1 /* |
|
2 dirs.c - dynamic directory diddling for dirstates |
|
3 |
|
4 Copyright 2013 Facebook |
|
5 |
|
6 This software may be used and distributed according to the terms of |
|
7 the GNU General Public License, incorporated herein by reference. |
|
8 */ |
|
9 |
|
10 #define PY_SSIZE_T_CLEAN |
|
11 #include <Python.h> |
|
12 #include "util.h" |
|
13 |
|
14 #ifdef IS_PY3K |
|
15 #define PYLONG_VALUE(o) ((PyLongObject *)o)->ob_digit[1] |
|
16 #else |
|
17 #define PYLONG_VALUE(o) PyInt_AS_LONG(o) |
|
18 #endif |
|
19 |
|
20 /* |
|
21 * This is a multiset of directory names, built from the files that |
|
22 * appear in a dirstate or manifest. |
|
23 * |
|
24 * A few implementation notes: |
|
25 * |
|
26 * We modify Python integers for refcounting, but those integers are |
|
27 * never visible to Python code. |
|
28 * |
|
29 * We mutate strings in-place, but leave them immutable once they can |
|
30 * be seen by Python code. |
|
31 */ |
|
32 typedef struct { |
|
33 PyObject_HEAD |
|
34 PyObject *dict; |
|
35 } dirsObject; |
|
36 |
|
37 static inline Py_ssize_t _finddir(const char *path, Py_ssize_t pos) |
|
38 { |
|
39 while (pos != -1) { |
|
40 if (path[pos] == '/') |
|
41 break; |
|
42 pos -= 1; |
|
43 } |
|
44 |
|
45 return pos; |
|
46 } |
|
47 |
|
48 static int _addpath(PyObject *dirs, PyObject *path) |
|
49 { |
|
50 const char *cpath = PyBytes_AS_STRING(path); |
|
51 Py_ssize_t pos = PyBytes_GET_SIZE(path); |
|
52 PyObject *key = NULL; |
|
53 int ret = -1; |
|
54 |
|
55 /* This loop is super critical for performance. That's why we inline |
|
56 * access to Python structs instead of going through a supported API. |
|
57 * The implementation, therefore, is heavily dependent on CPython |
|
58 * implementation details. We also commit violations of the Python |
|
59 * "protocol" such as mutating immutable objects. But since we only |
|
60 * mutate objects created in this function or in other well-defined |
|
61 * locations, the references are known so these violations should go |
|
62 * unnoticed. The code for adjusting the length of a PyBytesObject is |
|
63 * essentially a minimal version of _PyBytes_Resize. */ |
|
64 while ((pos = _finddir(cpath, pos - 1)) != -1) { |
|
65 PyObject *val; |
|
66 |
|
67 /* It's likely that every prefix already has an entry |
|
68 in our dict. Try to avoid allocating and |
|
69 deallocating a string for each prefix we check. */ |
|
70 if (key != NULL) |
|
71 ((PyBytesObject *)key)->ob_shash = -1; |
|
72 else { |
|
73 /* Force Python to not reuse a small shared string. */ |
|
74 key = PyBytes_FromStringAndSize(cpath, |
|
75 pos < 2 ? 2 : pos); |
|
76 if (key == NULL) |
|
77 goto bail; |
|
78 } |
|
79 /* Py_SIZE(o) refers to the ob_size member of the struct. Yes, |
|
80 * assigning to what looks like a function seems wrong. */ |
|
81 Py_SIZE(key) = pos; |
|
82 ((PyBytesObject *)key)->ob_sval[pos] = '\0'; |
|
83 |
|
84 val = PyDict_GetItem(dirs, key); |
|
85 if (val != NULL) { |
|
86 PYLONG_VALUE(val) += 1; |
|
87 break; |
|
88 } |
|
89 |
|
90 /* Force Python to not reuse a small shared int. */ |
|
91 #ifdef IS_PY3K |
|
92 val = PyLong_FromLong(0x1eadbeef); |
|
93 #else |
|
94 val = PyInt_FromLong(0x1eadbeef); |
|
95 #endif |
|
96 |
|
97 if (val == NULL) |
|
98 goto bail; |
|
99 |
|
100 PYLONG_VALUE(val) = 1; |
|
101 ret = PyDict_SetItem(dirs, key, val); |
|
102 Py_DECREF(val); |
|
103 if (ret == -1) |
|
104 goto bail; |
|
105 Py_CLEAR(key); |
|
106 } |
|
107 ret = 0; |
|
108 |
|
109 bail: |
|
110 Py_XDECREF(key); |
|
111 |
|
112 return ret; |
|
113 } |
|
114 |
|
115 static int _delpath(PyObject *dirs, PyObject *path) |
|
116 { |
|
117 char *cpath = PyBytes_AS_STRING(path); |
|
118 Py_ssize_t pos = PyBytes_GET_SIZE(path); |
|
119 PyObject *key = NULL; |
|
120 int ret = -1; |
|
121 |
|
122 while ((pos = _finddir(cpath, pos - 1)) != -1) { |
|
123 PyObject *val; |
|
124 |
|
125 key = PyBytes_FromStringAndSize(cpath, pos); |
|
126 |
|
127 if (key == NULL) |
|
128 goto bail; |
|
129 |
|
130 val = PyDict_GetItem(dirs, key); |
|
131 if (val == NULL) { |
|
132 PyErr_SetString(PyExc_ValueError, |
|
133 "expected a value, found none"); |
|
134 goto bail; |
|
135 } |
|
136 |
|
137 if (--PYLONG_VALUE(val) <= 0) { |
|
138 if (PyDict_DelItem(dirs, key) == -1) |
|
139 goto bail; |
|
140 } else |
|
141 break; |
|
142 Py_CLEAR(key); |
|
143 } |
|
144 ret = 0; |
|
145 |
|
146 bail: |
|
147 Py_XDECREF(key); |
|
148 |
|
149 return ret; |
|
150 } |
|
151 |
|
152 static int dirs_fromdict(PyObject *dirs, PyObject *source, char skipchar) |
|
153 { |
|
154 PyObject *key, *value; |
|
155 Py_ssize_t pos = 0; |
|
156 |
|
157 while (PyDict_Next(source, &pos, &key, &value)) { |
|
158 if (!PyBytes_Check(key)) { |
|
159 PyErr_SetString(PyExc_TypeError, "expected string key"); |
|
160 return -1; |
|
161 } |
|
162 if (skipchar) { |
|
163 if (!dirstate_tuple_check(value)) { |
|
164 PyErr_SetString(PyExc_TypeError, |
|
165 "expected a dirstate tuple"); |
|
166 return -1; |
|
167 } |
|
168 if (((dirstateTupleObject *)value)->state == skipchar) |
|
169 continue; |
|
170 } |
|
171 |
|
172 if (_addpath(dirs, key) == -1) |
|
173 return -1; |
|
174 } |
|
175 |
|
176 return 0; |
|
177 } |
|
178 |
|
179 static int dirs_fromiter(PyObject *dirs, PyObject *source) |
|
180 { |
|
181 PyObject *iter, *item = NULL; |
|
182 int ret; |
|
183 |
|
184 iter = PyObject_GetIter(source); |
|
185 if (iter == NULL) |
|
186 return -1; |
|
187 |
|
188 while ((item = PyIter_Next(iter)) != NULL) { |
|
189 if (!PyBytes_Check(item)) { |
|
190 PyErr_SetString(PyExc_TypeError, "expected string"); |
|
191 break; |
|
192 } |
|
193 |
|
194 if (_addpath(dirs, item) == -1) |
|
195 break; |
|
196 Py_CLEAR(item); |
|
197 } |
|
198 |
|
199 ret = PyErr_Occurred() ? -1 : 0; |
|
200 Py_DECREF(iter); |
|
201 Py_XDECREF(item); |
|
202 return ret; |
|
203 } |
|
204 |
|
205 /* |
|
206 * Calculate a refcounted set of directory names for the files in a |
|
207 * dirstate. |
|
208 */ |
|
209 static int dirs_init(dirsObject *self, PyObject *args) |
|
210 { |
|
211 PyObject *dirs = NULL, *source = NULL; |
|
212 char skipchar = 0; |
|
213 int ret = -1; |
|
214 |
|
215 self->dict = NULL; |
|
216 |
|
217 if (!PyArg_ParseTuple(args, "|Oc:__init__", &source, &skipchar)) |
|
218 return -1; |
|
219 |
|
220 dirs = PyDict_New(); |
|
221 |
|
222 if (dirs == NULL) |
|
223 return -1; |
|
224 |
|
225 if (source == NULL) |
|
226 ret = 0; |
|
227 else if (PyDict_Check(source)) |
|
228 ret = dirs_fromdict(dirs, source, skipchar); |
|
229 else if (skipchar) |
|
230 PyErr_SetString(PyExc_ValueError, |
|
231 "skip character is only supported " |
|
232 "with a dict source"); |
|
233 else |
|
234 ret = dirs_fromiter(dirs, source); |
|
235 |
|
236 if (ret == -1) |
|
237 Py_XDECREF(dirs); |
|
238 else |
|
239 self->dict = dirs; |
|
240 |
|
241 return ret; |
|
242 } |
|
243 |
|
244 PyObject *dirs_addpath(dirsObject *self, PyObject *args) |
|
245 { |
|
246 PyObject *path; |
|
247 |
|
248 if (!PyArg_ParseTuple(args, "O!:addpath", &PyBytes_Type, &path)) |
|
249 return NULL; |
|
250 |
|
251 if (_addpath(self->dict, path) == -1) |
|
252 return NULL; |
|
253 |
|
254 Py_RETURN_NONE; |
|
255 } |
|
256 |
|
257 static PyObject *dirs_delpath(dirsObject *self, PyObject *args) |
|
258 { |
|
259 PyObject *path; |
|
260 |
|
261 if (!PyArg_ParseTuple(args, "O!:delpath", &PyBytes_Type, &path)) |
|
262 return NULL; |
|
263 |
|
264 if (_delpath(self->dict, path) == -1) |
|
265 return NULL; |
|
266 |
|
267 Py_RETURN_NONE; |
|
268 } |
|
269 |
|
270 static int dirs_contains(dirsObject *self, PyObject *value) |
|
271 { |
|
272 return PyBytes_Check(value) ? PyDict_Contains(self->dict, value) : 0; |
|
273 } |
|
274 |
|
275 static void dirs_dealloc(dirsObject *self) |
|
276 { |
|
277 Py_XDECREF(self->dict); |
|
278 PyObject_Del(self); |
|
279 } |
|
280 |
|
281 static PyObject *dirs_iter(dirsObject *self) |
|
282 { |
|
283 return PyObject_GetIter(self->dict); |
|
284 } |
|
285 |
|
286 static PySequenceMethods dirs_sequence_methods; |
|
287 |
|
288 static PyMethodDef dirs_methods[] = { |
|
289 {"addpath", (PyCFunction)dirs_addpath, METH_VARARGS, "add a path"}, |
|
290 {"delpath", (PyCFunction)dirs_delpath, METH_VARARGS, "remove a path"}, |
|
291 {NULL} /* Sentinel */ |
|
292 }; |
|
293 |
|
294 static PyTypeObject dirsType = { PyVarObject_HEAD_INIT(NULL, 0) }; |
|
295 |
|
296 void dirs_module_init(PyObject *mod) |
|
297 { |
|
298 dirs_sequence_methods.sq_contains = (objobjproc)dirs_contains; |
|
299 dirsType.tp_name = "parsers.dirs"; |
|
300 dirsType.tp_new = PyType_GenericNew; |
|
301 dirsType.tp_basicsize = sizeof(dirsObject); |
|
302 dirsType.tp_dealloc = (destructor)dirs_dealloc; |
|
303 dirsType.tp_as_sequence = &dirs_sequence_methods; |
|
304 dirsType.tp_flags = Py_TPFLAGS_DEFAULT; |
|
305 dirsType.tp_doc = "dirs"; |
|
306 dirsType.tp_iter = (getiterfunc)dirs_iter; |
|
307 dirsType.tp_methods = dirs_methods; |
|
308 dirsType.tp_init = (initproc)dirs_init; |
|
309 |
|
310 if (PyType_Ready(&dirsType) < 0) |
|
311 return; |
|
312 Py_INCREF(&dirsType); |
|
313 |
|
314 PyModule_AddObject(mod, "dirs", (PyObject *)&dirsType); |
|
315 } |
|