
233238
*/
234239
PyLongObject* small_ints[_PY_NSMALLNEGINTS + _PY_NSMALLPOSINTS];
235240
#endif
241+
struct _Py_bytes_state bytes;
236242
struct _Py_unicode_state unicode;
237243
struct _Py_float_state float_state;
238244
/* Using a cache is very effective since typically only a single slice is

‎Include/internal/pycore_pylifecycle.h

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -63,7 +63,7 @@ extern void _PyDict_Fini(PyThreadState *tstate);
6363
extern void _PyTuple_Fini(PyThreadState *tstate);
6464
extern void _PyList_Fini(PyThreadState *tstate);
6565
extern void _PySet_Fini(PyThreadState *tstate);
66-
extern void _PyBytes_Fini(void);
66+
extern void _PyBytes_Fini(PyThreadState *tstate);
6767
extern void _PyFloat_Fini(PyThreadState *tstate);
6868
extern void _PySlice_Fini(PyThreadState *tstate);
6969
extern void _PyAsyncGen_Fini(PyThreadState *tstate);

‎Misc/NEWS.d/next/Core and Builtins/2020-05-20-01-17-34.bpo-40521.wvAehI.rst

Lines changed: 9 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,9 @@
1-
The tuple free lists, the empty tuple singleton, the list free list, the empty
2-
frozenset singleton, the float free list, the slice cache, the dict free lists,
3-
the frame free list, the asynchronous generator free lists, and the context
4-
free list are no longer shared by all interpreters: each interpreter now its
5-
has own free lists and caches.
1+
Each interpreter now its has own free lists, singletons and caches:
2+
3+
* Free lists: float, tuple, list, dict, frame, context,
4+
asynchronous generator.
5+
* Singletons: empty tuple, empty frozenset, empty bytes string,
6+
single byte character.
7+
* Slice cache.
8+
9+
They are no longer shared by all interpreters.

‎Objects/bytesobject.c

Lines changed: 55 additions & 27 deletions
Original file line numberDiff line numberDiff line change
@@ -18,9 +18,6 @@ class bytes "PyBytesObject *" "&PyBytes_Type"
1818

1919
#include "clinic/bytesobject.c.h"
2020

21-
static PyBytesObject *characters[UCHAR_MAX + 1];
22-
static PyBytesObject *nullstring;
23-
2421
_Py_IDENTIFIER(__bytes__);
2522

2623
/* PyBytesObject_SIZE gives the basic size of a string; any memory allocation
@@ -35,6 +32,15 @@ _Py_IDENTIFIER(__bytes__);
3532
Py_LOCAL_INLINE(Py_ssize_t) _PyBytesWriter_GetSize(_PyBytesWriter *writer,
3633
char *str);
3734

35+
36+
static struct _Py_bytes_state*
37+
get_bytes_state(void)
38+
{
39+
PyInterpreterState *interp = _PyInterpreterState_GET();
40+
return &interp->bytes;
41+
}
42+
43+
3844
/*
3945
For PyBytes_FromString(), the parameter `str' points to a null-terminated
4046
string containing exactly `size' bytes.
@@ -63,9 +69,13 @@ _PyBytes_FromSize(Py_ssize_t size, int use_calloc)
6369
PyBytesObject *op;
6470
assert(size >= 0);
6571

66-
if (size == 0 && (op = nullstring) != NULL) {
67-
Py_INCREF(op);
68-
return (PyObject *)op;
72+
if (size == 0) {
73+
struct _Py_bytes_state *state = get_bytes_state();
74+
op = state->empty_string;
75+
if (op != NULL) {
76+
Py_INCREF(op);
77+
return (PyObject *)op;
78+
}
6979
}
7080

7181
if ((size_t)size > (size_t)PY_SSIZE_T_MAX - PyBytesObject_SIZE) {
@@ -88,8 +98,9 @@ _PyBytes_FromSize(Py_ssize_t size, int use_calloc)
8898
op->ob_sval[size] = '\0';
8999
/* empty byte string singleton */
90100
if (size == 0) {
91-
nullstring = op;
101+
struct _Py_bytes_state *state = get_bytes_state();
92102
Py_INCREF(op);
103+
state->empty_string = op;
93104
}
94105
return (PyObject *) op;
95106
}
@@ -103,11 +114,13 @@ PyBytes_FromStringAndSize(const char *str, Py_ssize_t size)
103114
"Negative size passed to PyBytes_FromStringAndSize");
104115
return NULL;
105116
}
106-
if (size == 1 && str != NULL &&
107-
(op = characters[*str & UCHAR_MAX]) != NULL)
108-
{
109-
Py_INCREF(op);
110-
return (PyObject *)op;
117+
if (size == 1 && str != NULL) {
118+
struct _Py_bytes_state *state = get_bytes_state();
119+
op = state->characters[*str & UCHAR_MAX];
120+
if (op != NULL) {
121+
Py_INCREF(op);
122+
return (PyObject *)op;
123+
}
111124
}
112125

113126
op = (PyBytesObject *)_PyBytes_FromSize(size, 0);
@@ -119,8 +132,9 @@ PyBytes_FromStringAndSize(const char *str, Py_ssize_t size)
119132
memcpy(op->ob_sval, str, size);
120133
/* share short strings */
121134
if (size == 1) {
122-
characters[*str & UCHAR_MAX] = op;
135+
struct _Py_bytes_state *state = get_bytes_state();
123136
Py_INCREF(op);
137+
state->characters[*str & UCHAR_MAX] = op;
124138
}
125139
return (PyObject *) op;
126140
}
@@ -138,13 +152,21 @@ PyBytes_FromString(const char *str)
138152
"byte string is too long");
139153
return NULL;
140154
}
141-
if (size == 0 && (op = nullstring) != NULL) {
142-
Py_INCREF(op);
143-
return (PyObject *)op;
155+
156+
struct _Py_bytes_state *state = get_bytes_state();
157+
if (size == 0) {
158+
op = state->empty_string;
159+
if (op != NULL) {
160+
Py_INCREF(op);
161+
return (PyObject *)op;
162+
}
144163
}
145-
if (size == 1 && (op = characters[*str & UCHAR_MAX]) != NULL) {
146-
Py_INCREF(op);
147-
return (PyObject *)op;
164+
else if (size == 1) {
165+
op = state->characters[*str & UCHAR_MAX];
166+
if (op != NULL) {
167+
Py_INCREF(op);
168+
return (PyObject *)op;
169+
}
148170
}
149171

150172
/* Inline PyObject_NewVar */
@@ -157,11 +179,12 @@ PyBytes_FromString(const char *str)
157179
memcpy(op->ob_sval, str, size+1);
158180
/* share short strings */
159181
if (size == 0) {
160-
nullstring = op;
161182
Py_INCREF(op);
162-
} else if (size == 1) {
163-
characters[*str & UCHAR_MAX] = op;
183+
state->empty_string = op;
184+
}
185+
else if (size == 1) {
164186
Py_INCREF(op);
187+
state->characters[*str & UCHAR_MAX] = op;
165188
}
166189
return (PyObject *) op;
167190
}
@@ -1249,6 +1272,8 @@ PyBytes_AsStringAndSize(PyObject *obj,
12491272
/* -------------------------------------------------------------------- */
12501273
/* Methods */
12511274

1275+
#define STRINGLIB_GET_EMPTY() get_bytes_state()->empty_string
1276+
12521277
#include "stringlib/stringdefs.h"
12531278

12541279
#include "stringlib/fastsearch.h"
@@ -1261,6 +1286,8 @@ PyBytes_AsStringAndSize(PyObject *obj,
12611286

12621287
#include "stringlib/transmogrify.h"
12631288

1289+
#undef STRINGLIB_GET_EMPTY
1290+
12641291
PyObject *
12651292
PyBytes_Repr(PyObject *obj, int smartquotes)
12661293
{
@@ -3058,12 +3085,13 @@ _PyBytes_Resize(PyObject **pv, Py_ssize_t newsize)
30583085
}
30593086

30603087
void
3061-
_PyBytes_Fini(void)
3088+
_PyBytes_Fini(PyThreadState *tstate)
30623089
{
3063-
int i;
3064-
for (i = 0; i < UCHAR_MAX + 1; i++)
3065-
Py_CLEAR(characters[i]);
3066-
Py_CLEAR(nullstring);
3090+
struct _Py_bytes_state* state = &tstate->interp->bytes;
3091+
for (int i = 0; i < UCHAR_MAX + 1; i++) {
3092+
Py_CLEAR(state->characters[i]);
3093+
}
3094+
Py_CLEAR(state->empty_string);
30673095
}
30683096

30693097
/*********************** Bytes Iterator ****************************/

‎Objects/stringlib/README.txt

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -11,10 +11,10 @@ STRINGLIB_CHAR
1111

1212
the type used to hold a character (char or Py_UNICODE)
1313

14-
STRINGLIB_EMPTY
14+
STRINGLIB_GET_EMPTY()
1515

16-
a PyObject representing the empty string, only to be used if
17-
STRINGLIB_MUTABLE is 0
16+
returns a PyObject representing the empty string, only to be used if
17+
STRINGLIB_MUTABLE is 0. It must not be NULL.
1818

1919
Py_ssize_t STRINGLIB_LEN(PyObject*)
2020

‎Objects/stringlib/asciilib.h

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -11,7 +11,7 @@
1111
#define STRINGLIB_CHAR Py_UCS1
1212
#define STRINGLIB_TYPE_NAME "unicode"
1313
#define STRINGLIB_PARSE_CODE "U"
14-
#define STRINGLIB_EMPTY unicode_empty
14+
#define STRINGLIB_GET_EMPTY() unicode_empty
1515
#define STRINGLIB_ISSPACE Py_UNICODE_ISSPACE
1616
#define STRINGLIB_ISLINEBREAK BLOOM_LINEBREAK
1717
#define STRINGLIB_ISDECIMAL Py_UNICODE_ISDECIMAL

‎Objects/stringlib/partition.h

Lines changed: 12 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -37,10 +37,12 @@ STRINGLIB(partition)(PyObject* str_obj,
3737
#else
3838
Py_INCREF(str_obj);
3939
PyTuple_SET_ITEM(out, 0, (PyObject*) str_obj);
40-
Py_INCREF(STRINGLIB_EMPTY);
41-
PyTuple_SET_ITEM(out, 1, (PyObject*) STRINGLIB_EMPTY);
42-
Py_INCREF(STRINGLIB_EMPTY);
43-
PyTuple_SET_ITEM(out, 2, (PyObject*) STRINGLIB_EMPTY);
40+
PyObject *empty = (PyObject*)STRINGLIB_GET_EMPTY();
41+
assert(empty != NULL);
42+
Py_INCREF(empty);
43+
PyTuple_SET_ITEM(out, 1, empty);
44+
Py_INCREF(empty);
45+
PyTuple_SET_ITEM(out, 2, empty);
4446
#endif
4547
return out;
4648
}
@@ -90,10 +92,12 @@ STRINGLIB(rpartition)(PyObject* str_obj,
9092
return NULL;
9193
}
9294
#else
93-
Py_INCREF(STRINGLIB_EMPTY);
94-
PyTuple_SET_ITEM(out, 0, (PyObject*) STRINGLIB_EMPTY);
95-
Py_INCREF(STRINGLIB_EMPTY);
96-
PyTuple_SET_ITEM(out, 1, (PyObject*) STRINGLIB_EMPTY);
95+
PyObject *empty = (PyObject*)STRINGLIB_GET_EMPTY();
96+
assert(empty != NULL);
97+
Py_INCREF(empty);
98+
PyTuple_SET_ITEM(out, 0, empty);
99+
Py_INCREF(empty);
100+
PyTuple_SET_ITEM(out, 1, empty);
97101
Py_INCREF(str_obj);
98102
PyTuple_SET_ITEM(out, 2, (PyObject*) str_obj);
99103
#endif

‎Objects/stringlib/stringdefs.h

Lines changed: 4 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,10 @@
11
#ifndef STRINGLIB_STRINGDEFS_H
22
#define STRINGLIB_STRINGDEFS_H
33

4+
#ifndef STRINGLIB_GET_EMPTY
5+
# error "STRINGLIB_GET_EMPTY macro must be defined"
6+
#endif
7+
48
/* this is sort of a hack. there's at least one place (formatting
59
floats) where some stringlib code takes a different path if it's
610
compiled as unicode. */
@@ -13,7 +17,6 @@
1317
#define STRINGLIB_CHAR char
1418
#define STRINGLIB_TYPE_NAME "string"
1519
#define STRINGLIB_PARSE_CODE "S"
16-
#define STRINGLIB_EMPTY nullstring
1720
#define STRINGLIB_ISSPACE Py_ISSPACE
1821
#define STRINGLIB_ISLINEBREAK(x) ((x == '\n') || (x == '\r'))
1922
#define STRINGLIB_ISDECIMAL(x) ((x >= '0') && (x <= '9'))

‎Objects/stringlib/ucs1lib.h

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -11,7 +11,7 @@
1111
#define STRINGLIB_CHAR Py_UCS1
1212
#define STRINGLIB_TYPE_NAME "unicode"
1313
#define STRINGLIB_PARSE_CODE "U"
14-
#define STRINGLIB_EMPTY unicode_empty
14+
#define STRINGLIB_GET_EMPTY() unicode_empty
1515
#define STRINGLIB_ISSPACE Py_UNICODE_ISSPACE
1616
#define STRINGLIB_ISLINEBREAK BLOOM_LINEBREAK
1717
#define STRINGLIB_ISDECIMAL Py_UNICODE_ISDECIMAL

‎Objects/stringlib/ucs2lib.h

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -11,7 +11,7 @@
1111
#define STRINGLIB_CHAR Py_UCS2
1212
#define STRINGLIB_TYPE_NAME "unicode"
1313
#define STRINGLIB_PARSE_CODE "U"
14-
#define STRINGLIB_EMPTY unicode_empty
14+
#define STRINGLIB_GET_EMPTY() unicode_empty
1515
#define STRINGLIB_ISSPACE Py_UNICODE_ISSPACE
1616
#define STRINGLIB_ISLINEBREAK BLOOM_LINEBREAK
1717
#define STRINGLIB_ISDECIMAL Py_UNICODE_ISDECIMAL

‎Objects/stringlib/ucs4lib.h

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -11,7 +11,7 @@
1111
#define STRINGLIB_CHAR Py_UCS4
1212
#define STRINGLIB_TYPE_NAME "unicode"
1313
#define STRINGLIB_PARSE_CODE "U"
14-
#define STRINGLIB_EMPTY unicode_empty
14+
#define STRINGLIB_GET_EMPTY() unicode_empty
1515
#define STRINGLIB_ISSPACE Py_UNICODE_ISSPACE
1616
#define STRINGLIB_ISLINEBREAK BLOOM_LINEBREAK
1717
#define STRINGLIB_ISDECIMAL Py_UNICODE_ISDECIMAL

‎Objects/stringlib/unicodedefs.h

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -13,7 +13,7 @@
1313
#define STRINGLIB_CHAR Py_UNICODE
1414
#define STRINGLIB_TYPE_NAME "unicode"
1515
#define STRINGLIB_PARSE_CODE "U"
16-
#define STRINGLIB_EMPTY unicode_empty
16+
#define STRINGLIB_GET_EMPTY() unicode_empty
1717
#define STRINGLIB_ISSPACE Py_UNICODE_ISSPACE
1818
#define STRINGLIB_ISLINEBREAK BLOOM_LINEBREAK
1919
#define STRINGLIB_ISDECIMAL Py_UNICODE_ISDECIMAL

‎Python/pylifecycle.c

Lines changed: 1 addition & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1262,9 +1262,7 @@ finalize_interp_types(PyThreadState *tstate, int is_main_interp)
12621262

12631263
_PySlice_Fini(tstate);
12641264

1265-
if (is_main_interp) {
1266-
_PyBytes_Fini();
1267-
}
1265+
_PyBytes_Fini(tstate);
12681266
_PyUnicode_Fini(tstate);
12691267
_PyFloat_Fini(tstate);
12701268
_PyLong_Fini(tstate);

0 commit comments

Comments
 (0)