1
# Copyright (C) 2008, 2009 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18
"""Pyrex extensions to btree node parsing."""
21
cdef extern from "python-compat.h":
24
cdef extern from "stdlib.h":
25
ctypedef unsigned size_t
27
cdef extern from "Python.h":
28
ctypedef int Py_ssize_t # Required for older pyrex versions
29
ctypedef struct PyObject:
31
int PyList_Append(object lst, object item) except -1
33
char *PyString_AsString(object p) except NULL
34
object PyString_FromStringAndSize(char *, Py_ssize_t)
35
PyObject *PyString_FromStringAndSize_ptr "PyString_FromStringAndSize" (char *, Py_ssize_t)
36
int PyString_CheckExact(object s)
37
int PyString_CheckExact_ptr "PyString_CheckExact" (PyObject *)
38
Py_ssize_t PyString_Size(object p)
39
Py_ssize_t PyString_GET_SIZE_ptr "PyString_GET_SIZE" (PyObject *)
40
char * PyString_AS_STRING_ptr "PyString_AS_STRING" (PyObject *)
41
int PyString_AsStringAndSize_ptr(PyObject *, char **buf, Py_ssize_t *len)
42
void PyString_InternInPlace(PyObject **)
43
int PyTuple_CheckExact(object t)
44
Py_ssize_t PyTuple_GET_SIZE(object t)
45
PyObject *PyTuple_GET_ITEM_ptr_object "PyTuple_GET_ITEM" (object tpl, int index)
46
void Py_DECREF_ptr "Py_DECREF" (PyObject *)
48
cdef extern from "string.h":
49
void *memcpy(void *dest, void *src, size_t n)
50
void *memchr(void *s, int c, size_t n)
52
# void *memrchr(void *s, int c, size_t n)
53
int strncmp(char *s1, char *s2, size_t n)
56
# TODO: Find some way to import this from _dirstate_helpers
57
cdef void* _my_memrchr(void *s, int c, size_t n): # cannot_raise
58
# memrchr seems to be a GNU extension, so we have to implement it ourselves
59
# It is not present in any win32 standard library
71
# TODO: Import this from _dirstate_helpers when it is merged
72
cdef object safe_string_from_size(char *s, Py_ssize_t size):
75
'tried to create a string with an invalid size: %d @0x%x'
77
return PyString_FromStringAndSize(s, size)
80
cdef object safe_interned_string_from_size(char *s, Py_ssize_t size):
84
'tried to create a string with an invalid size: %d @0x%x'
86
py_str = PyString_FromStringAndSize_ptr(s, size)
87
PyString_InternInPlace(&py_str)
88
result = <object>py_str
89
# Casting a PyObject* to an <object> triggers an INCREF from Pyrex, so we
90
# DECREF it to avoid geting immortal strings
95
cdef class BTreeLeafParser:
96
"""Parse the leaf nodes of a BTree index.
98
:ivar bytes: The PyString object containing the uncompressed text for the
100
:ivar key_length: An integer describing how many pieces the keys have for
102
:ivar ref_list_length: An integer describing how many references this index
104
:ivar keys: A PyList of keys found in this node.
106
:ivar _cur_str: A pointer to the start of the next line to parse
107
:ivar _end_str: A pointer to the end of bytes
108
:ivar _start: Pointer to the location within the current line while
110
:ivar _header_found: True when we have parsed the header for this node
115
cdef int ref_list_length
120
# The current start point for parsing
123
cdef int _header_found
125
def __init__(self, bytes, key_length, ref_list_length):
127
self.key_length = key_length
128
self.ref_list_length = ref_list_length
132
self._header_found = 0
134
cdef extract_key(self, char * last):
137
:param last: points at the byte after the last byte permitted for the
141
cdef int loop_counter
145
while loop_counter < self.key_length:
146
loop_counter = loop_counter + 1
148
temp_ptr = <char*>memchr(self._start, c'\0', last - self._start)
150
if loop_counter == self.key_length:
155
failure_string = ("invalid key, wanted segment from " +
156
repr(safe_string_from_size(self._start,
157
last - self._start)))
158
raise AssertionError(failure_string)
159
# capture the key string
160
# TODO: Consider using PyIntern_FromString, the only caveat is that
161
# it assumes a NULL-terminated string, so we have to check if
162
# temp_ptr[0] == c'\0' or some other char.
163
key_element = safe_interned_string_from_size(self._start,
164
temp_ptr - self._start)
165
# advance our pointer
166
self._start = temp_ptr + 1
167
PyList_Append(key_segments, key_element)
168
return tuple(key_segments)
170
cdef int process_line(self) except -1:
171
"""Process a line in the bytes."""
175
cdef char *next_start
176
cdef int loop_counter
178
self._start = self._cur_str
179
# Find the next newline
180
last = <char*>memchr(self._start, c'\n', self._end_str - self._start)
182
# Process until the end of the file
184
self._cur_str = self._end_str
186
# And the next string is right after it
187
self._cur_str = last + 1
188
# The last character is right before the '\n'
190
if last == self._start:
193
if last < self._start:
194
# Unexpected error condition - fail
195
raise AssertionError("last < self._start")
196
if 0 == self._header_found:
197
# The first line in a leaf node is the header "type=leaf\n"
198
if strncmp("type=leaf", self._start, last - self._start) == 0:
199
self._header_found = 1
202
raise AssertionError('Node did not start with "type=leaf": %r'
203
% (safe_string_from_size(self._start, last - self._start)))
205
key = self.extract_key(last)
206
# find the value area
207
temp_ptr = <char*>_my_memrchr(self._start, c'\0', last - self._start)
210
raise AssertionError("Failed to find the value area")
212
# capture the value string
213
value = safe_string_from_size(temp_ptr + 1, last - temp_ptr - 1)
214
# shrink the references end point
216
if self.ref_list_length:
219
while loop_counter < self.ref_list_length:
221
# extract a reference list
222
loop_counter = loop_counter + 1
223
if last < self._start:
224
raise AssertionError("last < self._start")
225
# find the next reference list end point:
226
temp_ptr = <char*>memchr(self._start, c'\t', last - self._start)
228
# Only valid for the last list
229
if loop_counter != self.ref_list_length:
231
raise AssertionError(
232
"invalid key, loop_counter != self.ref_list_length")
234
# scan to the end of the ref list area
238
# scan to the end of this ref list
240
next_start = temp_ptr + 1
241
# Now, there may be multiple keys in the ref list.
242
while self._start < ref_ptr:
243
# loop finding keys and extracting them
244
temp_ptr = <char*>memchr(self._start, c'\r',
245
ref_ptr - self._start)
247
# key runs to the end
249
PyList_Append(ref_list, self.extract_key(temp_ptr))
250
PyList_Append(ref_lists, tuple(ref_list))
251
# prepare for the next reference list
252
self._start = next_start
253
ref_lists = tuple(ref_lists)
254
node_value = (value, ref_lists)
256
if last != self._start:
257
# unexpected reference data present
258
raise AssertionError("unexpected reference data present")
259
node_value = (value, ())
260
PyList_Append(self.keys, (key, node_value))
264
cdef Py_ssize_t byte_count
265
if not PyString_CheckExact(self.bytes):
266
raise AssertionError('self.bytes is not a string.')
267
byte_count = PyString_Size(self.bytes)
268
self._cur_str = PyString_AsString(self.bytes)
269
# This points to the last character in the string
270
self._end_str = self._cur_str + byte_count
271
while self._cur_str < self._end_str:
276
def _parse_leaf_lines(bytes, key_length, ref_list_length):
277
parser = BTreeLeafParser(bytes, key_length, ref_list_length)
278
return parser.parse()
281
def _flatten_node(node, reference_lists):
282
"""Convert a node into the serialized form.
284
:param node: A tuple representing a node:
285
(index, key_tuple, value, references)
286
:param reference_lists: Does this index have reference lists?
287
:return: (string_key, flattened)
288
string_key The serialized key for referencing this node
289
flattened A string with the serialized form for the contents
291
cdef int have_reference_lists
292
cdef Py_ssize_t flat_len
293
cdef Py_ssize_t key_len
294
cdef Py_ssize_t node_len
297
cdef Py_ssize_t value_len
299
cdef Py_ssize_t refs_len
300
cdef Py_ssize_t next_len
301
cdef int first_ref_list
302
cdef int first_reference
304
cdef PyObject *ref_bit
305
cdef Py_ssize_t ref_bit_len
307
if not PyTuple_CheckExact(node):
308
raise TypeError('We expected a tuple() for node not: %s'
310
node_len = PyTuple_GET_SIZE(node)
311
have_reference_lists = reference_lists
312
if have_reference_lists:
314
raise ValueError('With ref_lists, we expected 4 entries not: %s'
317
raise ValueError('Without ref_lists, we need at least 3 entries not: %s'
319
# I don't expect that we can do faster than string.join()
320
string_key = '\0'.join(<object>PyTuple_GET_ITEM_ptr_object(node, 1))
322
# TODO: instead of using string joins, precompute the final string length,
323
# and then malloc a single string and copy everything in.
325
# TODO: We probably want to use PySequenceFast, because we have lists and
326
# tuples, but we aren't sure which we will get.
328
# line := string_key NULL flat_refs NULL value LF
329
# string_key := BYTES (NULL BYTES)*
330
# flat_refs := ref_list (TAB ref_list)*
331
# ref_list := ref (CR ref)*
332
# ref := BYTES (NULL BYTES)*
335
if have_reference_lists:
336
# Figure out how many bytes it will take to store the references
337
ref_lists = <object>PyTuple_GET_ITEM_ptr_object(node, 3)
338
next_len = len(ref_lists) # TODO: use a Py function
340
# If there are no nodes, we don't need to do any work
341
# Otherwise we will need (len - 1) '\t' characters to separate
342
# the reference lists
343
refs_len = refs_len + (next_len - 1)
344
for ref_list in ref_lists:
345
next_len = len(ref_list)
347
# We will need (len - 1) '\r' characters to separate the
349
refs_len = refs_len + (next_len - 1)
350
for reference in ref_list:
351
if not PyTuple_CheckExact(reference):
353
'We expect references to be tuples not: %s'
355
next_len = PyTuple_GET_SIZE(reference)
357
# We will need (len - 1) '\x00' characters to
358
# separate the reference key
359
refs_len = refs_len + (next_len - 1)
360
for i from 0 <= i < next_len:
361
ref_bit = PyTuple_GET_ITEM_ptr_object(reference, i)
362
if not PyString_CheckExact_ptr(ref_bit):
363
raise TypeError('We expect reference bits'
364
' to be strings not: %s'
365
% type(<object>ref_bit))
366
refs_len = refs_len + PyString_GET_SIZE_ptr(ref_bit)
368
# So we have the (key NULL refs NULL value LF)
369
key_len = PyString_Size(string_key)
370
val = PyTuple_GET_ITEM_ptr_object(node, 2)
371
if not PyString_CheckExact_ptr(val):
372
raise TypeError('Expected a plain str for value not: %s'
374
value = PyString_AS_STRING_ptr(val)
375
value_len = PyString_GET_SIZE_ptr(val)
376
flat_len = (key_len + 1 + refs_len + 1 + value_len + 1)
377
line = PyString_FromStringAndSize(NULL, flat_len)
378
# Get a pointer to the new buffer
379
out = PyString_AsString(line)
380
memcpy(out, PyString_AsString(string_key), key_len)
386
for ref_list in ref_lists:
387
if first_ref_list == 0:
392
for reference in ref_list:
393
if first_reference == 0:
397
next_len = PyTuple_GET_SIZE(reference)
398
for i from 0 <= i < next_len:
402
ref_bit = PyTuple_GET_ITEM_ptr_object(reference, i)
403
ref_bit_len = PyString_GET_SIZE_ptr(ref_bit)
404
memcpy(out, PyString_AS_STRING_ptr(ref_bit), ref_bit_len)
405
out = out + ref_bit_len
408
memcpy(out, value, value_len)
409
out = out + value_len
411
return string_key, line