Thanks to visit codestin.com
Credit goes to github.com

Skip to content

gh-119396: Optimize PyUnicode_FromFormat() UTF-8 decoder #119398

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
vstinner merged 3 commits into from
May 22, 2024
+141 −62
Merged

gh-119396: Optimize PyUnicode_FromFormat() UTF-8 decoder #119398

Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
33b1c4a
gh-119182: Optimize PyUnicode_FromFormat() UTF-8 decoder
vstinner May 22, 2024
99f4b13
Fix unicode_decode_utf8() perf regression
vstinner May 22, 2024
d496db8
Address review
vstinner May 22, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Fix unicode_decode_utf8() perf regression
  • Loading branch information
@vstinner
vstinner committed May 22, 2024
commit 99f4b131c412df480aa4667accf8c9ba95a1c587
108 changes: 75 additions & 33 deletions Objects/unicodeobject.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4750,35 +4750,12 @@ ascii_decode(const char *start, const char *end, Py_UCS1 *dest)


static int
unicode_decode_utf8_writer(_PyUnicodeWriter *writer,
const char *s, Py_ssize_t size,
_Py_error_handler error_handler, const char *errors,
Py_ssize_t *consumed)
unicode_decode_utf8_impl(_PyUnicodeWriter *writer,
const char *starts, const char *s, const char *end,
_Py_error_handler error_handler,
const char *errors,
Py_ssize_t *consumed)
{
const char *starts = s;
const char *end = s + size;

// fast path: try ASCII string.
if (_PyUnicodeWriter_Prepare(writer, size, 127) < 0) {
return -1;
}

Py_UCS1 *dest = (Py_UCS1*)writer->data + writer->pos * writer->kind;
if (writer->kind == PyUnicode_1BYTE_KIND
&& _Py_IS_ALIGNED(dest, ALIGNOF_SIZE_T))
{
Py_ssize_t decoded = ascii_decode(s, end, dest);
writer->pos += decoded;

if (decoded == size) {
if (consumed) {
*consumed = size;
}
return 0;
}
s += decoded;
}

Py_ssize_t startinpos, endinpos;
const char *errmsg = "";
PyObject *error_handler_obj = NULL;
Expand Down Expand Up @@ -4828,6 +4805,8 @@ unicode_decode_utf8_writer(_PyUnicodeWriter *writer,
endinpos = startinpos + ch - 1;
break;
default:
// ch doesn't fit into kind, so change the buffer kind to write
// the character
if (_PyUnicodeWriter_WriteCharInline(writer, ch) < 0)
goto onError;
continue;
Expand Down Expand Up @@ -4899,8 +4878,9 @@ unicode_decode_utf8(const char *s, Py_ssize_t size,
Py_ssize_t *consumed)
{
if (size == 0) {
if (consumed)
if (consumed) {
*consumed = 0;
}
_Py_RETURN_UNICODE_EMPTY();
}

Expand All @@ -4912,19 +4892,81 @@ unicode_decode_utf8(const char *s, Py_ssize_t size,
return get_latin1_char((unsigned char)s[0]);
}

// fast path: try ASCII string.
const char *starts = s;
const char *end = s + size;
PyObject *u = PyUnicode_New(size, 127);
if (u == NULL) {
return NULL;
}
Py_ssize_t decoded = ascii_decode(s, end, PyUnicode_1BYTE_DATA(u));
if (decoded == size) {
if (consumed) {
*consumed = size;
}
return u;
}
s += decoded;
size -= decoded;

// Use _PyUnicodeWriter after fast path is failed.
_PyUnicodeWriter writer;
_PyUnicodeWriter_Init(&writer);
_PyUnicodeWriter_InitWithBuffer(&writer, u);
writer.pos = decoded;

if (unicode_decode_utf8_writer(&writer, s, size,
error_handler, errors,
consumed) < 0) {
if (unicode_decode_utf8_impl(&writer, starts, s, end,
error_handler, errors,
consumed) < 0) {
_PyUnicodeWriter_Dealloc(&writer);
return NULL;
}
return _PyUnicodeWriter_Finish(&writer);
}


static int
unicode_decode_utf8_writer(_PyUnicodeWriter *writer,
const char *s, Py_ssize_t size,
_Py_error_handler error_handler, const char *errors,
Py_ssize_t *consumed)
{
if (size == 0) {
if (consumed) {
*consumed = 0;
}
return 0;
}

// fast path: try ASCII string.
if (_PyUnicodeWriter_Prepare(writer, size, 127) < 0) {
return -1;
}

const char *starts = s;
const char *end = s + size;
Py_ssize_t decoded = 0;
Py_UCS1 *dest = (Py_UCS1*)writer->data + writer->pos * writer->kind;
if (writer->kind == PyUnicode_1BYTE_KIND
&& _Py_IS_ALIGNED(dest, ALIGNOF_SIZE_T))
{
decoded = ascii_decode(s, end, dest);
writer->pos += decoded;

if (decoded == size) {
if (consumed) {
*consumed = size;
}
return 0;
}
s += decoded;
size -= decoded;
}

return unicode_decode_utf8_impl(writer, starts, s, end,
error_handler, errors, consumed);
}


PyObject *
PyUnicode_DecodeUTF8Stateful(const char *s,
Py_ssize_t size,
Expand Down