Implementation of PEP 3101, Advanced String Formatting.

Known issues:

The string.Formatter class, as discussed in the PEP, is incomplete.

Error handling needs to conform to the PEP.

Need to fix this warning that I introduced in Python/formatter_unicode.c:
Objects/stringlib/unicodedefs.h:26: warning: `STRINGLIB_CMP' defined but not used

Need to make sure sign formatting is correct, more tests needed.

Need to remove '()' sign formatting, left over from an earlier version of the PEP.

4 files changed, 121 insertions, 2 deletions

diff --git a/Python/Python-ast.c b/Python/Python-ast.c
index 605a152..4c6f42c 100644
--- a/Python/Python-ast.c
+++ b/Python/Python-ast.c
@@ -2,7 +2,7 @@
 
 
 /*
-   __version__ 56266.
+   __version__ .
 
    This module must be committed separately after each AST grammar change;
    The __version__ number is set to the revision number of the commit
@@ -3179,7 +3179,7 @@ init_ast(void)
         if (PyDict_SetItemString(d, "AST", (PyObject*)AST_type) < 0) return;
         if (PyModule_AddIntConstant(m, "PyCF_ONLY_AST", PyCF_ONLY_AST) < 0)
                 return;
-        if (PyModule_AddStringConstant(m, "__version__", "56266") < 0)
+        if (PyModule_AddStringConstant(m, "__version__", "") < 0)
                 return;
         if (PyDict_SetItemString(d, "mod", (PyObject*)mod_type) < 0) return;
         if (PyDict_SetItemString(d, "Module", (PyObject*)Module_type) < 0)
diff --git a/Python/bltinmodule.c b/Python/bltinmodule.c
index d087e9c..17f5b59 100644
--- a/Python/bltinmodule.c
+++ b/Python/bltinmodule.c
@@ -275,6 +275,61 @@ for which the predicate (a Boolean function) returns true.\n\
 If the predicate is None, 'lambda x: bool(x)' is assumed.\n\
 (This is identical to itertools.ifilter().)");
 
+static PyObject *
+builtin_format(PyObject *self, PyObject *args)
+{
+        static PyObject * format_str = NULL;
+        PyObject *value;
+        PyObject *spec;
+        PyObject *meth;
+        PyObject *result;
+
+        /* Initialize cached value */
+        if (format_str == NULL) {
+                /* Initialize static variable needed by _PyType_Lookup */
+                format_str = PyUnicode_FromString("__format__");
+                if (format_str == NULL)
+                        return NULL;
+        }
+
+        if (!PyArg_ParseTuple(args, "OO:format", &value, &spec))
+               return NULL;
+
+        /* Make sure the type is initialized.  float gets initialized late */
+        if (Py_Type(value)->tp_dict == NULL)
+                if (PyType_Ready(Py_Type(value)) < 0)
+                    return NULL;
+
+        /* Find the (unbound!) __format__ method (a borrowed reference) */
+        meth = _PyType_Lookup(Py_Type(value), format_str);
+        if (meth == NULL) {
+                PyErr_Format(PyExc_TypeError,
+                             "Type %.100s doesn't define __format__",
+                             Py_Type(value)->tp_name);
+                return NULL;
+        }
+
+        /* And call it, binding it to the value */
+        result = PyObject_CallFunctionObjArgs(meth, value, spec, NULL);
+
+#if 0
+        /* XXX this is segfaulting, not sure why.  find out later! */
+	if (!PyUnicode_Check(result)) {
+                PyErr_SetString(PyExc_TypeError,
+                                "__format__ method did not return string");
+                Py_DECREF(result);
+                return NULL;
+        }
+#endif
+
+        return result;
+}
+
+
+PyDoc_STRVAR(format_doc,
+"format(value, format_spec) -> string\n\
+\n\
+Returns value.__format__(format_spec).");
 
 static PyObject *
 builtin_chr8(PyObject *self, PyObject *args)
@@ -1676,6 +1731,7 @@ static PyMethodDef builtin_methods[] = {
  	{"eval",	builtin_eval,       METH_VARARGS, eval_doc},
 	{"exec",        builtin_exec,       METH_VARARGS, exec_doc},
  	{"filter",	builtin_filter,     METH_VARARGS, filter_doc},
+ 	{"format",	builtin_format,     METH_VARARGS, format_doc},
  	{"getattr",	builtin_getattr,    METH_VARARGS, getattr_doc},
  	{"globals",	(PyCFunction)builtin_globals,    METH_NOARGS, globals_doc},
  	{"hasattr",	builtin_hasattr,    METH_VARARGS, hasattr_doc},
diff --git a/Python/formatter_unicode.c b/Python/formatter_unicode.c
new file mode 100644
index 0000000..114fe30
--- /dev/null
+++ b/Python/formatter_unicode.c
@@ -0,0 +1,13 @@
+/* implements the unicode (as opposed to string) version of the
+   built-in formatters for string, int, float.  that is, the versions
+   of int.__float__, etc., that take and return unicode objects */
+
+#include "Python.h"
+#include "formatter_unicode.h"
+
+#include "../Objects/stringlib/unicodedefs.h"
+
+#define FORMAT_STRING unicode_unicode__format__
+#define FORMAT_LONG   unicode_long__format__
+#define FORMAT_FLOAT  unicode_float__format__
+#include "../Objects/stringlib/formatter.h"
diff --git a/Python/sysmodule.c b/Python/sysmodule.c
index 6ccd3e9..10a7427 100644
--- a/Python/sysmodule.c
+++ b/Python/sysmodule.c
@@ -660,6 +660,54 @@ sys_current_frames(PyObject *self, PyObject *noargs)
 	return _PyThread_CurrentFrames();
 }
 
+/* sys_formatter_iterator is used to implement
+   string.Formatter.vformat.  it parses a string and returns tuples
+   describing the parsed elements.  see unicodeobject.c's
+   _unicodeformatter_iterator for details */
+static PyObject *
+sys_formatter_iterator(PyObject *self, PyObject *args)
+{
+        /* in 2.6, check type and dispatch to unicode or string
+           accordingly */
+        PyObject *str;
+
+        if (!PyArg_ParseTuple(args, "O:_formatter_iterator", &str))
+                return NULL;
+
+        if (!PyUnicode_Check(str)) {
+                PyErr_SetString(PyExc_TypeError,
+                                "_formatter_iterator expects unicode object");
+                return NULL;
+        }
+
+        return _unicodeformatter_iterator(str);
+}
+
+/* sys_formatter_lookup is used to implement string.Formatter.vformat.
+   it takes an PEP 3101 "field name", args, and kwargs, and returns a
+   tuple (index, name, object).  see unicodeobject.c's
+   _unicodeformatter_lookup for details */
+static PyObject *
+sys_formatter_lookup(PyObject *self, PyObject *args)
+{
+        PyObject *field_name;
+        PyObject *arg_args;
+        PyObject *kwargs;
+
+        if (!PyArg_ParseTuple(args, "OOO:_formatter_lookup", &field_name,
+                              &arg_args, &kwargs))
+                return NULL;
+
+        if (!PyUnicode_Check(field_name)) {
+                PyErr_SetString(PyExc_TypeError,
+                                "_formatter_lookup expects unicode object");
+                return NULL;
+        }
+
+        return _unicodeformatter_lookup(field_name, arg_args, kwargs);
+}
+
+
 PyDoc_STRVAR(call_tracing_doc,
 "call_tracing(func, args) -> object\n\
 \n\
@@ -724,6 +772,8 @@ static PyMethodDef sys_methods[] = {
 	 callstats_doc},
 	{"_current_frames", sys_current_frames, METH_NOARGS,
 	 current_frames_doc},
+        {"_formatter_parser", sys_formatter_iterator, METH_VARARGS},
+        {"_formatter_lookup", sys_formatter_lookup, METH_VARARGS},
 	{"displayhook",	sys_displayhook, METH_O, displayhook_doc},
 	{"exc_info",	sys_exc_info, METH_NOARGS, exc_info_doc},
 	{"excepthook",	sys_excepthook, METH_VARARGS, excepthook_doc},