Add a format specifier %R to PyUnicode_FromFormat(), which embeds
the result of a call to PyObject_Repr() into the string. This makes
it possible to simplify many repr implementations.
PyUnicode_FromFormat() uses two steps to create the final string: A first
pass through the format string determines the size of the final string and
a second pass creates the string. To avoid calling PyObject_Repr() twice
for each %R specifier, PyObject_Repr() is called during the size
calculation step and the results are stored in an array (whose size is
determined at the start by counting %R specifiers).
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
index b46093e..f2c5afa 100644
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -484,6 +484,9 @@
PyUnicode_FromFormatV(const char *format, va_list vargs)
{
va_list count;
+ Py_ssize_t callcount = 0;
+ PyObject **callresults = NULL;
+ PyObject **callresult;
Py_ssize_t n = 0;
const char* f;
Py_UNICODE *s;
@@ -501,7 +504,23 @@
count = vargs;
#endif
#endif
- /* step 1: figure out how large a buffer we need */
+ /* step 1: count the number of %R format specifications
+ * (we call PyObject_Repr() for these objects once during step 3
+ * and put the result in an array) */
+ for (f = format; *f; f++) {
+ if (*f == '%' && *(f+1)=='R')
+ ++callcount;
+ }
+ /* step 2: allocate memory for the results of PyObject_Repr() calls */
+ if (callcount) {
+ callresults = PyMem_Malloc(sizeof(PyObject *)*callcount);
+ if (!callresults) {
+ PyErr_NoMemory();
+ return NULL;
+ }
+ callresult = callresults;
+ }
+ /* step 3: figure out how large a buffer we need */
for (f = format; *f; f++) {
if (*f == '%') {
const char* p = f;
@@ -539,6 +558,19 @@
n += PyUnicode_GET_SIZE(obj);
break;
}
+ case 'R':
+ {
+ PyObject *obj = va_arg(count, PyObject *);
+ PyObject *repr;
+ assert(obj);
+ repr = PyObject_Repr(obj);
+ if (!repr)
+ goto fail;
+ n += PyUnicode_GET_SIZE(repr);
+ /* Remember the repr and switch to the next slot */
+ *callresult++ = repr;
+ break;
+ }
case 'p':
(void) va_arg(count, int);
/* maximum 64-bit pointer representation:
@@ -562,14 +594,16 @@
n++;
}
expand:
- /* step 2: fill the buffer */
+ /* step 4: fill the buffer */
/* Since we've analyzed how much space we need for the worst case,
- we don't have to resize the string. */
+ we don't have to resize the string.
+ There can be no errors beyond this point. */
string = PyUnicode_FromUnicode(NULL, n);
if (!string)
return NULL;
s = PyUnicode_AS_UNICODE(string);
+ callresult = callresults;
for (f = format; *f; f++) {
if (*f == '%') {
@@ -649,6 +683,21 @@
*s++ = ucopy[upos++];
break;
}
+ case 'R':
+ {
+ /* unused, since we already have the result */
+ (void) va_arg(vargs, PyObject *);
+ Py_UNICODE *ucopy = PyUnicode_AS_UNICODE(*callresult);
+ Py_ssize_t usize = PyUnicode_GET_SIZE(*callresult);
+ Py_ssize_t upos;
+ for (upos = 0; upos<usize;)
+ *s++ = ucopy[upos++];
+ /* We're done with the repr() => forget it */
+ Py_DECREF(*callresult);
+ /* switch to next repr() result */
+ ++callresult;
+ break;
+ }
case 'p':
sprintf(buffer, "%p", va_arg(vargs, void*));
/* %p is ill-defined: ensure leading 0x. */
@@ -673,8 +722,20 @@
}
end:
+ if (callresults)
+ PyMem_Free(callresults);
_PyUnicode_Resize(&string, s - PyUnicode_AS_UNICODE(string));
return string;
+ fail:
+ if (callresults) {
+ PyObject **callresult2 = callresults;
+ while (callresult2 <= callresult) {
+ Py_DECREF(*callresult2);
+ ++callresult2;
+ }
+ PyMem_Free(callresults);
+ }
+ return NULL;
}
#undef appendstring