Mailman 3 October 2011 - Python-checkins

cpython: Optimize PyUnicode_Copy(): don't recompute maximum character
by victor.stinner 30 Sep '11

30 Sep '11

http://hg.python.org/cpython/rev/b47e8c50a6a0 changeset: 72551:b47e8c50a6a0 user: Victor Stinner <victor.stinner(a)haypocalc.com> date: Sat Oct 01 01:34:32 2011 +0200 summary: Optimize PyUnicode_Copy(): don't recompute maximum character files: Objects/unicodeobject.c | 31 ++++++++++++++++++++++++++-- 1 files changed, 28 insertions(+), 3 deletions(-) diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c --- a/Objects/unicodeobject.c +++ b/Objects/unicodeobject.c @@ -1212,15 +1212,40 @@ PyObject* PyUnicode_Copy(PyObject *unicode) { + Py_ssize_t size; + PyObject *copy; + void *data; + if (!PyUnicode_Check(unicode)) { PyErr_BadInternalCall(); return NULL; } if (PyUnicode_READY(unicode)) return NULL; - return PyUnicode_FromKindAndData(PyUnicode_KIND(unicode), - PyUnicode_DATA(unicode), - PyUnicode_GET_LENGTH(unicode)); + + size = PyUnicode_GET_LENGTH(unicode); + copy = PyUnicode_New(size, PyUnicode_MAX_CHAR_VALUE(unicode)); + if (!copy) + return NULL; + assert(PyUnicode_KIND(copy) == PyUnicode_KIND(unicode)); + + data = PyUnicode_DATA(unicode); + switch (PyUnicode_KIND(unicode)) + { + case PyUnicode_1BYTE_KIND: + memcpy(PyUnicode_1BYTE_DATA(copy), data, size); + break; + case PyUnicode_2BYTE_KIND: + memcpy(PyUnicode_2BYTE_DATA(copy), data, sizeof(Py_UCS2) * size); + break; + case PyUnicode_4BYTE_KIND: + memcpy(PyUnicode_4BYTE_DATA(copy), data, sizeof(Py_UCS4) * size); + break; + default: + assert(0); + break; + } + return copy; } -- Repository URL: http://hg.python.org/cpython

1 0

cpython: PyUnicode_FromObject() reuses PyUnicode_Copy()
by victor.stinner 30 Sep '11

30 Sep '11

http://hg.python.org/cpython/rev/d94b0b371878 changeset: 72549:d94b0b371878 user: Victor Stinner <victor.stinner(a)haypocalc.com> date: Sat Oct 01 01:16:59 2011 +0200 summary: PyUnicode_FromObject() reuses PyUnicode_Copy() * PyUnicode_Copy() is faster than substring() * Fix also a compiler warning files: Objects/unicodeobject.c | 6 ++---- 1 files changed, 2 insertions(+), 4 deletions(-) diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c --- a/Objects/unicodeobject.c +++ b/Objects/unicodeobject.c @@ -2052,9 +2052,7 @@ if (PyUnicode_Check(obj)) { /* For a Unicode subtype that's not a Unicode object, return a true Unicode object with the same data. */ - if (PyUnicode_READY(obj) == -1) - return NULL; - return substring((PyUnicodeObject *)obj, 0, PyUnicode_GET_LENGTH(obj)); + return PyUnicode_Copy(obj); } PyErr_Format(PyExc_TypeError, "Can't convert '%.100s' object to str implicitly", @@ -11465,7 +11463,7 @@ return (PyObject*) self; } else - return PyUnicode_Copy(self); + return PyUnicode_Copy((PyObject*)self); } fill = width - _PyUnicode_LENGTH(self); -- Repository URL: http://hg.python.org/cpython

1 0

cpython: Remove commented code: str+=str is no more super-optimized
by victor.stinner 30 Sep '11

30 Sep '11

http://hg.python.org/cpython/rev/df6deb7bb772 changeset: 72550:df6deb7bb772 user: Victor Stinner <victor.stinner(a)haypocalc.com> date: Sat Oct 01 01:26:08 2011 +0200 summary: Remove commented code: str+=str is no more super-optimized files: Python/ceval.c | 118 +----------------------------------- 1 files changed, 6 insertions(+), 112 deletions(-) diff --git a/Python/ceval.c b/Python/ceval.c --- a/Python/ceval.c +++ b/Python/ceval.c @@ -136,8 +136,6 @@ static int import_all_from(PyObject *, PyObject *); static void format_exc_check_arg(PyObject *, const char *, PyObject *); static void format_exc_unbound(PyCodeObject *co, int oparg); -static PyObject * unicode_concatenate(PyObject *, PyObject *, - PyFrameObject *, unsigned char *); static PyObject * special_lookup(PyObject *, char *, PyObject **); #define NAME_ERROR_MSG \ @@ -1509,17 +1507,11 @@ TARGET(BINARY_ADD) w = POP(); v = TOP(); - if (PyUnicode_CheckExact(v) && - PyUnicode_CheckExact(w)) { - x = unicode_concatenate(v, w, f, next_instr); - /* unicode_concatenate consumed the ref to v */ - goto skip_decref_vx; - } - else { + if (PyUnicode_Check(v) && PyUnicode_Check(w)) + x = PyUnicode_Concat(v, w); + else x = PyNumber_Add(v, w); - } Py_DECREF(v); - skip_decref_vx: Py_DECREF(w); SET_TOP(x); if (x != NULL) DISPATCH(); @@ -1670,17 +1662,11 @@ TARGET(INPLACE_ADD) w = POP(); v = TOP(); - if (PyUnicode_CheckExact(v) && - PyUnicode_CheckExact(w)) { - x = unicode_concatenate(v, w, f, next_instr); - /* unicode_concatenate consumed the ref to v */ - goto skip_decref_v; - } - else { + if (PyUnicode_Check(v) && PyUnicode_Check(w)) + x = PyUnicode_Concat(v, w); + else x = PyNumber_InPlaceAdd(v, w); - } Py_DECREF(v); - skip_decref_v: Py_DECREF(w); SET_TOP(x); if (x != NULL) DISPATCH(); @@ -4515,98 +4501,6 @@ } } -static PyObject * -unicode_concatenate(PyObject *v, PyObject *w, - PyFrameObject *f, unsigned char *next_instr) -{ - /* This function implements 'variable += expr' when both arguments - are (Unicode) strings. */ - - w = PyUnicode_Concat(v, w); - Py_DECREF(v); - return w; - - /* XXX: This optimization is currently disabled as unicode objects in the - new flexible representation are not in-place resizable anymore. */ -#if 0 - Py_ssize_t v_len = PyUnicode_GET_SIZE(v); - Py_ssize_t w_len = PyUnicode_GET_SIZE(w); - Py_ssize_t new_len = v_len + w_len; - if (new_len < 0) { - PyErr_SetString(PyExc_OverflowError, - "strings are too large to concat"); - return NULL; - } - - if (Py_REFCNT(v) == 2) { - /* In the common case, there are 2 references to the value - * stored in 'variable' when the += is performed: one on the - * value stack (in 'v') and one still stored in the - * 'variable'. We try to delete the variable now to reduce - * the refcnt to 1. - */ - switch (*next_instr) { - case STORE_FAST: - { - int oparg = PEEKARG(); - PyObject **fastlocals = f->f_localsplus; - if (GETLOCAL(oparg) == v) - SETLOCAL(oparg, NULL); - break; - } - case STORE_DEREF: - { - PyObject **freevars = (f->f_localsplus + - f->f_code->co_nlocals); - PyObject *c = freevars[PEEKARG()]; - if (PyCell_GET(c) == v) - PyCell_Set(c, NULL); - break; - } - case STORE_NAME: - { - PyObject *names = f->f_code->co_names; - PyObject *name = GETITEM(names, PEEKARG()); - PyObject *locals = f->f_locals; - if (PyDict_CheckExact(locals) && - PyDict_GetItem(locals, name) == v) { - if (PyDict_DelItem(locals, name) != 0) { - PyErr_Clear(); - } - } - break; - } - } - } - - if (Py_REFCNT(v) == 1 && !PyUnicode_CHECK_INTERNED(v) && - !PyUnicode_IS_COMPACT((PyUnicodeObject *)v)) { - /* Now we own the last reference to 'v', so we can resize it - * in-place. - */ - if (PyUnicode_Resize(&v, new_len) != 0) { - /* XXX if PyUnicode_Resize() fails, 'v' has been - * deallocated so it cannot be put back into - * 'variable'. The MemoryError is raised when there - * is no value in 'variable', which might (very - * remotely) be a cause of incompatibilities. - */ - return NULL; - } - /* copy 'w' into the newly allocated area of 'v' */ - memcpy(PyUnicode_AS_UNICODE(v) + v_len, - PyUnicode_AS_UNICODE(w), w_len*sizeof(Py_UNICODE)); - return v; - } - else { - /* When in-place resizing is not an option. */ - w = PyUnicode_Concat(v, w); - Py_DECREF(v); - return w; - } -#endif -} - #ifdef DYNAMIC_EXECUTION_PROFILE static PyObject * -- Repository URL: http://hg.python.org/cpython

1 0