bpo-36365: Rewrite structseq_repr() using _PyUnicodeWriter (GH-12440)

No longer limit repr(structseq) to 512 bytes. Use _PyUnicodeWriter
for better performance and to write directly Unicode rather than
encoding repr() value to UTF-8 and then decoding from UTF-8.
This commit is contained in:
Victor Stinner 2019-03-20 00:05:51 +01:00 committed by GitHub
parent fd23cfa464
commit c70ab02df2
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 69 additions and 58 deletions

View File

@ -0,0 +1 @@
repr(structseq) is no longer limited to 512 bytes.

View File

@ -168,78 +168,88 @@ structseq_new_impl(PyTypeObject *type, PyObject *arg, PyObject *dict)
static PyObject *
structseq_repr(PyStructSequence *obj)
{
/* buffer and type size were chosen well considered. */
#define REPR_BUFFER_SIZE 512
#define TYPE_MAXSIZE 100
PyTypeObject *typ = Py_TYPE(obj);
Py_ssize_t i;
int removelast = 0;
Py_ssize_t len;
char buf[REPR_BUFFER_SIZE];
char *endofbuf, *pbuf = buf;
_PyUnicodeWriter writer;
/* pointer to end of writeable buffer; safes space for "...)\0" */
endofbuf= &buf[REPR_BUFFER_SIZE-5];
/* Write "typename(" */
PyObject *type_name = PyUnicode_DecodeUTF8(typ->tp_name,
strlen(typ->tp_name),
NULL);
if (type_name == NULL) {
goto error;
}
/* "typename(", limited to TYPE_MAXSIZE */
len = strlen(typ->tp_name);
len = Py_MIN(len, TYPE_MAXSIZE);
memcpy(pbuf, typ->tp_name, len);
pbuf += len;
*pbuf++ = '(';
_PyUnicodeWriter_Init(&writer);
writer.overallocate = 1;
/* count 5 characters per item: "x=1, " */
writer.min_length = (PyUnicode_GET_LENGTH(type_name) + 1
+ VISIBLE_SIZE(obj) * 5 + 1);
for (i=0; i < VISIBLE_SIZE(obj); i++) {
PyObject *val, *repr;
const char *cname, *crepr;
if (_PyUnicodeWriter_WriteStr(&writer, type_name) < 0) {
Py_DECREF(type_name);
goto error;
}
Py_DECREF(type_name);
cname = typ->tp_members[i].name;
if (cname == NULL) {
if (_PyUnicodeWriter_WriteChar(&writer, '(') < 0) {
goto error;
}
for (Py_ssize_t i=0; i < VISIBLE_SIZE(obj); i++) {
if (i > 0) {
/* Write ", " */
if (_PyUnicodeWriter_WriteASCIIString(&writer, ", ", 2) < 0) {
goto error;
}
}
/* Write "name=repr" */
const char *name_utf8 = typ->tp_members[i].name;
if (name_utf8 == NULL) {
PyErr_Format(PyExc_SystemError, "In structseq_repr(), member %zd name is NULL"
" for type %.500s", i, typ->tp_name);
return NULL;
}
val = PyStructSequence_GET_ITEM(obj, i);
repr = PyObject_Repr(val);
if (repr == NULL)
return NULL;
crepr = PyUnicode_AsUTF8(repr);
if (crepr == NULL) {
Py_DECREF(repr);
return NULL;
goto error;
}
/* + 3: keep space for "=" and ", " */
len = strlen(cname) + strlen(crepr) + 3;
if ((pbuf+len) <= endofbuf) {
strcpy(pbuf, cname);
pbuf += strlen(cname);
*pbuf++ = '=';
strcpy(pbuf, crepr);
pbuf += strlen(crepr);
*pbuf++ = ',';
*pbuf++ = ' ';
removelast = 1;
Py_DECREF(repr);
PyObject *name = PyUnicode_DecodeUTF8(name_utf8, strlen(name_utf8), NULL);
if (name == NULL) {
goto error;
}
else {
strcpy(pbuf, "...");
pbuf += 3;
removelast = 0;
Py_DECREF(repr);
break;
if (_PyUnicodeWriter_WriteStr(&writer, name) < 0) {
Py_DECREF(name);
goto error;
}
}
if (removelast) {
/* overwrite last ", " */
pbuf-=2;
}
*pbuf++ = ')';
*pbuf = '\0';
Py_DECREF(name);
return PyUnicode_FromString(buf);
if (_PyUnicodeWriter_WriteChar(&writer, '=') < 0) {
goto error;
}
PyObject *value = PyStructSequence_GET_ITEM(obj, i);
assert(value != NULL);
PyObject *repr = PyObject_Repr(value);
if (repr == NULL) {
goto error;
}
if (_PyUnicodeWriter_WriteStr(&writer, repr) < 0) {
Py_DECREF(repr);
goto error;
}
Py_DECREF(repr);
}
if (_PyUnicodeWriter_WriteChar(&writer, ')') < 0) {
goto error;
}
return _PyUnicodeWriter_Finish(&writer);
error:
_PyUnicodeWriter_Dealloc(&writer);
return NULL;
}
static PyObject *
structseq_reduce(PyStructSequence* self, PyObject *Py_UNUSED(ignored))
{