Show More
This diff has been collapsed as it changes many lines, (606 lines changed) Show them Hide them | |||||
@@ -1,1035 +1,1195 b'' | |||||
1 | # Copyright 2014-present Facebook, Inc. |
|
1 | # Copyright 2014-present Facebook, Inc. | |
2 | # All rights reserved. |
|
2 | # All rights reserved. | |
3 | # |
|
3 | # | |
4 | # Redistribution and use in source and binary forms, with or without |
|
4 | # Redistribution and use in source and binary forms, with or without | |
5 | # modification, are permitted provided that the following conditions are met: |
|
5 | # modification, are permitted provided that the following conditions are met: | |
6 | # |
|
6 | # | |
7 | # * Redistributions of source code must retain the above copyright notice, |
|
7 | # * Redistributions of source code must retain the above copyright notice, | |
8 | # this list of conditions and the following disclaimer. |
|
8 | # this list of conditions and the following disclaimer. | |
9 | # |
|
9 | # | |
10 | # * Redistributions in binary form must reproduce the above copyright notice, |
|
10 | # * Redistributions in binary form must reproduce the above copyright notice, | |
11 | # this list of conditions and the following disclaimer in the documentation |
|
11 | # this list of conditions and the following disclaimer in the documentation | |
12 | # and/or other materials provided with the distribution. |
|
12 | # and/or other materials provided with the distribution. | |
13 | # |
|
13 | # | |
14 | # * Neither the name Facebook nor the names of its contributors may be used to |
|
14 | # * Neither the name Facebook nor the names of its contributors may be used to | |
15 | # endorse or promote products derived from this software without specific |
|
15 | # endorse or promote products derived from this software without specific | |
16 | # prior written permission. |
|
16 | # prior written permission. | |
17 | # |
|
17 | # | |
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
|
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" | |
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
|
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
|
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE | |
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
|
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE | |
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
|
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
|
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR | |
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
|
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER | |
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
|
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, | |
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
|
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
|
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
28 |
|
28 | |||
29 | from __future__ import absolute_import |
|
|||
30 | from __future__ import division |
|
|||
31 | from __future__ import print_function |
|
|||
32 | # no unicode literals |
|
29 | # no unicode literals | |
|
30 | from __future__ import absolute_import, division, print_function | |||
33 |
|
31 | |||
34 | import inspect |
|
32 | import inspect | |
35 | import math |
|
33 | import math | |
36 | import os |
|
34 | import os | |
37 | import socket |
|
35 | import socket | |
38 | import subprocess |
|
36 | import subprocess | |
39 | import time |
|
37 | import time | |
40 |
|
38 | |||
|
39 | from . import capabilities, compat, encoding, load | |||
|
40 | ||||
|
41 | ||||
41 | # Sometimes it's really hard to get Python extensions to compile, |
|
42 | # Sometimes it's really hard to get Python extensions to compile, | |
42 | # so fall back to a pure Python implementation. |
|
43 | # so fall back to a pure Python implementation. | |
43 | try: |
|
44 | try: | |
44 | from . import bser |
|
45 | from . import bser | |
|
46 | ||||
45 | # Demandimport causes modules to be loaded lazily. Force the load now |
|
47 | # Demandimport causes modules to be loaded lazily. Force the load now | |
46 | # so that we can fall back on pybser if bser doesn't exist |
|
48 | # so that we can fall back on pybser if bser doesn't exist | |
47 | bser.pdu_info |
|
49 | bser.pdu_info | |
48 | except ImportError: |
|
50 | except ImportError: | |
49 | from . import pybser as bser |
|
51 | from . import pybser as bser | |
50 |
|
52 | |||
51 | from mercurial.utils import ( |
|
|||
52 | procutil, |
|
|||
53 | ) |
|
|||
54 |
|
53 | |||
55 | from mercurial import ( |
|
54 | if os.name == "nt": | |
56 | pycompat, |
|
|||
57 | ) |
|
|||
58 |
|
||||
59 | from . import ( |
|
|||
60 | capabilities, |
|
|||
61 | compat, |
|
|||
62 | encoding, |
|
|||
63 | load, |
|
|||
64 | ) |
|
|||
65 |
|
||||
66 |
|
||||
67 | if os.name == 'nt': |
|
|||
68 | import ctypes |
|
55 | import ctypes | |
69 | import ctypes.wintypes |
|
56 | import ctypes.wintypes | |
70 |
|
57 | |||
71 | wintypes = ctypes.wintypes |
|
58 | wintypes = ctypes.wintypes | |
72 | GENERIC_READ = 0x80000000 |
|
59 | GENERIC_READ = 0x80000000 | |
73 | GENERIC_WRITE = 0x40000000 |
|
60 | GENERIC_WRITE = 0x40000000 | |
74 | FILE_FLAG_OVERLAPPED = 0x40000000 |
|
61 | FILE_FLAG_OVERLAPPED = 0x40000000 | |
75 | OPEN_EXISTING = 3 |
|
62 | OPEN_EXISTING = 3 | |
76 | INVALID_HANDLE_VALUE = -1 |
|
63 | INVALID_HANDLE_VALUE = ctypes.c_void_p(-1).value | |
77 | FORMAT_MESSAGE_FROM_SYSTEM = 0x00001000 |
|
64 | FORMAT_MESSAGE_FROM_SYSTEM = 0x00001000 | |
78 | FORMAT_MESSAGE_ALLOCATE_BUFFER = 0x00000100 |
|
65 | FORMAT_MESSAGE_ALLOCATE_BUFFER = 0x00000100 | |
79 | FORMAT_MESSAGE_IGNORE_INSERTS = 0x00000200 |
|
66 | FORMAT_MESSAGE_IGNORE_INSERTS = 0x00000200 | |
80 | WAIT_FAILED = 0xFFFFFFFF |
|
67 | WAIT_FAILED = 0xFFFFFFFF | |
81 | WAIT_TIMEOUT = 0x00000102 |
|
68 | WAIT_TIMEOUT = 0x00000102 | |
82 | WAIT_OBJECT_0 = 0x00000000 |
|
69 | WAIT_OBJECT_0 = 0x00000000 | |
83 | WAIT_IO_COMPLETION = 0x000000C0 |
|
70 | WAIT_IO_COMPLETION = 0x000000C0 | |
84 | INFINITE = 0xFFFFFFFF |
|
71 | INFINITE = 0xFFFFFFFF | |
85 |
|
72 | |||
86 | # Overlapped I/O operation is in progress. (997) |
|
73 | # Overlapped I/O operation is in progress. (997) | |
87 | ERROR_IO_PENDING = 0x000003E5 |
|
74 | ERROR_IO_PENDING = 0x000003E5 | |
88 |
|
75 | |||
89 | # The pointer size follows the architecture |
|
76 | # The pointer size follows the architecture | |
90 | # We use WPARAM since this type is already conditionally defined |
|
77 | # We use WPARAM since this type is already conditionally defined | |
91 | ULONG_PTR = ctypes.wintypes.WPARAM |
|
78 | ULONG_PTR = ctypes.wintypes.WPARAM | |
92 |
|
79 | |||
93 | class OVERLAPPED(ctypes.Structure): |
|
80 | class OVERLAPPED(ctypes.Structure): | |
94 | _fields_ = [ |
|
81 | _fields_ = [ | |
95 |
("Internal", ULONG_PTR), |
|
82 | ("Internal", ULONG_PTR), | |
96 | ("Offset", wintypes.DWORD), ("OffsetHigh", wintypes.DWORD), |
|
83 | ("InternalHigh", ULONG_PTR), | |
97 |
(" |
|
84 | ("Offset", wintypes.DWORD), | |
|
85 | ("OffsetHigh", wintypes.DWORD), | |||
|
86 | ("hEvent", wintypes.HANDLE), | |||
98 | ] |
|
87 | ] | |
99 |
|
88 | |||
100 | def __init__(self): |
|
89 | def __init__(self): | |
101 | self.Internal = 0 |
|
90 | self.Internal = 0 | |
102 | self.InternalHigh = 0 |
|
91 | self.InternalHigh = 0 | |
103 | self.Offset = 0 |
|
92 | self.Offset = 0 | |
104 | self.OffsetHigh = 0 |
|
93 | self.OffsetHigh = 0 | |
105 | self.hEvent = 0 |
|
94 | self.hEvent = 0 | |
106 |
|
95 | |||
107 | LPDWORD = ctypes.POINTER(wintypes.DWORD) |
|
96 | LPDWORD = ctypes.POINTER(wintypes.DWORD) | |
108 |
|
97 | |||
109 | CreateFile = ctypes.windll.kernel32.CreateFileA |
|
98 | CreateFile = ctypes.windll.kernel32.CreateFileA | |
110 | CreateFile.argtypes = [wintypes.LPSTR, wintypes.DWORD, wintypes.DWORD, |
|
99 | CreateFile.argtypes = [ | |
111 | wintypes.LPVOID, wintypes.DWORD, wintypes.DWORD, |
|
100 | wintypes.LPSTR, | |
112 | wintypes.HANDLE] |
|
101 | wintypes.DWORD, | |
|
102 | wintypes.DWORD, | |||
|
103 | wintypes.LPVOID, | |||
|
104 | wintypes.DWORD, | |||
|
105 | wintypes.DWORD, | |||
|
106 | wintypes.HANDLE, | |||
|
107 | ] | |||
113 | CreateFile.restype = wintypes.HANDLE |
|
108 | CreateFile.restype = wintypes.HANDLE | |
114 |
|
109 | |||
115 | CloseHandle = ctypes.windll.kernel32.CloseHandle |
|
110 | CloseHandle = ctypes.windll.kernel32.CloseHandle | |
116 | CloseHandle.argtypes = [wintypes.HANDLE] |
|
111 | CloseHandle.argtypes = [wintypes.HANDLE] | |
117 | CloseHandle.restype = wintypes.BOOL |
|
112 | CloseHandle.restype = wintypes.BOOL | |
118 |
|
113 | |||
119 | ReadFile = ctypes.windll.kernel32.ReadFile |
|
114 | ReadFile = ctypes.windll.kernel32.ReadFile | |
120 | ReadFile.argtypes = [wintypes.HANDLE, wintypes.LPVOID, wintypes.DWORD, |
|
115 | ReadFile.argtypes = [ | |
121 | LPDWORD, ctypes.POINTER(OVERLAPPED)] |
|
116 | wintypes.HANDLE, | |
|
117 | wintypes.LPVOID, | |||
|
118 | wintypes.DWORD, | |||
|
119 | LPDWORD, | |||
|
120 | ctypes.POINTER(OVERLAPPED), | |||
|
121 | ] | |||
122 | ReadFile.restype = wintypes.BOOL |
|
122 | ReadFile.restype = wintypes.BOOL | |
123 |
|
123 | |||
124 | WriteFile = ctypes.windll.kernel32.WriteFile |
|
124 | WriteFile = ctypes.windll.kernel32.WriteFile | |
125 | WriteFile.argtypes = [wintypes.HANDLE, wintypes.LPVOID, wintypes.DWORD, |
|
125 | WriteFile.argtypes = [ | |
126 | LPDWORD, ctypes.POINTER(OVERLAPPED)] |
|
126 | wintypes.HANDLE, | |
|
127 | wintypes.LPVOID, | |||
|
128 | wintypes.DWORD, | |||
|
129 | LPDWORD, | |||
|
130 | ctypes.POINTER(OVERLAPPED), | |||
|
131 | ] | |||
127 | WriteFile.restype = wintypes.BOOL |
|
132 | WriteFile.restype = wintypes.BOOL | |
128 |
|
133 | |||
129 | GetLastError = ctypes.windll.kernel32.GetLastError |
|
134 | GetLastError = ctypes.windll.kernel32.GetLastError | |
130 | GetLastError.argtypes = [] |
|
135 | GetLastError.argtypes = [] | |
131 | GetLastError.restype = wintypes.DWORD |
|
136 | GetLastError.restype = wintypes.DWORD | |
132 |
|
137 | |||
133 | SetLastError = ctypes.windll.kernel32.SetLastError |
|
138 | SetLastError = ctypes.windll.kernel32.SetLastError | |
134 | SetLastError.argtypes = [wintypes.DWORD] |
|
139 | SetLastError.argtypes = [wintypes.DWORD] | |
135 | SetLastError.restype = None |
|
140 | SetLastError.restype = None | |
136 |
|
141 | |||
137 | FormatMessage = ctypes.windll.kernel32.FormatMessageA |
|
142 | FormatMessage = ctypes.windll.kernel32.FormatMessageA | |
138 | FormatMessage.argtypes = [wintypes.DWORD, wintypes.LPVOID, wintypes.DWORD, |
|
143 | FormatMessage.argtypes = [ | |
139 | wintypes.DWORD, ctypes.POINTER(wintypes.LPSTR), |
|
144 | wintypes.DWORD, | |
140 | wintypes.DWORD, wintypes.LPVOID] |
|
145 | wintypes.LPVOID, | |
|
146 | wintypes.DWORD, | |||
|
147 | wintypes.DWORD, | |||
|
148 | ctypes.POINTER(wintypes.LPSTR), | |||
|
149 | wintypes.DWORD, | |||
|
150 | wintypes.LPVOID, | |||
|
151 | ] | |||
141 | FormatMessage.restype = wintypes.DWORD |
|
152 | FormatMessage.restype = wintypes.DWORD | |
142 |
|
153 | |||
143 | LocalFree = ctypes.windll.kernel32.LocalFree |
|
154 | LocalFree = ctypes.windll.kernel32.LocalFree | |
144 |
|
155 | |||
145 | GetOverlappedResult = ctypes.windll.kernel32.GetOverlappedResult |
|
156 | GetOverlappedResult = ctypes.windll.kernel32.GetOverlappedResult | |
146 |
GetOverlappedResult.argtypes = [ |
|
157 | GetOverlappedResult.argtypes = [ | |
147 | ctypes.POINTER(OVERLAPPED), LPDWORD, |
|
158 | wintypes.HANDLE, | |
148 | wintypes.BOOL] |
|
159 | ctypes.POINTER(OVERLAPPED), | |
|
160 | LPDWORD, | |||
|
161 | wintypes.BOOL, | |||
|
162 | ] | |||
149 | GetOverlappedResult.restype = wintypes.BOOL |
|
163 | GetOverlappedResult.restype = wintypes.BOOL | |
150 |
|
164 | |||
151 |
GetOverlappedResultEx = getattr( |
|
165 | GetOverlappedResultEx = getattr( | |
152 | 'GetOverlappedResultEx', None) |
|
166 | ctypes.windll.kernel32, "GetOverlappedResultEx", None | |
|
167 | ) | |||
153 | if GetOverlappedResultEx is not None: |
|
168 | if GetOverlappedResultEx is not None: | |
154 |
GetOverlappedResultEx.argtypes = [ |
|
169 | GetOverlappedResultEx.argtypes = [ | |
155 | ctypes.POINTER(OVERLAPPED), LPDWORD, |
|
170 | wintypes.HANDLE, | |
156 | wintypes.DWORD, wintypes.BOOL] |
|
171 | ctypes.POINTER(OVERLAPPED), | |
|
172 | LPDWORD, | |||
|
173 | wintypes.DWORD, | |||
|
174 | wintypes.BOOL, | |||
|
175 | ] | |||
157 | GetOverlappedResultEx.restype = wintypes.BOOL |
|
176 | GetOverlappedResultEx.restype = wintypes.BOOL | |
158 |
|
177 | |||
159 | WaitForSingleObjectEx = ctypes.windll.kernel32.WaitForSingleObjectEx |
|
178 | WaitForSingleObjectEx = ctypes.windll.kernel32.WaitForSingleObjectEx | |
160 |
WaitForSingleObjectEx.argtypes = [ |
|
179 | WaitForSingleObjectEx.argtypes = [ | |
|
180 | wintypes.HANDLE, | |||
|
181 | wintypes.DWORD, | |||
|
182 | wintypes.BOOL, | |||
|
183 | ] | |||
161 | WaitForSingleObjectEx.restype = wintypes.DWORD |
|
184 | WaitForSingleObjectEx.restype = wintypes.DWORD | |
162 |
|
185 | |||
163 | CreateEvent = ctypes.windll.kernel32.CreateEventA |
|
186 | CreateEvent = ctypes.windll.kernel32.CreateEventA | |
164 | CreateEvent.argtypes = [LPDWORD, wintypes.BOOL, wintypes.BOOL, |
|
187 | CreateEvent.argtypes = [ | |
165 | wintypes.LPSTR] |
|
188 | LPDWORD, | |
|
189 | wintypes.BOOL, | |||
|
190 | wintypes.BOOL, | |||
|
191 | wintypes.LPSTR, | |||
|
192 | ] | |||
166 | CreateEvent.restype = wintypes.HANDLE |
|
193 | CreateEvent.restype = wintypes.HANDLE | |
167 |
|
194 | |||
168 | # Windows Vista is the minimum supported client for CancelIoEx. |
|
195 | # Windows Vista is the minimum supported client for CancelIoEx. | |
169 | CancelIoEx = ctypes.windll.kernel32.CancelIoEx |
|
196 | CancelIoEx = ctypes.windll.kernel32.CancelIoEx | |
170 | CancelIoEx.argtypes = [wintypes.HANDLE, ctypes.POINTER(OVERLAPPED)] |
|
197 | CancelIoEx.argtypes = [wintypes.HANDLE, ctypes.POINTER(OVERLAPPED)] | |
171 | CancelIoEx.restype = wintypes.BOOL |
|
198 | CancelIoEx.restype = wintypes.BOOL | |
172 |
|
199 | |||
173 | # 2 bytes marker, 1 byte int size, 8 bytes int64 value |
|
200 | # 2 bytes marker, 1 byte int size, 8 bytes int64 value | |
174 | sniff_len = 13 |
|
201 | sniff_len = 13 | |
175 |
|
202 | |||
176 | # This is a helper for debugging the client. |
|
203 | # This is a helper for debugging the client. | |
177 | _debugging = False |
|
204 | _debugging = False | |
178 | if _debugging: |
|
205 | if _debugging: | |
179 |
|
206 | |||
180 | def log(fmt, *args): |
|
207 | def log(fmt, *args): | |
181 |
print( |
|
208 | print( | |
182 | (time.strftime("%a, %d %b %Y %H:%M:%S", time.gmtime()), |
|
209 | "[%s] %s" | |
183 |
|
|
210 | % ( | |
|
211 | time.strftime("%a, %d %b %Y %H:%M:%S", time.gmtime()), | |||
|
212 | fmt % args[:], | |||
|
213 | ) | |||
|
214 | ) | |||
|
215 | ||||
|
216 | ||||
184 | else: |
|
217 | else: | |
185 |
|
218 | |||
186 | def log(fmt, *args): |
|
219 | def log(fmt, *args): | |
187 | pass |
|
220 | pass | |
188 |
|
221 | |||
189 |
|
222 | |||
190 | def _win32_strerror(err): |
|
223 | def _win32_strerror(err): | |
191 | """ expand a win32 error code into a human readable message """ |
|
224 | """ expand a win32 error code into a human readable message """ | |
192 |
|
225 | |||
193 | # FormatMessage will allocate memory and assign it here |
|
226 | # FormatMessage will allocate memory and assign it here | |
194 | buf = ctypes.c_char_p() |
|
227 | buf = ctypes.c_char_p() | |
195 | FormatMessage( |
|
228 | FormatMessage( | |
196 |
FORMAT_MESSAGE_FROM_SYSTEM |
|
229 | FORMAT_MESSAGE_FROM_SYSTEM | |
197 | | FORMAT_MESSAGE_IGNORE_INSERTS, None, err, 0, buf, 0, None) |
|
230 | | FORMAT_MESSAGE_ALLOCATE_BUFFER | |
|
231 | | FORMAT_MESSAGE_IGNORE_INSERTS, | |||
|
232 | None, | |||
|
233 | err, | |||
|
234 | 0, | |||
|
235 | buf, | |||
|
236 | 0, | |||
|
237 | None, | |||
|
238 | ) | |||
198 | try: |
|
239 | try: | |
199 | return buf.value |
|
240 | return buf.value | |
200 | finally: |
|
241 | finally: | |
201 | LocalFree(buf) |
|
242 | LocalFree(buf) | |
202 |
|
243 | |||
203 |
|
244 | |||
204 | class WatchmanError(Exception): |
|
245 | class WatchmanError(Exception): | |
205 | def __init__(self, msg=None, cmd=None): |
|
246 | def __init__(self, msg=None, cmd=None): | |
206 | self.msg = msg |
|
247 | self.msg = msg | |
207 | self.cmd = cmd |
|
248 | self.cmd = cmd | |
208 |
|
249 | |||
209 | def setCommand(self, cmd): |
|
250 | def setCommand(self, cmd): | |
210 | self.cmd = cmd |
|
251 | self.cmd = cmd | |
211 |
|
252 | |||
212 | def __str__(self): |
|
253 | def __str__(self): | |
213 | if self.cmd: |
|
254 | if self.cmd: | |
214 |
return |
|
255 | return "%s, while executing %s" % (self.msg, self.cmd) | |
215 | return self.msg |
|
256 | return self.msg | |
216 |
|
257 | |||
217 |
|
258 | |||
|
259 | class BSERv1Unsupported(WatchmanError): | |||
|
260 | pass | |||
|
261 | ||||
|
262 | ||||
|
263 | class UseAfterFork(WatchmanError): | |||
|
264 | pass | |||
|
265 | ||||
|
266 | ||||
218 | class WatchmanEnvironmentError(WatchmanError): |
|
267 | class WatchmanEnvironmentError(WatchmanError): | |
219 | def __init__(self, msg, errno, errmsg, cmd=None): |
|
268 | def __init__(self, msg, errno, errmsg, cmd=None): | |
220 | super(WatchmanEnvironmentError, self).__init__( |
|
269 | super(WatchmanEnvironmentError, self).__init__( | |
221 |
|
|
270 | "{0}: errno={1} errmsg={2}".format(msg, errno, errmsg), cmd | |
222 |
|
|
271 | ) | |
223 |
|
272 | |||
224 |
|
273 | |||
225 | class SocketConnectError(WatchmanError): |
|
274 | class SocketConnectError(WatchmanError): | |
226 | def __init__(self, sockpath, exc): |
|
275 | def __init__(self, sockpath, exc): | |
227 | super(SocketConnectError, self).__init__( |
|
276 | super(SocketConnectError, self).__init__( | |
228 |
|
|
277 | "unable to connect to %s: %s" % (sockpath, exc) | |
|
278 | ) | |||
229 | self.sockpath = sockpath |
|
279 | self.sockpath = sockpath | |
230 | self.exc = exc |
|
280 | self.exc = exc | |
231 |
|
281 | |||
232 |
|
282 | |||
233 | class SocketTimeout(WatchmanError): |
|
283 | class SocketTimeout(WatchmanError): | |
234 | """A specialized exception raised for socket timeouts during communication to/from watchman. |
|
284 | """A specialized exception raised for socket timeouts during communication to/from watchman. | |
235 | This makes it easier to implement non-blocking loops as callers can easily distinguish |
|
285 | This makes it easier to implement non-blocking loops as callers can easily distinguish | |
236 | between a routine timeout and an actual error condition. |
|
286 | between a routine timeout and an actual error condition. | |
237 |
|
287 | |||
238 | Note that catching WatchmanError will also catch this as it is a super-class, so backwards |
|
288 | Note that catching WatchmanError will also catch this as it is a super-class, so backwards | |
239 | compatibility in exception handling is preserved. |
|
289 | compatibility in exception handling is preserved. | |
240 | """ |
|
290 | """ | |
241 |
|
291 | |||
242 |
|
292 | |||
243 | class CommandError(WatchmanError): |
|
293 | class CommandError(WatchmanError): | |
244 | """error returned by watchman |
|
294 | """error returned by watchman | |
245 |
|
295 | |||
246 | self.msg is the message returned by watchman. |
|
296 | self.msg is the message returned by watchman. | |
247 | """ |
|
297 | """ | |
|
298 | ||||
248 | def __init__(self, msg, cmd=None): |
|
299 | def __init__(self, msg, cmd=None): | |
249 | super(CommandError, self).__init__( |
|
300 | super(CommandError, self).__init__( | |
250 |
|
|
301 | "watchman command error: %s" % (msg,), cmd | |
251 | cmd, |
|
|||
252 | ) |
|
302 | ) | |
253 |
|
303 | |||
254 |
|
304 | |||
255 | class Transport(object): |
|
305 | class Transport(object): | |
256 | """ communication transport to the watchman server """ |
|
306 | """ communication transport to the watchman server """ | |
|
307 | ||||
257 | buf = None |
|
308 | buf = None | |
258 |
|
309 | |||
259 | def close(self): |
|
310 | def close(self): | |
260 | """ tear it down """ |
|
311 | """ tear it down """ | |
261 | raise NotImplementedError() |
|
312 | raise NotImplementedError() | |
262 |
|
313 | |||
263 | def readBytes(self, size): |
|
314 | def readBytes(self, size): | |
264 | """ read size bytes """ |
|
315 | """ read size bytes """ | |
265 | raise NotImplementedError() |
|
316 | raise NotImplementedError() | |
266 |
|
317 | |||
267 | def write(self, buf): |
|
318 | def write(self, buf): | |
268 | """ write some data """ |
|
319 | """ write some data """ | |
269 | raise NotImplementedError() |
|
320 | raise NotImplementedError() | |
270 |
|
321 | |||
271 | def setTimeout(self, value): |
|
322 | def setTimeout(self, value): | |
272 | pass |
|
323 | pass | |
273 |
|
324 | |||
274 | def readLine(self): |
|
325 | def readLine(self): | |
275 | """ read a line |
|
326 | """ read a line | |
276 | Maintains its own buffer, callers of the transport should not mix |
|
327 | Maintains its own buffer, callers of the transport should not mix | |
277 | calls to readBytes and readLine. |
|
328 | calls to readBytes and readLine. | |
278 | """ |
|
329 | """ | |
279 | if self.buf is None: |
|
330 | if self.buf is None: | |
280 | self.buf = [] |
|
331 | self.buf = [] | |
281 |
|
332 | |||
282 | # Buffer may already have a line if we've received unilateral |
|
333 | # Buffer may already have a line if we've received unilateral | |
283 | # response(s) from the server |
|
334 | # response(s) from the server | |
284 | if len(self.buf) == 1 and b"\n" in self.buf[0]: |
|
335 | if len(self.buf) == 1 and b"\n" in self.buf[0]: | |
285 | (line, b) = self.buf[0].split(b"\n", 1) |
|
336 | (line, b) = self.buf[0].split(b"\n", 1) | |
286 | self.buf = [b] |
|
337 | self.buf = [b] | |
287 | return line |
|
338 | return line | |
288 |
|
339 | |||
289 | while True: |
|
340 | while True: | |
290 | b = self.readBytes(4096) |
|
341 | b = self.readBytes(4096) | |
291 | if b"\n" in b: |
|
342 | if b"\n" in b: | |
292 |
result = b |
|
343 | result = b"".join(self.buf) | |
293 | (line, b) = b.split(b"\n", 1) |
|
344 | (line, b) = b.split(b"\n", 1) | |
294 | self.buf = [b] |
|
345 | self.buf = [b] | |
295 | return result + line |
|
346 | return result + line | |
296 | self.buf.append(b) |
|
347 | self.buf.append(b) | |
297 |
|
348 | |||
298 |
|
349 | |||
299 | class Codec(object): |
|
350 | class Codec(object): | |
300 | """ communication encoding for the watchman server """ |
|
351 | """ communication encoding for the watchman server """ | |
|
352 | ||||
301 | transport = None |
|
353 | transport = None | |
302 |
|
354 | |||
303 | def __init__(self, transport): |
|
355 | def __init__(self, transport): | |
304 | self.transport = transport |
|
356 | self.transport = transport | |
305 |
|
357 | |||
306 | def receive(self): |
|
358 | def receive(self): | |
307 | raise NotImplementedError() |
|
359 | raise NotImplementedError() | |
308 |
|
360 | |||
309 | def send(self, *args): |
|
361 | def send(self, *args): | |
310 | raise NotImplementedError() |
|
362 | raise NotImplementedError() | |
311 |
|
363 | |||
312 | def setTimeout(self, value): |
|
364 | def setTimeout(self, value): | |
313 | self.transport.setTimeout(value) |
|
365 | self.transport.setTimeout(value) | |
314 |
|
366 | |||
315 |
|
367 | |||
316 | class UnixSocketTransport(Transport): |
|
368 | class UnixSocketTransport(Transport): | |
317 | """ local unix domain socket transport """ |
|
369 | """ local unix domain socket transport """ | |
|
370 | ||||
318 | sock = None |
|
371 | sock = None | |
319 |
|
372 | |||
320 |
def __init__(self, sockpath, timeout |
|
373 | def __init__(self, sockpath, timeout): | |
321 | self.sockpath = sockpath |
|
374 | self.sockpath = sockpath | |
322 | self.timeout = timeout |
|
375 | self.timeout = timeout | |
323 |
|
376 | |||
324 | sock = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) |
|
377 | sock = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) | |
325 | try: |
|
378 | try: | |
326 | sock.settimeout(self.timeout) |
|
379 | sock.settimeout(self.timeout) | |
327 | sock.connect(self.sockpath) |
|
380 | sock.connect(self.sockpath) | |
328 | self.sock = sock |
|
381 | self.sock = sock | |
329 | except socket.error as e: |
|
382 | except socket.error as e: | |
330 | sock.close() |
|
383 | sock.close() | |
331 | raise SocketConnectError(self.sockpath, e) |
|
384 | raise SocketConnectError(self.sockpath, e) | |
332 |
|
385 | |||
333 | def close(self): |
|
386 | def close(self): | |
334 |
self.sock |
|
387 | if self.sock: | |
335 |
self.sock |
|
388 | self.sock.close() | |
|
389 | self.sock = None | |||
336 |
|
390 | |||
337 | def setTimeout(self, value): |
|
391 | def setTimeout(self, value): | |
338 | self.timeout = value |
|
392 | self.timeout = value | |
339 | self.sock.settimeout(self.timeout) |
|
393 | self.sock.settimeout(self.timeout) | |
340 |
|
394 | |||
341 | def readBytes(self, size): |
|
395 | def readBytes(self, size): | |
342 | try: |
|
396 | try: | |
343 | buf = [self.sock.recv(size)] |
|
397 | buf = [self.sock.recv(size)] | |
344 | if not buf[0]: |
|
398 | if not buf[0]: | |
345 |
raise WatchmanError( |
|
399 | raise WatchmanError("empty watchman response") | |
346 | return buf[0] |
|
400 | return buf[0] | |
347 | except socket.timeout: |
|
401 | except socket.timeout: | |
348 |
raise SocketTimeout( |
|
402 | raise SocketTimeout("timed out waiting for response") | |
349 |
|
403 | |||
350 | def write(self, data): |
|
404 | def write(self, data): | |
351 | try: |
|
405 | try: | |
352 | self.sock.sendall(data) |
|
406 | self.sock.sendall(data) | |
353 | except socket.timeout: |
|
407 | except socket.timeout: | |
354 |
raise SocketTimeout( |
|
408 | raise SocketTimeout("timed out sending query command") | |
355 |
|
409 | |||
356 |
|
410 | |||
357 | def _get_overlapped_result_ex_impl(pipe, olap, nbytes, millis, alertable): |
|
411 | def _get_overlapped_result_ex_impl(pipe, olap, nbytes, millis, alertable): | |
358 | """ Windows 7 and earlier does not support GetOverlappedResultEx. The |
|
412 | """ Windows 7 and earlier does not support GetOverlappedResultEx. The | |
359 | alternative is to use GetOverlappedResult and wait for read or write |
|
413 | alternative is to use GetOverlappedResult and wait for read or write | |
360 | operation to complete. This is done be using CreateEvent and |
|
414 | operation to complete. This is done be using CreateEvent and | |
361 | WaitForSingleObjectEx. CreateEvent, WaitForSingleObjectEx |
|
415 | WaitForSingleObjectEx. CreateEvent, WaitForSingleObjectEx | |
362 | and GetOverlappedResult are all part of Windows API since WindowsXP. |
|
416 | and GetOverlappedResult are all part of Windows API since WindowsXP. | |
363 | This is the exact same implementation that can be found in the watchman |
|
417 | This is the exact same implementation that can be found in the watchman | |
364 | source code (see get_overlapped_result_ex_impl in stream_win.c). This |
|
418 | source code (see get_overlapped_result_ex_impl in stream_win.c). This | |
365 | way, maintenance should be simplified. |
|
419 | way, maintenance should be simplified. | |
366 | """ |
|
420 | """ | |
367 |
log( |
|
421 | log("Preparing to wait for maximum %dms", millis) | |
368 | if millis != 0: |
|
422 | if millis != 0: | |
369 | waitReturnCode = WaitForSingleObjectEx(olap.hEvent, millis, alertable) |
|
423 | waitReturnCode = WaitForSingleObjectEx(olap.hEvent, millis, alertable) | |
370 | if waitReturnCode == WAIT_OBJECT_0: |
|
424 | if waitReturnCode == WAIT_OBJECT_0: | |
371 | # Event is signaled, overlapped IO operation result should be available. |
|
425 | # Event is signaled, overlapped IO operation result should be available. | |
372 | pass |
|
426 | pass | |
373 | elif waitReturnCode == WAIT_IO_COMPLETION: |
|
427 | elif waitReturnCode == WAIT_IO_COMPLETION: | |
374 | # WaitForSingleObjectEx returnes because the system added an I/O completion |
|
428 | # WaitForSingleObjectEx returnes because the system added an I/O completion | |
375 | # routine or an asynchronous procedure call (APC) to the thread queue. |
|
429 | # routine or an asynchronous procedure call (APC) to the thread queue. | |
376 | SetLastError(WAIT_IO_COMPLETION) |
|
430 | SetLastError(WAIT_IO_COMPLETION) | |
377 | pass |
|
431 | pass | |
378 | elif waitReturnCode == WAIT_TIMEOUT: |
|
432 | elif waitReturnCode == WAIT_TIMEOUT: | |
379 | # We reached the maximum allowed wait time, the IO operation failed |
|
433 | # We reached the maximum allowed wait time, the IO operation failed | |
380 | # to complete in timely fashion. |
|
434 | # to complete in timely fashion. | |
381 | SetLastError(WAIT_TIMEOUT) |
|
435 | SetLastError(WAIT_TIMEOUT) | |
382 | return False |
|
436 | return False | |
383 | elif waitReturnCode == WAIT_FAILED: |
|
437 | elif waitReturnCode == WAIT_FAILED: | |
384 | # something went wrong calling WaitForSingleObjectEx |
|
438 | # something went wrong calling WaitForSingleObjectEx | |
385 | err = GetLastError() |
|
439 | err = GetLastError() | |
386 |
log( |
|
440 | log("WaitForSingleObjectEx failed: %s", _win32_strerror(err)) | |
387 | return False |
|
441 | return False | |
388 | else: |
|
442 | else: | |
389 | # unexpected situation deserving investigation. |
|
443 | # unexpected situation deserving investigation. | |
390 | err = GetLastError() |
|
444 | err = GetLastError() | |
391 |
log( |
|
445 | log("Unexpected error: %s", _win32_strerror(err)) | |
392 | return False |
|
446 | return False | |
393 |
|
447 | |||
394 | return GetOverlappedResult(pipe, olap, nbytes, False) |
|
448 | return GetOverlappedResult(pipe, olap, nbytes, False) | |
395 |
|
449 | |||
396 |
|
450 | |||
397 | class WindowsNamedPipeTransport(Transport): |
|
451 | class WindowsNamedPipeTransport(Transport): | |
398 | """ connect to a named pipe """ |
|
452 | """ connect to a named pipe """ | |
399 |
|
453 | |||
400 |
def __init__(self, sockpath, timeout |
|
454 | def __init__(self, sockpath, timeout): | |
401 | self.sockpath = sockpath |
|
455 | self.sockpath = sockpath | |
402 | self.timeout = int(math.ceil(timeout * 1000)) |
|
456 | self.timeout = int(math.ceil(timeout * 1000)) | |
403 | self._iobuf = None |
|
457 | self._iobuf = None | |
404 |
|
458 | |||
405 | self.pipe = CreateFile(sockpath, GENERIC_READ | GENERIC_WRITE, 0, None, |
|
459 | if compat.PYTHON3: | |
406 | OPEN_EXISTING, FILE_FLAG_OVERLAPPED, None) |
|
460 | sockpath = os.fsencode(sockpath) | |
|
461 | self.pipe = CreateFile( | |||
|
462 | sockpath, | |||
|
463 | GENERIC_READ | GENERIC_WRITE, | |||
|
464 | 0, | |||
|
465 | None, | |||
|
466 | OPEN_EXISTING, | |||
|
467 | FILE_FLAG_OVERLAPPED, | |||
|
468 | None, | |||
|
469 | ) | |||
407 |
|
470 | |||
408 | if self.pipe == INVALID_HANDLE_VALUE: |
|
471 | err = GetLastError() | |
|
472 | if self.pipe == INVALID_HANDLE_VALUE or self.pipe == 0: | |||
409 | self.pipe = None |
|
473 | self.pipe = None | |
410 | self._raise_win_err('failed to open pipe %s' % sockpath, |
|
474 | raise SocketConnectError(self.sockpath, self._make_win_err("", err)) | |
411 | GetLastError()) |
|
|||
412 |
|
475 | |||
413 | # event for the overlapped I/O operations |
|
476 | # event for the overlapped I/O operations | |
414 | self._waitable = CreateEvent(None, True, False, None) |
|
477 | self._waitable = CreateEvent(None, True, False, None) | |
|
478 | err = GetLastError() | |||
415 | if self._waitable is None: |
|
479 | if self._waitable is None: | |
416 |
self._raise_win_err( |
|
480 | self._raise_win_err("CreateEvent failed", err) | |
417 |
|
481 | |||
418 | self._get_overlapped_result_ex = GetOverlappedResultEx |
|
482 | self._get_overlapped_result_ex = GetOverlappedResultEx | |
419 | if (os.getenv('WATCHMAN_WIN7_COMPAT') == '1' or |
|
483 | if ( | |
420 | self._get_overlapped_result_ex is None): |
|
484 | os.getenv("WATCHMAN_WIN7_COMPAT") == "1" | |
|
485 | or self._get_overlapped_result_ex is None | |||
|
486 | ): | |||
421 | self._get_overlapped_result_ex = _get_overlapped_result_ex_impl |
|
487 | self._get_overlapped_result_ex = _get_overlapped_result_ex_impl | |
422 |
|
488 | |||
423 | def _raise_win_err(self, msg, err): |
|
489 | def _raise_win_err(self, msg, err): | |
424 | raise IOError('%s win32 error code: %d %s' % |
|
490 | raise self._make_win_err(msg, err) | |
425 | (msg, err, _win32_strerror(err))) |
|
491 | ||
|
492 | def _make_win_err(self, msg, err): | |||
|
493 | return IOError( | |||
|
494 | "%s win32 error code: %d %s" % (msg, err, _win32_strerror(err)) | |||
|
495 | ) | |||
426 |
|
496 | |||
427 | def close(self): |
|
497 | def close(self): | |
428 | if self.pipe: |
|
498 | if self.pipe: | |
429 |
log( |
|
499 | log("Closing pipe") | |
430 | CloseHandle(self.pipe) |
|
500 | CloseHandle(self.pipe) | |
431 | self.pipe = None |
|
501 | self.pipe = None | |
432 |
|
502 | |||
433 | if self._waitable is not None: |
|
503 | if self._waitable is not None: | |
434 | # We release the handle for the event |
|
504 | # We release the handle for the event | |
435 | CloseHandle(self._waitable) |
|
505 | CloseHandle(self._waitable) | |
436 | self._waitable = None |
|
506 | self._waitable = None | |
437 |
|
507 | |||
438 | def setTimeout(self, value): |
|
508 | def setTimeout(self, value): | |
439 | # convert to milliseconds |
|
509 | # convert to milliseconds | |
440 | self.timeout = int(value * 1000) |
|
510 | self.timeout = int(value * 1000) | |
441 |
|
511 | |||
442 | def readBytes(self, size): |
|
512 | def readBytes(self, size): | |
443 | """ A read can block for an unbounded amount of time, even if the |
|
513 | """ A read can block for an unbounded amount of time, even if the | |
444 | kernel reports that the pipe handle is signalled, so we need to |
|
514 | kernel reports that the pipe handle is signalled, so we need to | |
445 | always perform our reads asynchronously |
|
515 | always perform our reads asynchronously | |
446 | """ |
|
516 | """ | |
447 |
|
517 | |||
448 | # try to satisfy the read from any buffered data |
|
518 | # try to satisfy the read from any buffered data | |
449 | if self._iobuf: |
|
519 | if self._iobuf: | |
450 | if size >= len(self._iobuf): |
|
520 | if size >= len(self._iobuf): | |
451 | res = self._iobuf |
|
521 | res = self._iobuf | |
452 | self.buf = None |
|
522 | self.buf = None | |
453 | return res |
|
523 | return res | |
454 | res = self._iobuf[:size] |
|
524 | res = self._iobuf[:size] | |
455 | self._iobuf = self._iobuf[size:] |
|
525 | self._iobuf = self._iobuf[size:] | |
456 | return res |
|
526 | return res | |
457 |
|
527 | |||
458 | # We need to initiate a read |
|
528 | # We need to initiate a read | |
459 | buf = ctypes.create_string_buffer(size) |
|
529 | buf = ctypes.create_string_buffer(size) | |
460 | olap = OVERLAPPED() |
|
530 | olap = OVERLAPPED() | |
461 | olap.hEvent = self._waitable |
|
531 | olap.hEvent = self._waitable | |
462 |
|
532 | |||
463 |
log( |
|
533 | log("made read buff of size %d", size) | |
464 |
|
534 | |||
465 | # ReadFile docs warn against sending in the nread parameter for async |
|
535 | # ReadFile docs warn against sending in the nread parameter for async | |
466 | # operations, so we always collect it via GetOverlappedResultEx |
|
536 | # operations, so we always collect it via GetOverlappedResultEx | |
467 | immediate = ReadFile(self.pipe, buf, size, None, olap) |
|
537 | immediate = ReadFile(self.pipe, buf, size, None, olap) | |
468 |
|
538 | |||
469 | if not immediate: |
|
539 | if not immediate: | |
470 | err = GetLastError() |
|
540 | err = GetLastError() | |
471 | if err != ERROR_IO_PENDING: |
|
541 | if err != ERROR_IO_PENDING: | |
472 |
self._raise_win_err( |
|
542 | self._raise_win_err("failed to read %d bytes" % size, err) | |
473 | GetLastError()) |
|
|||
474 |
|
543 | |||
475 | nread = wintypes.DWORD() |
|
544 | nread = wintypes.DWORD() | |
476 |
if not self._get_overlapped_result_ex( |
|
545 | if not self._get_overlapped_result_ex( | |
477 |
|
|
546 | self.pipe, olap, nread, 0 if immediate else self.timeout, True | |
478 | True): |
|
547 | ): | |
479 | err = GetLastError() |
|
548 | err = GetLastError() | |
480 | CancelIoEx(self.pipe, olap) |
|
549 | CancelIoEx(self.pipe, olap) | |
481 |
|
550 | |||
482 | if err == WAIT_TIMEOUT: |
|
551 | if err == WAIT_TIMEOUT: | |
483 |
log( |
|
552 | log("GetOverlappedResultEx timedout") | |
484 |
raise SocketTimeout( |
|
553 | raise SocketTimeout( | |
485 |
|
|
554 | "timed out after waiting %dms for read" % self.timeout | |
|
555 | ) | |||
486 |
|
556 | |||
487 |
log( |
|
557 | log("GetOverlappedResultEx reports error %d", err) | |
488 |
self._raise_win_err( |
|
558 | self._raise_win_err("error while waiting for read", err) | |
489 |
|
559 | |||
490 | nread = nread.value |
|
560 | nread = nread.value | |
491 | if nread == 0: |
|
561 | if nread == 0: | |
492 | # Docs say that named pipes return 0 byte when the other end did |
|
562 | # Docs say that named pipes return 0 byte when the other end did | |
493 | # a zero byte write. Since we don't ever do that, the only |
|
563 | # a zero byte write. Since we don't ever do that, the only | |
494 | # other way this shows up is if the client has gotten in a weird |
|
564 | # other way this shows up is if the client has gotten in a weird | |
495 | # state, so let's bail out |
|
565 | # state, so let's bail out | |
496 | CancelIoEx(self.pipe, olap) |
|
566 | CancelIoEx(self.pipe, olap) | |
497 |
raise IOError( |
|
567 | raise IOError("Async read yielded 0 bytes; unpossible!") | |
498 |
|
568 | |||
499 | # Holds precisely the bytes that we read from the prior request |
|
569 | # Holds precisely the bytes that we read from the prior request | |
500 | buf = buf[:nread] |
|
570 | buf = buf[:nread] | |
501 |
|
571 | |||
502 | returned_size = min(nread, size) |
|
572 | returned_size = min(nread, size) | |
503 | if returned_size == nread: |
|
573 | if returned_size == nread: | |
504 | return buf |
|
574 | return buf | |
505 |
|
575 | |||
506 | # keep any left-overs around for a later read to consume |
|
576 | # keep any left-overs around for a later read to consume | |
507 | self._iobuf = buf[returned_size:] |
|
577 | self._iobuf = buf[returned_size:] | |
508 | return buf[:returned_size] |
|
578 | return buf[:returned_size] | |
509 |
|
579 | |||
510 | def write(self, data): |
|
580 | def write(self, data): | |
511 | olap = OVERLAPPED() |
|
581 | olap = OVERLAPPED() | |
512 | olap.hEvent = self._waitable |
|
582 | olap.hEvent = self._waitable | |
513 |
|
583 | |||
514 | immediate = WriteFile(self.pipe, ctypes.c_char_p(data), len(data), |
|
584 | immediate = WriteFile( | |
515 | None, olap) |
|
585 | self.pipe, ctypes.c_char_p(data), len(data), None, olap | |
|
586 | ) | |||
516 |
|
587 | |||
517 | if not immediate: |
|
588 | if not immediate: | |
518 | err = GetLastError() |
|
589 | err = GetLastError() | |
519 | if err != ERROR_IO_PENDING: |
|
590 | if err != ERROR_IO_PENDING: | |
520 |
self._raise_win_err( |
|
591 | self._raise_win_err( | |
521 | GetLastError()) |
|
592 | "failed to write %d bytes to handle %r" | |
|
593 | % (len(data), self.pipe), | |||
|
594 | err, | |||
|
595 | ) | |||
522 |
|
596 | |||
523 | # Obtain results, waiting if needed |
|
597 | # Obtain results, waiting if needed | |
524 | nwrote = wintypes.DWORD() |
|
598 | nwrote = wintypes.DWORD() | |
525 |
if self._get_overlapped_result_ex( |
|
599 | if self._get_overlapped_result_ex( | |
526 |
|
|
600 | self.pipe, olap, nwrote, 0 if immediate else self.timeout, True | |
527 | True): |
|
601 | ): | |
528 |
log( |
|
602 | log("made write of %d bytes", nwrote.value) | |
529 | return nwrote.value |
|
603 | return nwrote.value | |
530 |
|
604 | |||
531 | err = GetLastError() |
|
605 | err = GetLastError() | |
532 |
|
606 | |||
533 | # It's potentially unsafe to allow the write to continue after |
|
607 | # It's potentially unsafe to allow the write to continue after | |
534 | # we unwind, so let's make a best effort to avoid that happening |
|
608 | # we unwind, so let's make a best effort to avoid that happening | |
535 | CancelIoEx(self.pipe, olap) |
|
609 | CancelIoEx(self.pipe, olap) | |
536 |
|
610 | |||
537 | if err == WAIT_TIMEOUT: |
|
611 | if err == WAIT_TIMEOUT: | |
538 |
raise SocketTimeout( |
|
612 | raise SocketTimeout( | |
539 |
|
|
613 | "timed out after waiting %dms for write" % self.timeout | |
540 | self._raise_win_err('error while waiting for write of %d bytes' % |
|
614 | ) | |
541 | len(data), err) |
|
615 | self._raise_win_err( | |
|
616 | "error while waiting for write of %d bytes" % len(data), err | |||
|
617 | ) | |||
|
618 | ||||
|
619 | ||||
|
620 | def _default_binpath(binpath=None): | |||
|
621 | if binpath: | |||
|
622 | return binpath | |||
|
623 | # The test harness sets WATCHMAN_BINARY to the binary under test, | |||
|
624 | # so we use that by default, otherwise, allow resolving watchman | |||
|
625 | # from the users PATH. | |||
|
626 | return os.environ.get("WATCHMAN_BINARY", "watchman") | |||
542 |
|
627 | |||
543 |
|
628 | |||
544 | class CLIProcessTransport(Transport): |
|
629 | class CLIProcessTransport(Transport): | |
545 | """ open a pipe to the cli to talk to the service |
|
630 | """ open a pipe to the cli to talk to the service | |
546 | This intended to be used only in the test harness! |
|
631 | This intended to be used only in the test harness! | |
547 |
|
632 | |||
548 | The CLI is an oddball because we only support JSON input |
|
633 | The CLI is an oddball because we only support JSON input | |
549 | and cannot send multiple commands through the same instance, |
|
634 | and cannot send multiple commands through the same instance, | |
550 | so we spawn a new process for each command. |
|
635 | so we spawn a new process for each command. | |
551 |
|
636 | |||
552 | We disable server spawning for this implementation, again, because |
|
637 | We disable server spawning for this implementation, again, because | |
553 | it is intended to be used only in our test harness. You really |
|
638 | it is intended to be used only in our test harness. You really | |
554 | should not need to use the CLI transport for anything real. |
|
639 | should not need to use the CLI transport for anything real. | |
555 |
|
640 | |||
556 | While the CLI can output in BSER, our Transport interface doesn't |
|
641 | While the CLI can output in BSER, our Transport interface doesn't | |
557 | support telling this instance that it should do so. That effectively |
|
642 | support telling this instance that it should do so. That effectively | |
558 | limits this implementation to JSON input and output only at this time. |
|
643 | limits this implementation to JSON input and output only at this time. | |
559 |
|
644 | |||
560 | It is the responsibility of the caller to set the send and |
|
645 | It is the responsibility of the caller to set the send and | |
561 | receive codecs appropriately. |
|
646 | receive codecs appropriately. | |
562 | """ |
|
647 | """ | |
|
648 | ||||
563 | proc = None |
|
649 | proc = None | |
564 | closed = True |
|
650 | closed = True | |
565 |
|
651 | |||
566 |
def __init__(self, sockpath, timeout, |
|
652 | def __init__(self, sockpath, timeout, binpath=None): | |
567 | self.sockpath = sockpath |
|
653 | self.sockpath = sockpath | |
568 | self.timeout = timeout |
|
654 | self.timeout = timeout | |
569 | self.watchman_exe = watchman_exe |
|
655 | self.binpath = _default_binpath(binpath) | |
570 |
|
656 | |||
571 | def close(self): |
|
657 | def close(self): | |
572 | if self.proc: |
|
658 | if self.proc: | |
573 | if self.proc.pid is not None: |
|
659 | if self.proc.pid is not None: | |
574 | self.proc.kill() |
|
660 | self.proc.kill() | |
575 | self.proc.stdin.close() |
|
661 | self.proc.stdin.close() | |
576 | self.proc.stdout.close() |
|
662 | self.proc.stdout.close() | |
|
663 | self.proc.wait() | |||
577 | self.proc = None |
|
664 | self.proc = None | |
578 |
|
665 | |||
579 | def _connect(self): |
|
666 | def _connect(self): | |
580 | if self.proc: |
|
667 | if self.proc: | |
581 | return self.proc |
|
668 | return self.proc | |
582 | args = [ |
|
669 | args = [ | |
583 |
self. |
|
670 | self.binpath, | |
584 |
|
|
671 | "--sockname={0}".format(self.sockpath), | |
585 |
|
|
672 | "--logfile=/BOGUS", | |
586 |
|
|
673 | "--statefile=/BOGUS", | |
587 |
|
|
674 | "--no-spawn", | |
588 |
|
|
675 | "--no-local", | |
589 |
|
|
676 | "--no-pretty", | |
590 |
|
|
677 | "-j", | |
591 | ] |
|
678 | ] | |
592 |
self.proc = subprocess.Popen( |
|
679 | self.proc = subprocess.Popen( | |
593 | args), |
|
680 | args, stdin=subprocess.PIPE, stdout=subprocess.PIPE | |
594 | stdin=subprocess.PIPE, |
|
681 | ) | |
595 | stdout=subprocess.PIPE) |
|
|||
596 | return self.proc |
|
682 | return self.proc | |
597 |
|
683 | |||
598 | def readBytes(self, size): |
|
684 | def readBytes(self, size): | |
599 | self._connect() |
|
685 | self._connect() | |
600 | res = self.proc.stdout.read(size) |
|
686 | res = self.proc.stdout.read(size) | |
601 |
if res |
|
687 | if not res: | |
602 |
raise WatchmanError( |
|
688 | raise WatchmanError("EOF on CLI process transport") | |
603 | return res |
|
689 | return res | |
604 |
|
690 | |||
605 | def write(self, data): |
|
691 | def write(self, data): | |
606 | if self.closed: |
|
692 | if self.closed: | |
607 | self.close() |
|
693 | self.close() | |
608 | self.closed = False |
|
694 | self.closed = False | |
609 | self._connect() |
|
695 | self._connect() | |
610 | res = self.proc.stdin.write(data) |
|
696 | res = self.proc.stdin.write(data) | |
611 | self.proc.stdin.close() |
|
697 | self.proc.stdin.close() | |
612 | self.closed = True |
|
698 | self.closed = True | |
613 | return res |
|
699 | return res | |
614 |
|
700 | |||
615 |
|
701 | |||
616 | class BserCodec(Codec): |
|
702 | class BserCodec(Codec): | |
617 | """ use the BSER encoding. This is the default, preferred codec """ |
|
703 | """ use the BSER encoding. This is the default, preferred codec """ | |
618 |
|
704 | |||
|
705 | def __init__(self, transport, value_encoding, value_errors): | |||
|
706 | super(BserCodec, self).__init__(transport) | |||
|
707 | self._value_encoding = value_encoding | |||
|
708 | self._value_errors = value_errors | |||
|
709 | ||||
619 | def _loads(self, response): |
|
710 | def _loads(self, response): | |
620 |
return bser.loads( |
|
711 | return bser.loads( | |
|
712 | response, | |||
|
713 | value_encoding=self._value_encoding, | |||
|
714 | value_errors=self._value_errors, | |||
|
715 | ) | |||
621 |
|
716 | |||
622 | def receive(self): |
|
717 | def receive(self): | |
623 | buf = [self.transport.readBytes(sniff_len)] |
|
718 | buf = [self.transport.readBytes(sniff_len)] | |
624 | if not buf[0]: |
|
719 | if not buf[0]: | |
625 |
raise WatchmanError( |
|
720 | raise WatchmanError("empty watchman response") | |
626 |
|
721 | |||
627 | _1, _2, elen = bser.pdu_info(buf[0]) |
|
722 | _1, _2, elen = bser.pdu_info(buf[0]) | |
628 |
|
723 | |||
629 | rlen = len(buf[0]) |
|
724 | rlen = len(buf[0]) | |
630 | while elen > rlen: |
|
725 | while elen > rlen: | |
631 | buf.append(self.transport.readBytes(elen - rlen)) |
|
726 | buf.append(self.transport.readBytes(elen - rlen)) | |
632 | rlen += len(buf[-1]) |
|
727 | rlen += len(buf[-1]) | |
633 |
|
728 | |||
634 |
response = b |
|
729 | response = b"".join(buf) | |
635 | try: |
|
730 | try: | |
636 | res = self._loads(response) |
|
731 | res = self._loads(response) | |
637 | return res |
|
732 | return res | |
638 | except ValueError as e: |
|
733 | except ValueError as e: | |
639 |
raise WatchmanError( |
|
734 | raise WatchmanError("watchman response decode error: %s" % e) | |
640 |
|
735 | |||
641 | def send(self, *args): |
|
736 | def send(self, *args): | |
642 | cmd = bser.dumps(*args) # Defaults to BSER v1 |
|
737 | cmd = bser.dumps(*args) # Defaults to BSER v1 | |
643 | self.transport.write(cmd) |
|
738 | self.transport.write(cmd) | |
644 |
|
739 | |||
645 |
|
740 | |||
646 | class ImmutableBserCodec(BserCodec): |
|
741 | class ImmutableBserCodec(BserCodec): | |
647 | """ use the BSER encoding, decoding values using the newer |
|
742 | """ use the BSER encoding, decoding values using the newer | |
648 | immutable object support """ |
|
743 | immutable object support """ | |
649 |
|
744 | |||
650 | def _loads(self, response): |
|
745 | def _loads(self, response): | |
651 | return bser.loads(response, False) # Defaults to BSER v1 |
|
746 | return bser.loads( | |
|
747 | response, | |||
|
748 | False, | |||
|
749 | value_encoding=self._value_encoding, | |||
|
750 | value_errors=self._value_errors, | |||
|
751 | ) | |||
652 |
|
752 | |||
653 |
|
753 | |||
654 | class Bser2WithFallbackCodec(BserCodec): |
|
754 | class Bser2WithFallbackCodec(BserCodec): | |
655 | """ use BSER v2 encoding """ |
|
755 | """ use BSER v2 encoding """ | |
656 |
|
756 | |||
657 | def __init__(self, transport): |
|
757 | def __init__(self, transport, value_encoding, value_errors): | |
658 |
super(Bser2WithFallbackCodec, self).__init__( |
|
758 | super(Bser2WithFallbackCodec, self).__init__( | |
659 | # Once the server advertises support for bser-v2 we should switch this |
|
759 | transport, value_encoding, value_errors | |
660 | # to 'required' on Python 3. |
|
760 | ) | |
661 | self.send(["version", {"optional": ["bser-v2"]}]) |
|
761 | if compat.PYTHON3: | |
|
762 | bserv2_key = "required" | |||
|
763 | else: | |||
|
764 | bserv2_key = "optional" | |||
|
765 | ||||
|
766 | self.send(["version", {bserv2_key: ["bser-v2"]}]) | |||
662 |
|
767 | |||
663 | capabilities = self.receive() |
|
768 | capabilities = self.receive() | |
664 |
|
769 | |||
665 |
if |
|
770 | if "error" in capabilities: | |
666 |
raise |
|
771 | raise BSERv1Unsupported( | |
|
772 | "The watchman server version does not support Python 3. Please " | |||
|
773 | "upgrade your watchman server." | |||
|
774 | ) | |||
667 |
|
775 | |||
668 |
if capabilities[ |
|
776 | if capabilities["capabilities"]["bser-v2"]: | |
669 | self.bser_version = 2 |
|
777 | self.bser_version = 2 | |
670 | self.bser_capabilities = 0 |
|
778 | self.bser_capabilities = 0 | |
671 | else: |
|
779 | else: | |
672 | self.bser_version = 1 |
|
780 | self.bser_version = 1 | |
673 | self.bser_capabilities = 0 |
|
781 | self.bser_capabilities = 0 | |
674 |
|
782 | |||
675 | def _loads(self, response): |
|
|||
676 | return bser.loads(response) |
|
|||
677 |
|
||||
678 | def receive(self): |
|
783 | def receive(self): | |
679 | buf = [self.transport.readBytes(sniff_len)] |
|
784 | buf = [self.transport.readBytes(sniff_len)] | |
680 | if not buf[0]: |
|
785 | if not buf[0]: | |
681 |
raise WatchmanError( |
|
786 | raise WatchmanError("empty watchman response") | |
682 |
|
787 | |||
683 | recv_bser_version, recv_bser_capabilities, elen = bser.pdu_info(buf[0]) |
|
788 | recv_bser_version, recv_bser_capabilities, elen = bser.pdu_info(buf[0]) | |
684 |
|
789 | |||
685 |
if hasattr(self, |
|
790 | if hasattr(self, "bser_version"): | |
686 | # Readjust BSER version and capabilities if necessary |
|
791 | # Readjust BSER version and capabilities if necessary | |
687 | self.bser_version = max(self.bser_version, recv_bser_version) |
|
792 | self.bser_version = max(self.bser_version, recv_bser_version) | |
688 | self.capabilities = self.bser_capabilities & recv_bser_capabilities |
|
793 | self.capabilities = self.bser_capabilities & recv_bser_capabilities | |
689 |
|
794 | |||
690 | rlen = len(buf[0]) |
|
795 | rlen = len(buf[0]) | |
691 | while elen > rlen: |
|
796 | while elen > rlen: | |
692 | buf.append(self.transport.readBytes(elen - rlen)) |
|
797 | buf.append(self.transport.readBytes(elen - rlen)) | |
693 | rlen += len(buf[-1]) |
|
798 | rlen += len(buf[-1]) | |
694 |
|
799 | |||
695 |
response = b |
|
800 | response = b"".join(buf) | |
696 | try: |
|
801 | try: | |
697 | res = self._loads(response) |
|
802 | res = self._loads(response) | |
698 | return res |
|
803 | return res | |
699 | except ValueError as e: |
|
804 | except ValueError as e: | |
700 |
raise WatchmanError( |
|
805 | raise WatchmanError("watchman response decode error: %s" % e) | |
701 |
|
806 | |||
702 | def send(self, *args): |
|
807 | def send(self, *args): | |
703 |
if hasattr(self, |
|
808 | if hasattr(self, "bser_version"): | |
704 |
cmd = bser.dumps( |
|
809 | cmd = bser.dumps( | |
705 | capabilities=self.bser_capabilities) |
|
810 | *args, | |
|
811 | version=self.bser_version, | |||
|
812 | capabilities=self.bser_capabilities | |||
|
813 | ) | |||
706 | else: |
|
814 | else: | |
707 | cmd = bser.dumps(*args) |
|
815 | cmd = bser.dumps(*args) | |
708 | self.transport.write(cmd) |
|
816 | self.transport.write(cmd) | |
709 |
|
817 | |||
710 |
|
818 | |||
|
819 | class ImmutableBser2Codec(Bser2WithFallbackCodec, ImmutableBserCodec): | |||
|
820 | """ use the BSER encoding, decoding values using the newer | |||
|
821 | immutable object support """ | |||
|
822 | ||||
|
823 | pass | |||
|
824 | ||||
|
825 | ||||
711 | class JsonCodec(Codec): |
|
826 | class JsonCodec(Codec): | |
712 | """ Use json codec. This is here primarily for testing purposes """ |
|
827 | """ Use json codec. This is here primarily for testing purposes """ | |
|
828 | ||||
713 | json = None |
|
829 | json = None | |
714 |
|
830 | |||
715 | def __init__(self, transport): |
|
831 | def __init__(self, transport): | |
716 | super(JsonCodec, self).__init__(transport) |
|
832 | super(JsonCodec, self).__init__(transport) | |
717 | # optional dep on json, only if JsonCodec is used |
|
833 | # optional dep on json, only if JsonCodec is used | |
718 | import json |
|
834 | import json | |
|
835 | ||||
719 | self.json = json |
|
836 | self.json = json | |
720 |
|
837 | |||
721 | def receive(self): |
|
838 | def receive(self): | |
722 | line = self.transport.readLine() |
|
839 | line = self.transport.readLine() | |
723 | try: |
|
840 | try: | |
724 | # In Python 3, json.loads is a transformation from Unicode string to |
|
841 | # In Python 3, json.loads is a transformation from Unicode string to | |
725 | # objects possibly containing Unicode strings. We typically expect |
|
842 | # objects possibly containing Unicode strings. We typically expect | |
726 | # the JSON blob to be ASCII-only with non-ASCII characters escaped, |
|
843 | # the JSON blob to be ASCII-only with non-ASCII characters escaped, | |
727 | # but it's possible we might get non-ASCII bytes that are valid |
|
844 | # but it's possible we might get non-ASCII bytes that are valid | |
728 | # UTF-8. |
|
845 | # UTF-8. | |
729 | if compat.PYTHON3: |
|
846 | if compat.PYTHON3: | |
730 |
line = line.decode( |
|
847 | line = line.decode("utf-8") | |
731 | return self.json.loads(line) |
|
848 | return self.json.loads(line) | |
732 | except Exception as e: |
|
849 | except Exception as e: | |
733 | print(e, line) |
|
850 | print(e, line) | |
734 | raise |
|
851 | raise | |
735 |
|
852 | |||
736 | def send(self, *args): |
|
853 | def send(self, *args): | |
737 | cmd = self.json.dumps(*args) |
|
854 | cmd = self.json.dumps(*args) | |
738 | # In Python 3, json.dumps is a transformation from objects possibly |
|
855 | # In Python 3, json.dumps is a transformation from objects possibly | |
739 | # containing Unicode strings to Unicode string. Even with (the default) |
|
856 | # containing Unicode strings to Unicode string. Even with (the default) | |
740 | # ensure_ascii=True, dumps returns a Unicode string. |
|
857 | # ensure_ascii=True, dumps returns a Unicode string. | |
741 | if compat.PYTHON3: |
|
858 | if compat.PYTHON3: | |
742 |
cmd = cmd.encode( |
|
859 | cmd = cmd.encode("ascii") | |
743 | self.transport.write(cmd + b"\n") |
|
860 | self.transport.write(cmd + b"\n") | |
744 |
|
861 | |||
745 |
|
862 | |||
746 | class client(object): |
|
863 | class client(object): | |
747 | """ Handles the communication with the watchman service """ |
|
864 | """ Handles the communication with the watchman service """ | |
|
865 | ||||
748 | sockpath = None |
|
866 | sockpath = None | |
749 | transport = None |
|
867 | transport = None | |
750 | sendCodec = None |
|
868 | sendCodec = None | |
751 | recvCodec = None |
|
869 | recvCodec = None | |
752 | sendConn = None |
|
870 | sendConn = None | |
753 | recvConn = None |
|
871 | recvConn = None | |
754 | subs = {} # Keyed by subscription name |
|
872 | subs = {} # Keyed by subscription name | |
755 | sub_by_root = {} # Keyed by root, then by subscription name |
|
873 | sub_by_root = {} # Keyed by root, then by subscription name | |
756 | logs = [] # When log level is raised |
|
874 | logs = [] # When log level is raised | |
757 |
unilateral = [ |
|
875 | unilateral = ["log", "subscription"] | |
758 | tport = None |
|
876 | tport = None | |
759 | useImmutableBser = None |
|
877 | useImmutableBser = None | |
760 | watchman_exe = None |
|
878 | pid = None | |
761 |
|
879 | |||
762 |
def __init__( |
|
880 | def __init__( | |
763 | sockpath=None, |
|
881 | self, | |
764 | timeout=1.0, |
|
882 | sockpath=None, | |
765 | transport=None, |
|
883 | timeout=1.0, | |
766 | sendEncoding=None, |
|
884 | transport=None, | |
767 |
|
|
885 | sendEncoding=None, | |
768 | useImmutableBser=False, |
|
886 | recvEncoding=None, | |
769 | watchman_exe=None): |
|
887 | useImmutableBser=False, | |
|
888 | # use False for these two because None has a special | |||
|
889 | # meaning | |||
|
890 | valueEncoding=False, | |||
|
891 | valueErrors=False, | |||
|
892 | binpath=None, | |||
|
893 | ): | |||
770 | self.sockpath = sockpath |
|
894 | self.sockpath = sockpath | |
771 | self.timeout = timeout |
|
895 | self.timeout = timeout | |
772 | self.useImmutableBser = useImmutableBser |
|
896 | self.useImmutableBser = useImmutableBser | |
773 | self.watchman_exe = watchman_exe |
|
897 | self.binpath = _default_binpath(binpath) | |
774 |
|
898 | |||
775 | if inspect.isclass(transport) and issubclass(transport, Transport): |
|
899 | if inspect.isclass(transport) and issubclass(transport, Transport): | |
776 | self.transport = transport |
|
900 | self.transport = transport | |
777 | else: |
|
901 | else: | |
778 |
transport = transport or os.getenv( |
|
902 | transport = transport or os.getenv("WATCHMAN_TRANSPORT") or "local" | |
779 |
if transport == |
|
903 | if transport == "local" and os.name == "nt": | |
780 | self.transport = WindowsNamedPipeTransport |
|
904 | self.transport = WindowsNamedPipeTransport | |
781 |
elif transport == |
|
905 | elif transport == "local": | |
782 | self.transport = UnixSocketTransport |
|
906 | self.transport = UnixSocketTransport | |
783 |
elif transport == |
|
907 | elif transport == "cli": | |
784 | self.transport = CLIProcessTransport |
|
908 | self.transport = CLIProcessTransport | |
785 | if sendEncoding is None: |
|
909 | if sendEncoding is None: | |
786 |
sendEncoding = |
|
910 | sendEncoding = "json" | |
787 | if recvEncoding is None: |
|
911 | if recvEncoding is None: | |
788 | recvEncoding = sendEncoding |
|
912 | recvEncoding = sendEncoding | |
789 | else: |
|
913 | else: | |
790 |
raise WatchmanError( |
|
914 | raise WatchmanError("invalid transport %s" % transport) | |
791 |
|
915 | |||
792 | sendEncoding = str(sendEncoding or os.getenv('WATCHMAN_ENCODING') or |
|
916 | sendEncoding = str( | |
793 | 'bser') |
|
917 | sendEncoding or os.getenv("WATCHMAN_ENCODING") or "bser" | |
794 | recvEncoding = str(recvEncoding or os.getenv('WATCHMAN_ENCODING') or |
|
918 | ) | |
795 | 'bser') |
|
919 | recvEncoding = str( | |
|
920 | recvEncoding or os.getenv("WATCHMAN_ENCODING") or "bser" | |||
|
921 | ) | |||
796 |
|
922 | |||
797 | self.recvCodec = self._parseEncoding(recvEncoding) |
|
923 | self.recvCodec = self._parseEncoding(recvEncoding) | |
798 | self.sendCodec = self._parseEncoding(sendEncoding) |
|
924 | self.sendCodec = self._parseEncoding(sendEncoding) | |
799 |
|
925 | |||
|
926 | # We want to act like the native OS methods as much as possible. This | |||
|
927 | # means returning bytestrings on Python 2 by default and Unicode | |||
|
928 | # strings on Python 3. However we take an optional argument that lets | |||
|
929 | # users override this. | |||
|
930 | if valueEncoding is False: | |||
|
931 | if compat.PYTHON3: | |||
|
932 | self.valueEncoding = encoding.get_local_encoding() | |||
|
933 | self.valueErrors = encoding.default_local_errors | |||
|
934 | else: | |||
|
935 | self.valueEncoding = None | |||
|
936 | self.valueErrors = None | |||
|
937 | else: | |||
|
938 | self.valueEncoding = valueEncoding | |||
|
939 | if valueErrors is False: | |||
|
940 | self.valueErrors = encoding.default_local_errors | |||
|
941 | else: | |||
|
942 | self.valueErrors = valueErrors | |||
|
943 | ||||
|
944 | def _makeBSERCodec(self, codec): | |||
|
945 | def make_codec(transport): | |||
|
946 | return codec(transport, self.valueEncoding, self.valueErrors) | |||
|
947 | ||||
|
948 | return make_codec | |||
|
949 | ||||
800 | def _parseEncoding(self, enc): |
|
950 | def _parseEncoding(self, enc): | |
801 |
if enc == |
|
951 | if enc == "bser": | |
802 | if self.useImmutableBser: |
|
952 | if self.useImmutableBser: | |
803 | return ImmutableBserCodec |
|
953 | return self._makeBSERCodec(ImmutableBser2Codec) | |
804 | return BserCodec |
|
954 | return self._makeBSERCodec(Bser2WithFallbackCodec) | |
805 |
elif enc == |
|
955 | elif enc == "bser-v1": | |
806 | return Bser2WithFallbackCodec |
|
956 | if compat.PYTHON3: | |
807 | elif enc == 'json': |
|
957 | raise BSERv1Unsupported( | |
|
958 | "Python 3 does not support the BSER v1 encoding: specify " | |||
|
959 | '"bser" or omit the sendEncoding and recvEncoding ' | |||
|
960 | "arguments" | |||
|
961 | ) | |||
|
962 | if self.useImmutableBser: | |||
|
963 | return self._makeBSERCodec(ImmutableBserCodec) | |||
|
964 | return self._makeBSERCodec(BserCodec) | |||
|
965 | elif enc == "json": | |||
808 | return JsonCodec |
|
966 | return JsonCodec | |
809 | else: |
|
967 | else: | |
810 |
raise WatchmanError( |
|
968 | raise WatchmanError("invalid encoding %s" % enc) | |
811 |
|
969 | |||
812 | def _hasprop(self, result, name): |
|
970 | def _hasprop(self, result, name): | |
813 | if self.useImmutableBser: |
|
971 | if self.useImmutableBser: | |
814 | return hasattr(result, name) |
|
972 | return hasattr(result, name) | |
815 | return name in result |
|
973 | return name in result | |
816 |
|
974 | |||
817 | def _resolvesockname(self): |
|
975 | def _resolvesockname(self): | |
818 | # if invoked via a trigger, watchman will set this env var; we |
|
976 | # if invoked via a trigger, watchman will set this env var; we | |
819 | # should use it unless explicitly set otherwise |
|
977 | # should use it unless explicitly set otherwise | |
820 |
path = os.getenv( |
|
978 | path = os.getenv("WATCHMAN_SOCK") | |
821 | if path: |
|
979 | if path: | |
822 | return path |
|
980 | return path | |
823 |
|
981 | |||
824 |
cmd = [self. |
|
982 | cmd = [self.binpath, "--output-encoding=bser", "get-sockname"] | |
825 | try: |
|
983 | try: | |
826 |
args = dict( |
|
984 | args = dict( | |
827 |
|
|
985 | stdout=subprocess.PIPE, stderr=subprocess.PIPE | |
828 | close_fds=os.name != 'nt') |
|
986 | ) # noqa: C408 | |
829 |
|
987 | |||
830 |
if os.name == |
|
988 | if os.name == "nt": | |
831 | # if invoked via an application with graphical user interface, |
|
989 | # if invoked via an application with graphical user interface, | |
832 | # this call will cause a brief command window pop-up. |
|
990 | # this call will cause a brief command window pop-up. | |
833 | # Using the flag STARTF_USESHOWWINDOW to avoid this behavior. |
|
991 | # Using the flag STARTF_USESHOWWINDOW to avoid this behavior. | |
834 | startupinfo = subprocess.STARTUPINFO() |
|
992 | startupinfo = subprocess.STARTUPINFO() | |
835 | startupinfo.dwFlags |= subprocess.STARTF_USESHOWWINDOW |
|
993 | startupinfo.dwFlags |= subprocess.STARTF_USESHOWWINDOW | |
836 |
args[ |
|
994 | args["startupinfo"] = startupinfo | |
837 |
|
995 | |||
838 |
p = subprocess.Popen( |
|
996 | p = subprocess.Popen(cmd, **args) | |
839 | **args) |
|
|||
840 |
|
997 | |||
841 | except OSError as e: |
|
998 | except OSError as e: | |
842 |
raise WatchmanError('"watchman" executable not in PATH (%s)' |
|
999 | raise WatchmanError('"watchman" executable not in PATH (%s)', e) | |
843 |
|
1000 | |||
844 | stdout, stderr = p.communicate() |
|
1001 | stdout, stderr = p.communicate() | |
845 | exitcode = p.poll() |
|
1002 | exitcode = p.poll() | |
846 |
|
1003 | |||
847 | if exitcode: |
|
1004 | if exitcode: | |
848 | raise WatchmanError("watchman exited with code %d" % exitcode) |
|
1005 | raise WatchmanError("watchman exited with code %d" % exitcode) | |
849 |
|
1006 | |||
850 | result = bser.loads(stdout) |
|
1007 | result = bser.loads(stdout) | |
851 |
if |
|
1008 | if "error" in result: | |
852 |
raise WatchmanError( |
|
1009 | raise WatchmanError("get-sockname error: %s" % result["error"]) | |
853 |
|
1010 | |||
854 |
return result[ |
|
1011 | return result["sockname"] | |
855 |
|
1012 | |||
856 | def _connect(self): |
|
1013 | def _connect(self): | |
857 | """ establish transport connection """ |
|
1014 | """ establish transport connection """ | |
858 |
|
1015 | |||
859 | if self.recvConn: |
|
1016 | if self.recvConn: | |
|
1017 | if self.pid != os.getpid(): | |||
|
1018 | raise UseAfterFork( | |||
|
1019 | "do not re-use a connection after fork; open a new client instead" | |||
|
1020 | ) | |||
860 | return |
|
1021 | return | |
861 |
|
1022 | |||
862 | if self.sockpath is None: |
|
1023 | if self.sockpath is None: | |
863 | self.sockpath = self._resolvesockname() |
|
1024 | self.sockpath = self._resolvesockname() | |
864 |
|
1025 | |||
865 | self.tport = self.transport(self.sockpath, self.timeout, self.watchman_exe) |
|
1026 | kwargs = {} | |
|
1027 | if self.transport == CLIProcessTransport: | |||
|
1028 | kwargs["binpath"] = self.binpath | |||
|
1029 | ||||
|
1030 | self.tport = self.transport(self.sockpath, self.timeout, **kwargs) | |||
866 | self.sendConn = self.sendCodec(self.tport) |
|
1031 | self.sendConn = self.sendCodec(self.tport) | |
867 | self.recvConn = self.recvCodec(self.tport) |
|
1032 | self.recvConn = self.recvCodec(self.tport) | |
|
1033 | self.pid = os.getpid() | |||
868 |
|
1034 | |||
869 | def __del__(self): |
|
1035 | def __del__(self): | |
870 | self.close() |
|
1036 | self.close() | |
871 |
|
1037 | |||
|
1038 | def __enter__(self): | |||
|
1039 | self._connect() | |||
|
1040 | return self | |||
|
1041 | ||||
|
1042 | def __exit__(self, exc_type, exc_value, exc_traceback): | |||
|
1043 | self.close() | |||
|
1044 | ||||
872 | def close(self): |
|
1045 | def close(self): | |
873 | if self.tport: |
|
1046 | if self.tport: | |
874 | self.tport.close() |
|
1047 | self.tport.close() | |
875 | self.tport = None |
|
1048 | self.tport = None | |
876 | self.recvConn = None |
|
1049 | self.recvConn = None | |
877 | self.sendConn = None |
|
1050 | self.sendConn = None | |
878 |
|
1051 | |||
879 | def receive(self): |
|
1052 | def receive(self): | |
880 | """ receive the next PDU from the watchman service |
|
1053 | """ receive the next PDU from the watchman service | |
881 |
|
1054 | |||
882 | If the client has activated subscriptions or logs then |
|
1055 | If the client has activated subscriptions or logs then | |
883 | this PDU may be a unilateral PDU sent by the service to |
|
1056 | this PDU may be a unilateral PDU sent by the service to | |
884 | inform the client of a log event or subscription change. |
|
1057 | inform the client of a log event or subscription change. | |
885 |
|
1058 | |||
886 | It may also simply be the response portion of a request |
|
1059 | It may also simply be the response portion of a request | |
887 | initiated by query. |
|
1060 | initiated by query. | |
888 |
|
1061 | |||
889 | There are clients in production that subscribe and call |
|
1062 | There are clients in production that subscribe and call | |
890 | this in a loop to retrieve all subscription responses, |
|
1063 | this in a loop to retrieve all subscription responses, | |
891 | so care should be taken when making changes here. |
|
1064 | so care should be taken when making changes here. | |
892 | """ |
|
1065 | """ | |
893 |
|
1066 | |||
894 | self._connect() |
|
1067 | self._connect() | |
895 | result = self.recvConn.receive() |
|
1068 | result = self.recvConn.receive() | |
896 |
if self._hasprop(result, |
|
1069 | if self._hasprop(result, "error"): | |
897 |
|
|
1070 | raise CommandError(result["error"]) | |
898 | if compat.PYTHON3 and isinstance(self.recvConn, BserCodec): |
|
|||
899 | error = result['error'].decode('utf-8', 'surrogateescape') |
|
|||
900 | raise CommandError(error) |
|
|||
901 |
|
1071 | |||
902 |
if self._hasprop(result, |
|
1072 | if self._hasprop(result, "log"): | |
903 |
|
|
1073 | self.logs.append(result["log"]) | |
904 | if compat.PYTHON3 and isinstance(self.recvConn, BserCodec): |
|
|||
905 | log = log.decode('utf-8', 'surrogateescape') |
|
|||
906 | self.logs.append(log) |
|
|||
907 |
|
1074 | |||
908 |
if self._hasprop(result, |
|
1075 | if self._hasprop(result, "subscription"): | |
909 |
sub = result[ |
|
1076 | sub = result["subscription"] | |
910 | if not (sub in self.subs): |
|
1077 | if not (sub in self.subs): | |
911 | self.subs[sub] = [] |
|
1078 | self.subs[sub] = [] | |
912 | self.subs[sub].append(result) |
|
1079 | self.subs[sub].append(result) | |
913 |
|
1080 | |||
914 | # also accumulate in {root,sub} keyed store |
|
1081 | # also accumulate in {root,sub} keyed store | |
915 |
root = os.path.normcase(result[ |
|
1082 | root = os.path.normpath(os.path.normcase(result["root"])) | |
916 | if not root in self.sub_by_root: |
|
1083 | if not root in self.sub_by_root: | |
917 | self.sub_by_root[root] = {} |
|
1084 | self.sub_by_root[root] = {} | |
918 | if not sub in self.sub_by_root[root]: |
|
1085 | if not sub in self.sub_by_root[root]: | |
919 | self.sub_by_root[root][sub] = [] |
|
1086 | self.sub_by_root[root][sub] = [] | |
920 | self.sub_by_root[root][sub].append(result) |
|
1087 | self.sub_by_root[root][sub].append(result) | |
921 |
|
1088 | |||
922 | return result |
|
1089 | return result | |
923 |
|
1090 | |||
924 | def isUnilateralResponse(self, res): |
|
1091 | def isUnilateralResponse(self, res): | |
925 |
if |
|
1092 | if "unilateral" in res and res["unilateral"]: | |
926 | return True |
|
1093 | return True | |
927 | # Fall back to checking for known unilateral responses |
|
1094 | # Fall back to checking for known unilateral responses | |
928 | for k in self.unilateral: |
|
1095 | for k in self.unilateral: | |
929 | if k in res: |
|
1096 | if k in res: | |
930 | return True |
|
1097 | return True | |
931 | return False |
|
1098 | return False | |
932 |
|
1099 | |||
933 | def getLog(self, remove=True): |
|
1100 | def getLog(self, remove=True): | |
934 | """ Retrieve buffered log data |
|
1101 | """ Retrieve buffered log data | |
935 |
|
1102 | |||
936 | If remove is true the data will be removed from the buffer. |
|
1103 | If remove is true the data will be removed from the buffer. | |
937 | Otherwise it will be left in the buffer |
|
1104 | Otherwise it will be left in the buffer | |
938 | """ |
|
1105 | """ | |
939 | res = self.logs |
|
1106 | res = self.logs | |
940 | if remove: |
|
1107 | if remove: | |
941 | self.logs = [] |
|
1108 | self.logs = [] | |
942 | return res |
|
1109 | return res | |
943 |
|
1110 | |||
944 | def getSubscription(self, name, remove=True, root=None): |
|
1111 | def getSubscription(self, name, remove=True, root=None): | |
945 | """ Retrieve the data associated with a named subscription |
|
1112 | """ Retrieve the data associated with a named subscription | |
946 |
|
1113 | |||
947 | If remove is True (the default), the subscription data is removed |
|
1114 | If remove is True (the default), the subscription data is removed | |
948 | from the buffer. Otherwise the data is returned but left in |
|
1115 | from the buffer. Otherwise the data is returned but left in | |
949 | the buffer. |
|
1116 | the buffer. | |
950 |
|
1117 | |||
951 | Returns None if there is no data associated with `name` |
|
1118 | Returns None if there is no data associated with `name` | |
952 |
|
1119 | |||
953 | If root is not None, then only return the subscription |
|
1120 | If root is not None, then only return the subscription | |
954 | data that matches both root and name. When used in this way, |
|
1121 | data that matches both root and name. When used in this way, | |
955 | remove processing impacts both the unscoped and scoped stores |
|
1122 | remove processing impacts both the unscoped and scoped stores | |
956 | for the subscription data. |
|
1123 | for the subscription data. | |
957 | """ |
|
1124 | """ | |
958 | if compat.PYTHON3 and issubclass(self.recvCodec, BserCodec): |
|
|||
959 | # People may pass in Unicode strings here -- but currently BSER only |
|
|||
960 | # returns bytestrings. Deal with that. |
|
|||
961 | if isinstance(root, str): |
|
|||
962 | root = encoding.encode_local(root) |
|
|||
963 | if isinstance(name, str): |
|
|||
964 | name = name.encode('utf-8') |
|
|||
965 |
|
||||
966 | if root is not None: |
|
1125 | if root is not None: | |
967 | if not root in self.sub_by_root: |
|
1126 | root = os.path.normpath(os.path.normcase(root)) | |
|
1127 | if root not in self.sub_by_root: | |||
968 | return None |
|
1128 | return None | |
969 |
if not |
|
1129 | if name not in self.sub_by_root[root]: | |
970 | return None |
|
1130 | return None | |
971 | sub = self.sub_by_root[root][name] |
|
1131 | sub = self.sub_by_root[root][name] | |
972 | if remove: |
|
1132 | if remove: | |
973 | del self.sub_by_root[root][name] |
|
1133 | del self.sub_by_root[root][name] | |
974 | # don't let this grow unbounded |
|
1134 | # don't let this grow unbounded | |
975 | if name in self.subs: |
|
1135 | if name in self.subs: | |
976 | del self.subs[name] |
|
1136 | del self.subs[name] | |
977 | return sub |
|
1137 | return sub | |
978 |
|
1138 | |||
979 |
if not |
|
1139 | if name not in self.subs: | |
980 | return None |
|
1140 | return None | |
981 | sub = self.subs[name] |
|
1141 | sub = self.subs[name] | |
982 | if remove: |
|
1142 | if remove: | |
983 | del self.subs[name] |
|
1143 | del self.subs[name] | |
984 | return sub |
|
1144 | return sub | |
985 |
|
1145 | |||
986 | def query(self, *args): |
|
1146 | def query(self, *args): | |
987 | """ Send a query to the watchman service and return the response |
|
1147 | """ Send a query to the watchman service and return the response | |
988 |
|
1148 | |||
989 | This call will block until the response is returned. |
|
1149 | This call will block until the response is returned. | |
990 | If any unilateral responses are sent by the service in between |
|
1150 | If any unilateral responses are sent by the service in between | |
991 | the request-response they will be buffered up in the client object |
|
1151 | the request-response they will be buffered up in the client object | |
992 | and NOT returned via this method. |
|
1152 | and NOT returned via this method. | |
993 | """ |
|
1153 | """ | |
994 |
|
1154 | |||
995 |
log( |
|
1155 | log("calling client.query") | |
996 | self._connect() |
|
1156 | self._connect() | |
997 | try: |
|
1157 | try: | |
998 | self.sendConn.send(args) |
|
1158 | self.sendConn.send(args) | |
999 |
|
1159 | |||
1000 | res = self.receive() |
|
1160 | res = self.receive() | |
1001 | while self.isUnilateralResponse(res): |
|
1161 | while self.isUnilateralResponse(res): | |
1002 | res = self.receive() |
|
1162 | res = self.receive() | |
1003 |
|
1163 | |||
1004 | return res |
|
1164 | return res | |
1005 | except EnvironmentError as ee: |
|
1165 | except EnvironmentError as ee: | |
1006 | # When we can depend on Python 3, we can use PEP 3134 |
|
1166 | # When we can depend on Python 3, we can use PEP 3134 | |
1007 | # exception chaining here. |
|
1167 | # exception chaining here. | |
1008 | raise WatchmanEnvironmentError( |
|
1168 | raise WatchmanEnvironmentError( | |
1009 |
|
|
1169 | "I/O error communicating with watchman daemon", | |
1010 | ee.errno, |
|
1170 | ee.errno, | |
1011 | ee.strerror, |
|
1171 | ee.strerror, | |
1012 |
args |
|
1172 | args, | |
|
1173 | ) | |||
1013 | except WatchmanError as ex: |
|
1174 | except WatchmanError as ex: | |
1014 | ex.setCommand(args) |
|
1175 | ex.setCommand(args) | |
1015 | raise |
|
1176 | raise | |
1016 |
|
1177 | |||
1017 | def capabilityCheck(self, optional=None, required=None): |
|
1178 | def capabilityCheck(self, optional=None, required=None): | |
1018 | """ Perform a server capability check """ |
|
1179 | """ Perform a server capability check """ | |
1019 |
res = self.query( |
|
1180 | res = self.query( | |
1020 |
|
|
1181 | "version", {"optional": optional or [], "required": required or []} | |
1021 | 'required': required or [] |
|
1182 | ) | |
1022 | }) |
|
|||
1023 |
|
1183 | |||
1024 |
if not self._hasprop(res, |
|
1184 | if not self._hasprop(res, "capabilities"): | |
1025 | # Server doesn't support capabilities, so we need to |
|
1185 | # Server doesn't support capabilities, so we need to | |
1026 | # synthesize the results based on the version |
|
1186 | # synthesize the results based on the version | |
1027 | capabilities.synthesize(res, optional) |
|
1187 | capabilities.synthesize(res, optional) | |
1028 |
if |
|
1188 | if "error" in res: | |
1029 |
raise CommandError(res[ |
|
1189 | raise CommandError(res["error"]) | |
1030 |
|
1190 | |||
1031 | return res |
|
1191 | return res | |
1032 |
|
1192 | |||
1033 | def setTimeout(self, value): |
|
1193 | def setTimeout(self, value): | |
1034 | self.recvConn.setTimeout(value) |
|
1194 | self.recvConn.setTimeout(value) | |
1035 | self.sendConn.setTimeout(value) |
|
1195 | self.sendConn.setTimeout(value) |
@@ -1,1223 +1,1241 b'' | |||||
1 | /* |
|
1 | /* | |
2 | Copyright (c) 2013-2015, Facebook, Inc. |
|
2 | Copyright (c) 2013-2015, Facebook, Inc. | |
3 | All rights reserved. |
|
3 | All rights reserved. | |
4 |
|
4 | |||
5 | Redistribution and use in source and binary forms, with or without |
|
5 | Redistribution and use in source and binary forms, with or without | |
6 | modification, are permitted provided that the following conditions are met: |
|
6 | modification, are permitted provided that the following conditions are met: | |
7 |
|
7 | |||
8 | * Redistributions of source code must retain the above copyright notice, |
|
8 | * Redistributions of source code must retain the above copyright notice, | |
9 | this list of conditions and the following disclaimer. |
|
9 | this list of conditions and the following disclaimer. | |
10 |
|
10 | |||
11 | * Redistributions in binary form must reproduce the above copyright notice, |
|
11 | * Redistributions in binary form must reproduce the above copyright notice, | |
12 | this list of conditions and the following disclaimer in the documentation |
|
12 | this list of conditions and the following disclaimer in the documentation | |
13 | and/or other materials provided with the distribution. |
|
13 | and/or other materials provided with the distribution. | |
14 |
|
14 | |||
15 | * Neither the name Facebook nor the names of its contributors may be used to |
|
15 | * Neither the name Facebook nor the names of its contributors may be used to | |
16 | endorse or promote products derived from this software without specific |
|
16 | endorse or promote products derived from this software without specific | |
17 | prior written permission. |
|
17 | prior written permission. | |
18 |
|
18 | |||
19 | THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
|
19 | THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" | |
20 | AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
|
20 | AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
21 | IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
|
21 | IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE | |
22 | DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
|
22 | DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE | |
23 | FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
|
23 | FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
24 | DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
|
24 | DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR | |
25 | SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
|
25 | SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER | |
26 | CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
|
26 | CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, | |
27 | OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
|
27 | OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
28 | OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
|
28 | OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
29 | */ |
|
29 | */ | |
30 |
|
30 | |||
31 | #include <Python.h> |
|
31 | #include <Python.h> | |
32 | #include <bytesobject.h> |
|
32 | #include <bytesobject.h> | |
33 | #ifdef _MSC_VER |
|
33 | #ifdef _MSC_VER | |
34 | #define inline __inline |
|
34 | #define inline __inline | |
35 | #if _MSC_VER >= 1800 |
|
35 | #if _MSC_VER >= 1800 | |
36 | #include <stdint.h> |
|
36 | #include <stdint.h> | |
37 | #else |
|
37 | #else | |
38 | // The compiler associated with Python 2.7 on Windows doesn't ship |
|
38 | // The compiler associated with Python 2.7 on Windows doesn't ship | |
39 | // with stdint.h, so define the small subset that we use here. |
|
39 | // with stdint.h, so define the small subset that we use here. | |
40 | typedef __int8 int8_t; |
|
40 | typedef __int8 int8_t; | |
41 | typedef __int16 int16_t; |
|
41 | typedef __int16 int16_t; | |
42 | typedef __int32 int32_t; |
|
42 | typedef __int32 int32_t; | |
43 | typedef __int64 int64_t; |
|
43 | typedef __int64 int64_t; | |
44 | typedef unsigned __int8 uint8_t; |
|
44 | typedef unsigned __int8 uint8_t; | |
45 | typedef unsigned __int16 uint16_t; |
|
45 | typedef unsigned __int16 uint16_t; | |
46 | typedef unsigned __int32 uint32_t; |
|
46 | typedef unsigned __int32 uint32_t; | |
47 | typedef unsigned __int64 uint64_t; |
|
47 | typedef unsigned __int64 uint64_t; | |
48 | #define UINT32_MAX 4294967295U |
|
48 | #define UINT32_MAX 4294967295U | |
49 | #endif |
|
49 | #endif | |
50 | #endif |
|
50 | #endif | |
51 |
|
51 | |||
52 | // clang-format off |
|
52 | // clang-format off | |
53 | /* Return the smallest size int that can store the value */ |
|
53 | /* Return the smallest size int that can store the value */ | |
54 | #define INT_SIZE(x) (((x) == ((int8_t)x)) ? 1 : \ |
|
54 | #define INT_SIZE(x) (((x) == ((int8_t)x)) ? 1 : \ | |
55 | ((x) == ((int16_t)x)) ? 2 : \ |
|
55 | ((x) == ((int16_t)x)) ? 2 : \ | |
56 | ((x) == ((int32_t)x)) ? 4 : 8) |
|
56 | ((x) == ((int32_t)x)) ? 4 : 8) | |
57 |
|
57 | |||
58 | #define BSER_ARRAY 0x00 |
|
58 | #define BSER_ARRAY 0x00 | |
59 | #define BSER_OBJECT 0x01 |
|
59 | #define BSER_OBJECT 0x01 | |
60 | #define BSER_BYTESTRING 0x02 |
|
60 | #define BSER_BYTESTRING 0x02 | |
61 | #define BSER_INT8 0x03 |
|
61 | #define BSER_INT8 0x03 | |
62 | #define BSER_INT16 0x04 |
|
62 | #define BSER_INT16 0x04 | |
63 | #define BSER_INT32 0x05 |
|
63 | #define BSER_INT32 0x05 | |
64 | #define BSER_INT64 0x06 |
|
64 | #define BSER_INT64 0x06 | |
65 | #define BSER_REAL 0x07 |
|
65 | #define BSER_REAL 0x07 | |
66 | #define BSER_TRUE 0x08 |
|
66 | #define BSER_TRUE 0x08 | |
67 | #define BSER_FALSE 0x09 |
|
67 | #define BSER_FALSE 0x09 | |
68 | #define BSER_NULL 0x0a |
|
68 | #define BSER_NULL 0x0a | |
69 | #define BSER_TEMPLATE 0x0b |
|
69 | #define BSER_TEMPLATE 0x0b | |
70 | #define BSER_SKIP 0x0c |
|
70 | #define BSER_SKIP 0x0c | |
71 | #define BSER_UTF8STRING 0x0d |
|
71 | #define BSER_UTF8STRING 0x0d | |
72 | // clang-format on |
|
72 | // clang-format on | |
73 |
|
73 | |||
74 | // An immutable object representation of BSER_OBJECT. |
|
74 | // An immutable object representation of BSER_OBJECT. | |
75 | // Rather than build a hash table, key -> value are obtained |
|
75 | // Rather than build a hash table, key -> value are obtained | |
76 | // by walking the list of keys to determine the offset into |
|
76 | // by walking the list of keys to determine the offset into | |
77 | // the values array. The assumption is that the number of |
|
77 | // the values array. The assumption is that the number of | |
78 | // array elements will be typically small (~6 for the top |
|
78 | // array elements will be typically small (~6 for the top | |
79 | // level query result and typically 3-5 for the file entries) |
|
79 | // level query result and typically 3-5 for the file entries) | |
80 | // so that the time overhead for this is small compared to |
|
80 | // so that the time overhead for this is small compared to | |
81 | // using a proper hash table. Even with this simplistic |
|
81 | // using a proper hash table. Even with this simplistic | |
82 | // approach, this is still faster for the mercurial use case |
|
82 | // approach, this is still faster for the mercurial use case | |
83 | // as it helps to eliminate creating N other objects to |
|
83 | // as it helps to eliminate creating N other objects to | |
84 | // represent the stat information in the hgwatchman extension |
|
84 | // represent the stat information in the hgwatchman extension | |
85 | // clang-format off |
|
85 | // clang-format off | |
86 | typedef struct { |
|
86 | typedef struct { | |
87 | PyObject_HEAD |
|
87 | PyObject_HEAD | |
88 | PyObject *keys; // tuple of field names |
|
88 | PyObject *keys; // tuple of field names | |
89 | PyObject *values; // tuple of values |
|
89 | PyObject *values; // tuple of values | |
90 | } bserObject; |
|
90 | } bserObject; | |
91 | // clang-format on |
|
91 | // clang-format on | |
92 |
|
92 | |||
93 | static Py_ssize_t bserobj_tuple_length(PyObject* o) { |
|
93 | static Py_ssize_t bserobj_tuple_length(PyObject* o) { | |
94 | bserObject* obj = (bserObject*)o; |
|
94 | bserObject* obj = (bserObject*)o; | |
95 |
|
95 | |||
96 | return PySequence_Length(obj->keys); |
|
96 | return PySequence_Length(obj->keys); | |
97 | } |
|
97 | } | |
98 |
|
98 | |||
99 | static PyObject* bserobj_tuple_item(PyObject* o, Py_ssize_t i) { |
|
99 | static PyObject* bserobj_tuple_item(PyObject* o, Py_ssize_t i) { | |
100 | bserObject* obj = (bserObject*)o; |
|
100 | bserObject* obj = (bserObject*)o; | |
101 |
|
101 | |||
102 | return PySequence_GetItem(obj->values, i); |
|
102 | return PySequence_GetItem(obj->values, i); | |
103 | } |
|
103 | } | |
104 |
|
104 | |||
105 | // clang-format off |
|
105 | // clang-format off | |
106 | static PySequenceMethods bserobj_sq = { |
|
106 | static PySequenceMethods bserobj_sq = { | |
107 | bserobj_tuple_length, /* sq_length */ |
|
107 | bserobj_tuple_length, /* sq_length */ | |
108 | 0, /* sq_concat */ |
|
108 | 0, /* sq_concat */ | |
109 | 0, /* sq_repeat */ |
|
109 | 0, /* sq_repeat */ | |
110 | bserobj_tuple_item, /* sq_item */ |
|
110 | bserobj_tuple_item, /* sq_item */ | |
111 | 0, /* sq_ass_item */ |
|
111 | 0, /* sq_ass_item */ | |
112 | 0, /* sq_contains */ |
|
112 | 0, /* sq_contains */ | |
113 | 0, /* sq_inplace_concat */ |
|
113 | 0, /* sq_inplace_concat */ | |
114 | 0 /* sq_inplace_repeat */ |
|
114 | 0 /* sq_inplace_repeat */ | |
115 | }; |
|
115 | }; | |
116 | // clang-format on |
|
116 | // clang-format on | |
117 |
|
117 | |||
118 | static void bserobj_dealloc(PyObject* o) { |
|
118 | static void bserobj_dealloc(PyObject* o) { | |
119 | bserObject* obj = (bserObject*)o; |
|
119 | bserObject* obj = (bserObject*)o; | |
120 |
|
120 | |||
121 | Py_CLEAR(obj->keys); |
|
121 | Py_CLEAR(obj->keys); | |
122 | Py_CLEAR(obj->values); |
|
122 | Py_CLEAR(obj->values); | |
123 | PyObject_Del(o); |
|
123 | PyObject_Del(o); | |
124 | } |
|
124 | } | |
125 |
|
125 | |||
126 | static PyObject* bserobj_getattrro(PyObject* o, PyObject* name) { |
|
126 | static PyObject* bserobj_getattrro(PyObject* o, PyObject* name) { | |
127 | bserObject* obj = (bserObject*)o; |
|
127 | bserObject* obj = (bserObject*)o; | |
128 | Py_ssize_t i, n; |
|
128 | Py_ssize_t i, n; | |
129 | PyObject* name_bytes = NULL; |
|
129 | PyObject* name_bytes = NULL; | |
130 | PyObject* ret = NULL; |
|
130 | PyObject* ret = NULL; | |
131 |
const char* namestr |
|
131 | const char* namestr; | |
132 |
|
132 | |||
133 | if (PyIndex_Check(name)) { |
|
133 | if (PyIndex_Check(name)) { | |
134 | i = PyNumber_AsSsize_t(name, PyExc_IndexError); |
|
134 | i = PyNumber_AsSsize_t(name, PyExc_IndexError); | |
135 | if (i == -1 && PyErr_Occurred()) { |
|
135 | if (i == -1 && PyErr_Occurred()) { | |
136 | goto bail; |
|
136 | goto bail; | |
137 | } |
|
137 | } | |
|
138 | ret = PySequence_GetItem(obj->values, i); | |||
|
139 | goto bail; | |||
|
140 | } | |||
138 |
|
141 | |||
139 | if (i == 8 && PySequence_Size(obj->values) < 9) { |
|
142 | // We can be passed in Unicode objects here -- we don't support anything other | |
140 | // Hack alert: Python 3 removed support for os.stat().st_mtime |
|
143 | // than UTF-8 for keys. | |
141 | // being an integer.Instead, if you need an integer, you have to |
|
144 | if (PyUnicode_Check(name)) { | |
142 | // use os.stat()[stat.ST_MTIME] instead. stat.ST_MTIME is 8, and |
|
145 | name_bytes = PyUnicode_AsUTF8String(name); | |
143 | // our stat tuples are shorter than that, so we can detect |
|
146 | if (name_bytes == NULL) { | |
144 | // requests for index 8 on tuples shorter than that and return |
|
|||
145 | // st_mtime instead. |
|
|||
146 | namestr = "st_mtime"; |
|
|||
147 | } else { |
|
|||
148 | ret = PySequence_GetItem(obj->values, i); |
|
|||
149 | goto bail; |
|
147 | goto bail; | |
150 | } |
|
148 | } | |
|
149 | namestr = PyBytes_AsString(name_bytes); | |||
151 | } else { |
|
150 | } else { | |
152 | // We can be passed in Unicode objects here -- we don't support anything other |
|
151 | namestr = PyBytes_AsString(name); | |
153 | // than UTF-8 for keys. |
|
|||
154 | if (PyUnicode_Check(name)) { |
|
|||
155 | name_bytes = PyUnicode_AsUTF8String(name); |
|
|||
156 | if (name_bytes == NULL) { |
|
|||
157 | goto bail; |
|
|||
158 | } |
|
|||
159 | namestr = PyBytes_AsString(name_bytes); |
|
|||
160 | } else { |
|
|||
161 | namestr = PyBytes_AsString(name); |
|
|||
162 | } |
|
|||
163 | } |
|
152 | } | |
164 |
|
153 | |||
165 | if (namestr == NULL) { |
|
154 | if (namestr == NULL) { | |
166 | goto bail; |
|
155 | goto bail; | |
167 | } |
|
156 | } | |
168 | // hack^Wfeature to allow mercurial to use "st_size" to reference "size" |
|
157 | // hack^Wfeature to allow mercurial to use "st_size" to reference "size" | |
169 | if (!strncmp(namestr, "st_", 3)) { |
|
158 | if (!strncmp(namestr, "st_", 3)) { | |
170 | namestr += 3; |
|
159 | namestr += 3; | |
171 | } |
|
160 | } | |
172 |
|
161 | |||
173 | n = PyTuple_GET_SIZE(obj->keys); |
|
162 | n = PyTuple_GET_SIZE(obj->keys); | |
174 | for (i = 0; i < n; i++) { |
|
163 | for (i = 0; i < n; i++) { | |
175 | const char* item_name = NULL; |
|
164 | const char* item_name = NULL; | |
176 | PyObject* key = PyTuple_GET_ITEM(obj->keys, i); |
|
165 | PyObject* key = PyTuple_GET_ITEM(obj->keys, i); | |
177 |
|
166 | |||
178 | item_name = PyBytes_AsString(key); |
|
167 | item_name = PyBytes_AsString(key); | |
179 | if (!strcmp(item_name, namestr)) { |
|
168 | if (!strcmp(item_name, namestr)) { | |
180 | ret = PySequence_GetItem(obj->values, i); |
|
169 | ret = PySequence_GetItem(obj->values, i); | |
181 | goto bail; |
|
170 | goto bail; | |
182 | } |
|
171 | } | |
183 | } |
|
172 | } | |
184 |
|
173 | |||
185 | PyErr_Format( |
|
174 | PyErr_Format( | |
186 | PyExc_AttributeError, "bserobject has no attribute '%.400s'", namestr); |
|
175 | PyExc_AttributeError, "bserobject has no attribute '%.400s'", namestr); | |
187 | bail: |
|
176 | bail: | |
188 | Py_XDECREF(name_bytes); |
|
177 | Py_XDECREF(name_bytes); | |
189 | return ret; |
|
178 | return ret; | |
190 | } |
|
179 | } | |
191 |
|
180 | |||
192 | // clang-format off |
|
181 | // clang-format off | |
193 | static PyMappingMethods bserobj_map = { |
|
182 | static PyMappingMethods bserobj_map = { | |
194 | bserobj_tuple_length, /* mp_length */ |
|
183 | bserobj_tuple_length, /* mp_length */ | |
195 | bserobj_getattrro, /* mp_subscript */ |
|
184 | bserobj_getattrro, /* mp_subscript */ | |
196 | 0 /* mp_ass_subscript */ |
|
185 | 0 /* mp_ass_subscript */ | |
197 | }; |
|
186 | }; | |
198 |
|
187 | |||
199 | PyTypeObject bserObjectType = { |
|
188 | PyTypeObject bserObjectType = { | |
200 | PyVarObject_HEAD_INIT(NULL, 0) |
|
189 | PyVarObject_HEAD_INIT(NULL, 0) | |
201 | "bserobj_tuple", /* tp_name */ |
|
190 | "bserobj_tuple", /* tp_name */ | |
202 | sizeof(bserObject), /* tp_basicsize */ |
|
191 | sizeof(bserObject), /* tp_basicsize */ | |
203 | 0, /* tp_itemsize */ |
|
192 | 0, /* tp_itemsize */ | |
204 | bserobj_dealloc, /* tp_dealloc */ |
|
193 | bserobj_dealloc, /* tp_dealloc */ | |
205 | 0, /* tp_print */ |
|
194 | 0, /* tp_print */ | |
206 | 0, /* tp_getattr */ |
|
195 | 0, /* tp_getattr */ | |
207 | 0, /* tp_setattr */ |
|
196 | 0, /* tp_setattr */ | |
208 | 0, /* tp_compare */ |
|
197 | 0, /* tp_compare */ | |
209 | 0, /* tp_repr */ |
|
198 | 0, /* tp_repr */ | |
210 | 0, /* tp_as_number */ |
|
199 | 0, /* tp_as_number */ | |
211 | &bserobj_sq, /* tp_as_sequence */ |
|
200 | &bserobj_sq, /* tp_as_sequence */ | |
212 | &bserobj_map, /* tp_as_mapping */ |
|
201 | &bserobj_map, /* tp_as_mapping */ | |
213 | 0, /* tp_hash */ |
|
202 | 0, /* tp_hash */ | |
214 | 0, /* tp_call */ |
|
203 | 0, /* tp_call */ | |
215 | 0, /* tp_str */ |
|
204 | 0, /* tp_str */ | |
216 | bserobj_getattrro, /* tp_getattro */ |
|
205 | bserobj_getattrro, /* tp_getattro */ | |
217 | 0, /* tp_setattro */ |
|
206 | 0, /* tp_setattro */ | |
218 | 0, /* tp_as_buffer */ |
|
207 | 0, /* tp_as_buffer */ | |
219 | Py_TPFLAGS_DEFAULT, /* tp_flags */ |
|
208 | Py_TPFLAGS_DEFAULT, /* tp_flags */ | |
220 | "bserobj tuple", /* tp_doc */ |
|
209 | "bserobj tuple", /* tp_doc */ | |
221 | 0, /* tp_traverse */ |
|
210 | 0, /* tp_traverse */ | |
222 | 0, /* tp_clear */ |
|
211 | 0, /* tp_clear */ | |
223 | 0, /* tp_richcompare */ |
|
212 | 0, /* tp_richcompare */ | |
224 | 0, /* tp_weaklistoffset */ |
|
213 | 0, /* tp_weaklistoffset */ | |
225 | 0, /* tp_iter */ |
|
214 | 0, /* tp_iter */ | |
226 | 0, /* tp_iternext */ |
|
215 | 0, /* tp_iternext */ | |
227 | 0, /* tp_methods */ |
|
216 | 0, /* tp_methods */ | |
228 | 0, /* tp_members */ |
|
217 | 0, /* tp_members */ | |
229 | 0, /* tp_getset */ |
|
218 | 0, /* tp_getset */ | |
230 | 0, /* tp_base */ |
|
219 | 0, /* tp_base */ | |
231 | 0, /* tp_dict */ |
|
220 | 0, /* tp_dict */ | |
232 | 0, /* tp_descr_get */ |
|
221 | 0, /* tp_descr_get */ | |
233 | 0, /* tp_descr_set */ |
|
222 | 0, /* tp_descr_set */ | |
234 | 0, /* tp_dictoffset */ |
|
223 | 0, /* tp_dictoffset */ | |
235 | 0, /* tp_init */ |
|
224 | 0, /* tp_init */ | |
236 | 0, /* tp_alloc */ |
|
225 | 0, /* tp_alloc */ | |
237 | 0, /* tp_new */ |
|
226 | 0, /* tp_new */ | |
238 | }; |
|
227 | }; | |
239 | // clang-format on |
|
228 | // clang-format on | |
240 |
|
229 | |||
241 | typedef struct loads_ctx { |
|
230 | typedef struct loads_ctx { | |
242 | int mutable; |
|
231 | int mutable; | |
243 | const char* value_encoding; |
|
232 | const char* value_encoding; | |
244 | const char* value_errors; |
|
233 | const char* value_errors; | |
245 | uint32_t bser_version; |
|
234 | uint32_t bser_version; | |
246 | uint32_t bser_capabilities; |
|
235 | uint32_t bser_capabilities; | |
247 | } unser_ctx_t; |
|
236 | } unser_ctx_t; | |
248 |
|
237 | |||
249 | static PyObject* |
|
238 | static PyObject* | |
250 | bser_loads_recursive(const char** ptr, const char* end, const unser_ctx_t* ctx); |
|
239 | bser_loads_recursive(const char** ptr, const char* end, const unser_ctx_t* ctx); | |
251 |
|
240 | |||
252 | static const char bser_true = BSER_TRUE; |
|
241 | static const char bser_true = BSER_TRUE; | |
253 | static const char bser_false = BSER_FALSE; |
|
242 | static const char bser_false = BSER_FALSE; | |
254 | static const char bser_null = BSER_NULL; |
|
243 | static const char bser_null = BSER_NULL; | |
255 | static const char bser_bytestring_hdr = BSER_BYTESTRING; |
|
244 | static const char bser_bytestring_hdr = BSER_BYTESTRING; | |
256 | static const char bser_array_hdr = BSER_ARRAY; |
|
245 | static const char bser_array_hdr = BSER_ARRAY; | |
257 | static const char bser_object_hdr = BSER_OBJECT; |
|
246 | static const char bser_object_hdr = BSER_OBJECT; | |
258 |
|
247 | |||
259 | static inline uint32_t next_power_2(uint32_t n) { |
|
248 | static inline uint32_t next_power_2(uint32_t n) { | |
260 | n |= (n >> 16); |
|
249 | n |= (n >> 16); | |
261 | n |= (n >> 8); |
|
250 | n |= (n >> 8); | |
262 | n |= (n >> 4); |
|
251 | n |= (n >> 4); | |
263 | n |= (n >> 2); |
|
252 | n |= (n >> 2); | |
264 | n |= (n >> 1); |
|
253 | n |= (n >> 1); | |
265 | return n + 1; |
|
254 | return n + 1; | |
266 | } |
|
255 | } | |
267 |
|
256 | |||
268 | // A buffer we use for building up the serialized result |
|
257 | // A buffer we use for building up the serialized result | |
269 | struct bser_buffer { |
|
258 | struct bser_buffer { | |
270 | char* buf; |
|
259 | char* buf; | |
271 | int wpos, allocd; |
|
260 | int wpos, allocd; | |
272 | uint32_t bser_version; |
|
261 | uint32_t bser_version; | |
273 | uint32_t capabilities; |
|
262 | uint32_t capabilities; | |
274 | }; |
|
263 | }; | |
275 | typedef struct bser_buffer bser_t; |
|
264 | typedef struct bser_buffer bser_t; | |
276 |
|
265 | |||
277 | static int bser_append(bser_t* bser, const char* data, uint32_t len) { |
|
266 | static int bser_append(bser_t* bser, const char* data, uint32_t len) { | |
278 | int newlen = next_power_2(bser->wpos + len); |
|
267 | int newlen = next_power_2(bser->wpos + len); | |
279 | if (newlen > bser->allocd) { |
|
268 | if (newlen > bser->allocd) { | |
280 | char* nbuf = realloc(bser->buf, newlen); |
|
269 | char* nbuf = realloc(bser->buf, newlen); | |
281 | if (!nbuf) { |
|
270 | if (!nbuf) { | |
282 | return 0; |
|
271 | return 0; | |
283 | } |
|
272 | } | |
284 |
|
273 | |||
285 | bser->buf = nbuf; |
|
274 | bser->buf = nbuf; | |
286 | bser->allocd = newlen; |
|
275 | bser->allocd = newlen; | |
287 | } |
|
276 | } | |
288 |
|
277 | |||
289 | memcpy(bser->buf + bser->wpos, data, len); |
|
278 | memcpy(bser->buf + bser->wpos, data, len); | |
290 | bser->wpos += len; |
|
279 | bser->wpos += len; | |
291 | return 1; |
|
280 | return 1; | |
292 | } |
|
281 | } | |
293 |
|
282 | |||
294 | static int bser_init(bser_t* bser, uint32_t version, uint32_t capabilities) { |
|
283 | static int bser_init(bser_t* bser, uint32_t version, uint32_t capabilities) { | |
295 | bser->allocd = 8192; |
|
284 | bser->allocd = 8192; | |
296 | bser->wpos = 0; |
|
285 | bser->wpos = 0; | |
297 | bser->buf = malloc(bser->allocd); |
|
286 | bser->buf = malloc(bser->allocd); | |
298 | bser->bser_version = version; |
|
287 | bser->bser_version = version; | |
299 | bser->capabilities = capabilities; |
|
288 | bser->capabilities = capabilities; | |
300 | if (!bser->buf) { |
|
289 | if (!bser->buf) { | |
301 | return 0; |
|
290 | return 0; | |
302 | } |
|
291 | } | |
303 |
|
292 | |||
304 | // Leave room for the serialization header, which includes |
|
293 | // Leave room for the serialization header, which includes | |
305 | // our overall length. To make things simpler, we'll use an |
|
294 | // our overall length. To make things simpler, we'll use an | |
306 | // int32 for the header |
|
295 | // int32 for the header | |
307 | #define EMPTY_HEADER "\x00\x01\x05\x00\x00\x00\x00" |
|
296 | #define EMPTY_HEADER "\x00\x01\x05\x00\x00\x00\x00" | |
308 |
|
297 | |||
309 | // Version 2 also carries an integer indicating the capabilities. The |
|
298 | // Version 2 also carries an integer indicating the capabilities. The | |
310 | // capabilities integer comes before the PDU size. |
|
299 | // capabilities integer comes before the PDU size. | |
311 | #define EMPTY_HEADER_V2 "\x00\x02\x00\x00\x00\x00\x05\x00\x00\x00\x00" |
|
300 | #define EMPTY_HEADER_V2 "\x00\x02\x00\x00\x00\x00\x05\x00\x00\x00\x00" | |
312 | if (version == 2) { |
|
301 | if (version == 2) { | |
313 | bser_append(bser, EMPTY_HEADER_V2, sizeof(EMPTY_HEADER_V2) - 1); |
|
302 | bser_append(bser, EMPTY_HEADER_V2, sizeof(EMPTY_HEADER_V2) - 1); | |
314 | } else { |
|
303 | } else { | |
315 | bser_append(bser, EMPTY_HEADER, sizeof(EMPTY_HEADER) - 1); |
|
304 | bser_append(bser, EMPTY_HEADER, sizeof(EMPTY_HEADER) - 1); | |
316 | } |
|
305 | } | |
317 |
|
306 | |||
318 | return 1; |
|
307 | return 1; | |
319 | } |
|
308 | } | |
320 |
|
309 | |||
321 | static void bser_dtor(bser_t* bser) { |
|
310 | static void bser_dtor(bser_t* bser) { | |
322 | free(bser->buf); |
|
311 | free(bser->buf); | |
323 | bser->buf = NULL; |
|
312 | bser->buf = NULL; | |
324 | } |
|
313 | } | |
325 |
|
314 | |||
326 | static int bser_long(bser_t* bser, int64_t val) { |
|
315 | static int bser_long(bser_t* bser, int64_t val) { | |
327 | int8_t i8; |
|
316 | int8_t i8; | |
328 | int16_t i16; |
|
317 | int16_t i16; | |
329 | int32_t i32; |
|
318 | int32_t i32; | |
330 | int64_t i64; |
|
319 | int64_t i64; | |
331 | char sz; |
|
320 | char sz; | |
332 | int size = INT_SIZE(val); |
|
321 | int size = INT_SIZE(val); | |
333 | char* iptr; |
|
322 | char* iptr; | |
334 |
|
323 | |||
335 | switch (size) { |
|
324 | switch (size) { | |
336 | case 1: |
|
325 | case 1: | |
337 | sz = BSER_INT8; |
|
326 | sz = BSER_INT8; | |
338 | i8 = (int8_t)val; |
|
327 | i8 = (int8_t)val; | |
339 | iptr = (char*)&i8; |
|
328 | iptr = (char*)&i8; | |
340 | break; |
|
329 | break; | |
341 | case 2: |
|
330 | case 2: | |
342 | sz = BSER_INT16; |
|
331 | sz = BSER_INT16; | |
343 | i16 = (int16_t)val; |
|
332 | i16 = (int16_t)val; | |
344 | iptr = (char*)&i16; |
|
333 | iptr = (char*)&i16; | |
345 | break; |
|
334 | break; | |
346 | case 4: |
|
335 | case 4: | |
347 | sz = BSER_INT32; |
|
336 | sz = BSER_INT32; | |
348 | i32 = (int32_t)val; |
|
337 | i32 = (int32_t)val; | |
349 | iptr = (char*)&i32; |
|
338 | iptr = (char*)&i32; | |
350 | break; |
|
339 | break; | |
351 | case 8: |
|
340 | case 8: | |
352 | sz = BSER_INT64; |
|
341 | sz = BSER_INT64; | |
353 | i64 = (int64_t)val; |
|
342 | i64 = (int64_t)val; | |
354 | iptr = (char*)&i64; |
|
343 | iptr = (char*)&i64; | |
355 | break; |
|
344 | break; | |
356 | default: |
|
345 | default: | |
357 | PyErr_SetString(PyExc_RuntimeError, "Cannot represent this long value!?"); |
|
346 | PyErr_SetString(PyExc_RuntimeError, "Cannot represent this long value!?"); | |
358 | return 0; |
|
347 | return 0; | |
359 | } |
|
348 | } | |
360 |
|
349 | |||
361 | if (!bser_append(bser, &sz, sizeof(sz))) { |
|
350 | if (!bser_append(bser, &sz, sizeof(sz))) { | |
362 | return 0; |
|
351 | return 0; | |
363 | } |
|
352 | } | |
364 |
|
353 | |||
365 | return bser_append(bser, iptr, size); |
|
354 | return bser_append(bser, iptr, size); | |
366 | } |
|
355 | } | |
367 |
|
356 | |||
368 | static int bser_bytestring(bser_t* bser, PyObject* sval) { |
|
357 | static int bser_bytestring(bser_t* bser, PyObject* sval) { | |
369 | char* buf = NULL; |
|
358 | char* buf = NULL; | |
370 | Py_ssize_t len; |
|
359 | Py_ssize_t len; | |
371 | int res; |
|
360 | int res; | |
372 | PyObject* utf = NULL; |
|
361 | PyObject* utf = NULL; | |
373 |
|
362 | |||
374 | if (PyUnicode_Check(sval)) { |
|
363 | if (PyUnicode_Check(sval)) { | |
375 | utf = PyUnicode_AsEncodedString(sval, "utf-8", "ignore"); |
|
364 | utf = PyUnicode_AsEncodedString(sval, "utf-8", "ignore"); | |
376 | sval = utf; |
|
365 | sval = utf; | |
377 | } |
|
366 | } | |
378 |
|
367 | |||
379 | res = PyBytes_AsStringAndSize(sval, &buf, &len); |
|
368 | res = PyBytes_AsStringAndSize(sval, &buf, &len); | |
380 | if (res == -1) { |
|
369 | if (res == -1) { | |
381 | res = 0; |
|
370 | res = 0; | |
382 | goto out; |
|
371 | goto out; | |
383 | } |
|
372 | } | |
384 |
|
373 | |||
385 | if (!bser_append(bser, &bser_bytestring_hdr, sizeof(bser_bytestring_hdr))) { |
|
374 | if (!bser_append(bser, &bser_bytestring_hdr, sizeof(bser_bytestring_hdr))) { | |
386 | res = 0; |
|
375 | res = 0; | |
387 | goto out; |
|
376 | goto out; | |
388 | } |
|
377 | } | |
389 |
|
378 | |||
390 | if (!bser_long(bser, len)) { |
|
379 | if (!bser_long(bser, len)) { | |
391 | res = 0; |
|
380 | res = 0; | |
392 | goto out; |
|
381 | goto out; | |
393 | } |
|
382 | } | |
394 |
|
383 | |||
395 | if (len > UINT32_MAX) { |
|
384 | if (len > UINT32_MAX) { | |
396 | PyErr_Format(PyExc_ValueError, "string too big"); |
|
385 | PyErr_Format(PyExc_ValueError, "string too big"); | |
397 | res = 0; |
|
386 | res = 0; | |
398 | goto out; |
|
387 | goto out; | |
399 | } |
|
388 | } | |
400 |
|
389 | |||
401 | res = bser_append(bser, buf, (uint32_t)len); |
|
390 | res = bser_append(bser, buf, (uint32_t)len); | |
402 |
|
391 | |||
403 | out: |
|
392 | out: | |
404 | if (utf) { |
|
393 | if (utf) { | |
405 | Py_DECREF(utf); |
|
394 | Py_DECREF(utf); | |
406 | } |
|
395 | } | |
407 |
|
396 | |||
408 | return res; |
|
397 | return res; | |
409 | } |
|
398 | } | |
410 |
|
399 | |||
411 | static int bser_recursive(bser_t* bser, PyObject* val) { |
|
400 | static int bser_recursive(bser_t* bser, PyObject* val) { | |
412 | if (PyBool_Check(val)) { |
|
401 | if (PyBool_Check(val)) { | |
413 | if (val == Py_True) { |
|
402 | if (val == Py_True) { | |
414 | return bser_append(bser, &bser_true, sizeof(bser_true)); |
|
403 | return bser_append(bser, &bser_true, sizeof(bser_true)); | |
415 | } |
|
404 | } | |
416 | return bser_append(bser, &bser_false, sizeof(bser_false)); |
|
405 | return bser_append(bser, &bser_false, sizeof(bser_false)); | |
417 | } |
|
406 | } | |
418 |
|
407 | |||
419 | if (val == Py_None) { |
|
408 | if (val == Py_None) { | |
420 | return bser_append(bser, &bser_null, sizeof(bser_null)); |
|
409 | return bser_append(bser, &bser_null, sizeof(bser_null)); | |
421 | } |
|
410 | } | |
422 |
|
411 | |||
423 | // Python 3 has one integer type. |
|
412 | // Python 3 has one integer type. | |
424 | #if PY_MAJOR_VERSION < 3 |
|
413 | #if PY_MAJOR_VERSION < 3 | |
425 | if (PyInt_Check(val)) { |
|
414 | if (PyInt_Check(val)) { | |
426 | return bser_long(bser, PyInt_AS_LONG(val)); |
|
415 | return bser_long(bser, PyInt_AS_LONG(val)); | |
427 | } |
|
416 | } | |
428 | #endif // PY_MAJOR_VERSION < 3 |
|
417 | #endif // PY_MAJOR_VERSION < 3 | |
429 |
|
418 | |||
430 | if (PyLong_Check(val)) { |
|
419 | if (PyLong_Check(val)) { | |
431 | return bser_long(bser, PyLong_AsLongLong(val)); |
|
420 | return bser_long(bser, PyLong_AsLongLong(val)); | |
432 | } |
|
421 | } | |
433 |
|
422 | |||
434 | if (PyBytes_Check(val) || PyUnicode_Check(val)) { |
|
423 | if (PyBytes_Check(val) || PyUnicode_Check(val)) { | |
435 | return bser_bytestring(bser, val); |
|
424 | return bser_bytestring(bser, val); | |
436 | } |
|
425 | } | |
437 |
|
426 | |||
438 | if (PyFloat_Check(val)) { |
|
427 | if (PyFloat_Check(val)) { | |
439 | double dval = PyFloat_AS_DOUBLE(val); |
|
428 | double dval = PyFloat_AS_DOUBLE(val); | |
440 | char sz = BSER_REAL; |
|
429 | char sz = BSER_REAL; | |
441 |
|
430 | |||
442 | if (!bser_append(bser, &sz, sizeof(sz))) { |
|
431 | if (!bser_append(bser, &sz, sizeof(sz))) { | |
443 | return 0; |
|
432 | return 0; | |
444 | } |
|
433 | } | |
445 |
|
434 | |||
446 | return bser_append(bser, (char*)&dval, sizeof(dval)); |
|
435 | return bser_append(bser, (char*)&dval, sizeof(dval)); | |
447 | } |
|
436 | } | |
448 |
|
437 | |||
449 | if (PyList_Check(val)) { |
|
438 | if (PyList_Check(val)) { | |
450 | Py_ssize_t i, len = PyList_GET_SIZE(val); |
|
439 | Py_ssize_t i, len = PyList_GET_SIZE(val); | |
451 |
|
440 | |||
452 | if (!bser_append(bser, &bser_array_hdr, sizeof(bser_array_hdr))) { |
|
441 | if (!bser_append(bser, &bser_array_hdr, sizeof(bser_array_hdr))) { | |
453 | return 0; |
|
442 | return 0; | |
454 | } |
|
443 | } | |
455 |
|
444 | |||
456 | if (!bser_long(bser, len)) { |
|
445 | if (!bser_long(bser, len)) { | |
457 | return 0; |
|
446 | return 0; | |
458 | } |
|
447 | } | |
459 |
|
448 | |||
460 | for (i = 0; i < len; i++) { |
|
449 | for (i = 0; i < len; i++) { | |
461 | PyObject* ele = PyList_GET_ITEM(val, i); |
|
450 | PyObject* ele = PyList_GET_ITEM(val, i); | |
462 |
|
451 | |||
463 | if (!bser_recursive(bser, ele)) { |
|
452 | if (!bser_recursive(bser, ele)) { | |
464 | return 0; |
|
453 | return 0; | |
465 | } |
|
454 | } | |
466 | } |
|
455 | } | |
467 |
|
456 | |||
468 | return 1; |
|
457 | return 1; | |
469 | } |
|
458 | } | |
470 |
|
459 | |||
471 | if (PyTuple_Check(val)) { |
|
460 | if (PyTuple_Check(val)) { | |
472 | Py_ssize_t i, len = PyTuple_GET_SIZE(val); |
|
461 | Py_ssize_t i, len = PyTuple_GET_SIZE(val); | |
473 |
|
462 | |||
474 | if (!bser_append(bser, &bser_array_hdr, sizeof(bser_array_hdr))) { |
|
463 | if (!bser_append(bser, &bser_array_hdr, sizeof(bser_array_hdr))) { | |
475 | return 0; |
|
464 | return 0; | |
476 | } |
|
465 | } | |
477 |
|
466 | |||
478 | if (!bser_long(bser, len)) { |
|
467 | if (!bser_long(bser, len)) { | |
479 | return 0; |
|
468 | return 0; | |
480 | } |
|
469 | } | |
481 |
|
470 | |||
482 | for (i = 0; i < len; i++) { |
|
471 | for (i = 0; i < len; i++) { | |
483 | PyObject* ele = PyTuple_GET_ITEM(val, i); |
|
472 | PyObject* ele = PyTuple_GET_ITEM(val, i); | |
484 |
|
473 | |||
485 | if (!bser_recursive(bser, ele)) { |
|
474 | if (!bser_recursive(bser, ele)) { | |
486 | return 0; |
|
475 | return 0; | |
487 | } |
|
476 | } | |
488 | } |
|
477 | } | |
489 |
|
478 | |||
490 | return 1; |
|
479 | return 1; | |
491 | } |
|
480 | } | |
492 |
|
481 | |||
493 | if (PyMapping_Check(val)) { |
|
482 | if (PyMapping_Check(val)) { | |
494 | Py_ssize_t len = PyMapping_Length(val); |
|
483 | Py_ssize_t len = PyMapping_Length(val); | |
495 | Py_ssize_t pos = 0; |
|
484 | Py_ssize_t pos = 0; | |
496 | PyObject *key, *ele; |
|
485 | PyObject *key, *ele; | |
497 |
|
486 | |||
498 | if (!bser_append(bser, &bser_object_hdr, sizeof(bser_object_hdr))) { |
|
487 | if (!bser_append(bser, &bser_object_hdr, sizeof(bser_object_hdr))) { | |
499 | return 0; |
|
488 | return 0; | |
500 | } |
|
489 | } | |
501 |
|
490 | |||
502 | if (!bser_long(bser, len)) { |
|
491 | if (!bser_long(bser, len)) { | |
503 | return 0; |
|
492 | return 0; | |
504 | } |
|
493 | } | |
505 |
|
494 | |||
506 | while (PyDict_Next(val, &pos, &key, &ele)) { |
|
495 | while (PyDict_Next(val, &pos, &key, &ele)) { | |
507 | if (!bser_bytestring(bser, key)) { |
|
496 | if (!bser_bytestring(bser, key)) { | |
508 | return 0; |
|
497 | return 0; | |
509 | } |
|
498 | } | |
510 | if (!bser_recursive(bser, ele)) { |
|
499 | if (!bser_recursive(bser, ele)) { | |
511 | return 0; |
|
500 | return 0; | |
512 | } |
|
501 | } | |
513 | } |
|
502 | } | |
514 |
|
503 | |||
515 | return 1; |
|
504 | return 1; | |
516 | } |
|
505 | } | |
517 |
|
506 | |||
518 | PyErr_SetString(PyExc_ValueError, "Unsupported value type"); |
|
507 | PyErr_SetString(PyExc_ValueError, "Unsupported value type"); | |
519 | return 0; |
|
508 | return 0; | |
520 | } |
|
509 | } | |
521 |
|
510 | |||
522 | static PyObject* bser_dumps(PyObject* self, PyObject* args, PyObject* kw) { |
|
511 | static PyObject* bser_dumps(PyObject* self, PyObject* args, PyObject* kw) { | |
523 | PyObject *val = NULL, *res; |
|
512 | PyObject *val = NULL, *res; | |
524 | bser_t bser; |
|
513 | bser_t bser; | |
525 | uint32_t len, bser_version = 1, bser_capabilities = 0; |
|
514 | uint32_t len, bser_version = 1, bser_capabilities = 0; | |
526 |
|
515 | |||
527 | static char* kw_list[] = {"val", "version", "capabilities", NULL}; |
|
516 | static char* kw_list[] = {"val", "version", "capabilities", NULL}; | |
528 |
|
517 | |||
529 | if (!PyArg_ParseTupleAndKeywords( |
|
518 | if (!PyArg_ParseTupleAndKeywords( | |
530 | args, |
|
519 | args, | |
531 | kw, |
|
520 | kw, | |
532 | "O|ii:dumps", |
|
521 | "O|ii:dumps", | |
533 | kw_list, |
|
522 | kw_list, | |
534 | &val, |
|
523 | &val, | |
535 | &bser_version, |
|
524 | &bser_version, | |
536 | &bser_capabilities)) { |
|
525 | &bser_capabilities)) { | |
537 | return NULL; |
|
526 | return NULL; | |
538 | } |
|
527 | } | |
539 |
|
528 | |||
540 | if (!bser_init(&bser, bser_version, bser_capabilities)) { |
|
529 | if (!bser_init(&bser, bser_version, bser_capabilities)) { | |
541 | return PyErr_NoMemory(); |
|
530 | return PyErr_NoMemory(); | |
542 | } |
|
531 | } | |
543 |
|
532 | |||
544 | if (!bser_recursive(&bser, val)) { |
|
533 | if (!bser_recursive(&bser, val)) { | |
545 | bser_dtor(&bser); |
|
534 | bser_dtor(&bser); | |
546 | if (errno == ENOMEM) { |
|
535 | if (errno == ENOMEM) { | |
547 | return PyErr_NoMemory(); |
|
536 | return PyErr_NoMemory(); | |
548 | } |
|
537 | } | |
549 | // otherwise, we've already set the error to something reasonable |
|
538 | // otherwise, we've already set the error to something reasonable | |
550 | return NULL; |
|
539 | return NULL; | |
551 | } |
|
540 | } | |
552 |
|
541 | |||
553 | // Now fill in the overall length |
|
542 | // Now fill in the overall length | |
554 | if (bser_version == 1) { |
|
543 | if (bser_version == 1) { | |
555 | len = bser.wpos - (sizeof(EMPTY_HEADER) - 1); |
|
544 | len = bser.wpos - (sizeof(EMPTY_HEADER) - 1); | |
556 | memcpy(bser.buf + 3, &len, sizeof(len)); |
|
545 | memcpy(bser.buf + 3, &len, sizeof(len)); | |
557 | } else { |
|
546 | } else { | |
558 | len = bser.wpos - (sizeof(EMPTY_HEADER_V2) - 1); |
|
547 | len = bser.wpos - (sizeof(EMPTY_HEADER_V2) - 1); | |
559 | // The BSER capabilities block comes before the PDU length |
|
548 | // The BSER capabilities block comes before the PDU length | |
560 | memcpy(bser.buf + 2, &bser_capabilities, sizeof(bser_capabilities)); |
|
549 | memcpy(bser.buf + 2, &bser_capabilities, sizeof(bser_capabilities)); | |
561 | memcpy(bser.buf + 7, &len, sizeof(len)); |
|
550 | memcpy(bser.buf + 7, &len, sizeof(len)); | |
562 | } |
|
551 | } | |
563 |
|
552 | |||
564 | res = PyBytes_FromStringAndSize(bser.buf, bser.wpos); |
|
553 | res = PyBytes_FromStringAndSize(bser.buf, bser.wpos); | |
565 | bser_dtor(&bser); |
|
554 | bser_dtor(&bser); | |
566 |
|
555 | |||
567 | return res; |
|
556 | return res; | |
568 | } |
|
557 | } | |
569 |
|
558 | |||
570 | int bunser_int(const char** ptr, const char* end, int64_t* val) { |
|
559 | int bunser_int(const char** ptr, const char* end, int64_t* val) { | |
571 | int needed; |
|
560 | int needed; | |
572 | const char* buf = *ptr; |
|
561 | const char* buf = *ptr; | |
573 | int8_t i8; |
|
562 | int8_t i8; | |
574 | int16_t i16; |
|
563 | int16_t i16; | |
575 | int32_t i32; |
|
564 | int32_t i32; | |
576 | int64_t i64; |
|
565 | int64_t i64; | |
577 |
|
566 | |||
578 | switch (buf[0]) { |
|
567 | switch (buf[0]) { | |
579 | case BSER_INT8: |
|
568 | case BSER_INT8: | |
580 | needed = 2; |
|
569 | needed = 2; | |
581 | break; |
|
570 | break; | |
582 | case BSER_INT16: |
|
571 | case BSER_INT16: | |
583 | needed = 3; |
|
572 | needed = 3; | |
584 | break; |
|
573 | break; | |
585 | case BSER_INT32: |
|
574 | case BSER_INT32: | |
586 | needed = 5; |
|
575 | needed = 5; | |
587 | break; |
|
576 | break; | |
588 | case BSER_INT64: |
|
577 | case BSER_INT64: | |
589 | needed = 9; |
|
578 | needed = 9; | |
590 | break; |
|
579 | break; | |
591 | default: |
|
580 | default: | |
592 | PyErr_Format( |
|
581 | PyErr_Format( | |
593 | PyExc_ValueError, "invalid bser int encoding 0x%02x", buf[0]); |
|
582 | PyExc_ValueError, "invalid bser int encoding 0x%02x", buf[0]); | |
594 | return 0; |
|
583 | return 0; | |
595 | } |
|
584 | } | |
596 | if (end - buf < needed) { |
|
585 | if (end - buf < needed) { | |
597 | PyErr_SetString(PyExc_ValueError, "input buffer to small for int encoding"); |
|
586 | PyErr_SetString(PyExc_ValueError, "input buffer to small for int encoding"); | |
598 | return 0; |
|
587 | return 0; | |
599 | } |
|
588 | } | |
600 | *ptr = buf + needed; |
|
589 | *ptr = buf + needed; | |
601 | switch (buf[0]) { |
|
590 | switch (buf[0]) { | |
602 | case BSER_INT8: |
|
591 | case BSER_INT8: | |
603 | memcpy(&i8, buf + 1, sizeof(i8)); |
|
592 | memcpy(&i8, buf + 1, sizeof(i8)); | |
604 | *val = i8; |
|
593 | *val = i8; | |
605 | return 1; |
|
594 | return 1; | |
606 | case BSER_INT16: |
|
595 | case BSER_INT16: | |
607 | memcpy(&i16, buf + 1, sizeof(i16)); |
|
596 | memcpy(&i16, buf + 1, sizeof(i16)); | |
608 | *val = i16; |
|
597 | *val = i16; | |
609 | return 1; |
|
598 | return 1; | |
610 | case BSER_INT32: |
|
599 | case BSER_INT32: | |
611 | memcpy(&i32, buf + 1, sizeof(i32)); |
|
600 | memcpy(&i32, buf + 1, sizeof(i32)); | |
612 | *val = i32; |
|
601 | *val = i32; | |
613 | return 1; |
|
602 | return 1; | |
614 | case BSER_INT64: |
|
603 | case BSER_INT64: | |
615 | memcpy(&i64, buf + 1, sizeof(i64)); |
|
604 | memcpy(&i64, buf + 1, sizeof(i64)); | |
616 | *val = i64; |
|
605 | *val = i64; | |
617 | return 1; |
|
606 | return 1; | |
618 | default: |
|
607 | default: | |
619 | return 0; |
|
608 | return 0; | |
620 | } |
|
609 | } | |
621 | } |
|
610 | } | |
622 |
|
611 | |||
623 | static int bunser_bytestring( |
|
612 | static int bunser_bytestring( | |
624 | const char** ptr, |
|
613 | const char** ptr, | |
625 | const char* end, |
|
614 | const char* end, | |
626 | const char** start, |
|
615 | const char** start, | |
627 | int64_t* len) { |
|
616 | int64_t* len) { | |
628 | const char* buf = *ptr; |
|
617 | const char* buf = *ptr; | |
629 |
|
618 | |||
630 | // skip string marker |
|
619 | // skip string marker | |
631 | buf++; |
|
620 | buf++; | |
632 | if (!bunser_int(&buf, end, len)) { |
|
621 | if (!bunser_int(&buf, end, len)) { | |
633 | return 0; |
|
622 | return 0; | |
634 | } |
|
623 | } | |
635 |
|
624 | |||
636 | if (buf + *len > end) { |
|
625 | if (buf + *len > end) { | |
637 | PyErr_Format(PyExc_ValueError, "invalid string length in bser data"); |
|
626 | PyErr_Format(PyExc_ValueError, "invalid string length in bser data"); | |
638 | return 0; |
|
627 | return 0; | |
639 | } |
|
628 | } | |
640 |
|
629 | |||
641 | *ptr = buf + *len; |
|
630 | *ptr = buf + *len; | |
642 | *start = buf; |
|
631 | *start = buf; | |
643 | return 1; |
|
632 | return 1; | |
644 | } |
|
633 | } | |
645 |
|
634 | |||
646 | static PyObject* |
|
635 | static PyObject* | |
647 | bunser_array(const char** ptr, const char* end, const unser_ctx_t* ctx) { |
|
636 | bunser_array(const char** ptr, const char* end, const unser_ctx_t* ctx) { | |
648 | const char* buf = *ptr; |
|
637 | const char* buf = *ptr; | |
649 | int64_t nitems, i; |
|
638 | int64_t nitems, i; | |
650 | int mutable = ctx->mutable; |
|
639 | int mutable = ctx->mutable; | |
651 | PyObject* res; |
|
640 | PyObject* res; | |
652 |
|
641 | |||
653 | // skip array header |
|
642 | // skip array header | |
654 | buf++; |
|
643 | buf++; | |
655 | if (!bunser_int(&buf, end, &nitems)) { |
|
644 | if (!bunser_int(&buf, end, &nitems)) { | |
656 | return 0; |
|
645 | return 0; | |
657 | } |
|
646 | } | |
658 | *ptr = buf; |
|
647 | *ptr = buf; | |
659 |
|
648 | |||
660 | if (nitems > LONG_MAX) { |
|
649 | if (nitems > LONG_MAX) { | |
661 | PyErr_Format(PyExc_ValueError, "too many items for python array"); |
|
650 | PyErr_Format(PyExc_ValueError, "too many items for python array"); | |
662 | return NULL; |
|
651 | return NULL; | |
663 | } |
|
652 | } | |
664 |
|
653 | |||
665 | if (mutable) { |
|
654 | if (mutable) { | |
666 | res = PyList_New((Py_ssize_t)nitems); |
|
655 | res = PyList_New((Py_ssize_t)nitems); | |
667 | } else { |
|
656 | } else { | |
668 | res = PyTuple_New((Py_ssize_t)nitems); |
|
657 | res = PyTuple_New((Py_ssize_t)nitems); | |
669 | } |
|
658 | } | |
670 |
|
659 | |||
671 | for (i = 0; i < nitems; i++) { |
|
660 | for (i = 0; i < nitems; i++) { | |
672 | PyObject* ele = bser_loads_recursive(ptr, end, ctx); |
|
661 | PyObject* ele = bser_loads_recursive(ptr, end, ctx); | |
673 |
|
662 | |||
674 | if (!ele) { |
|
663 | if (!ele) { | |
675 | Py_DECREF(res); |
|
664 | Py_DECREF(res); | |
676 | return NULL; |
|
665 | return NULL; | |
677 | } |
|
666 | } | |
678 |
|
667 | |||
679 | if (mutable) { |
|
668 | if (mutable) { | |
680 | PyList_SET_ITEM(res, i, ele); |
|
669 | PyList_SET_ITEM(res, i, ele); | |
681 | } else { |
|
670 | } else { | |
682 | PyTuple_SET_ITEM(res, i, ele); |
|
671 | PyTuple_SET_ITEM(res, i, ele); | |
683 | } |
|
672 | } | |
684 | // DECREF(ele) not required as SET_ITEM steals the ref |
|
673 | // DECREF(ele) not required as SET_ITEM steals the ref | |
685 | } |
|
674 | } | |
686 |
|
675 | |||
687 | return res; |
|
676 | return res; | |
688 | } |
|
677 | } | |
689 |
|
678 | |||
690 | static PyObject* |
|
679 | static PyObject* | |
691 | bunser_object(const char** ptr, const char* end, const unser_ctx_t* ctx) { |
|
680 | bunser_object(const char** ptr, const char* end, const unser_ctx_t* ctx) { | |
692 | const char* buf = *ptr; |
|
681 | const char* buf = *ptr; | |
693 | int64_t nitems, i; |
|
682 | int64_t nitems, i; | |
694 | int mutable = ctx->mutable; |
|
683 | int mutable = ctx->mutable; | |
695 | PyObject* res; |
|
684 | PyObject* res; | |
696 | bserObject* obj; |
|
685 | bserObject* obj; | |
697 |
|
686 | |||
698 | // skip array header |
|
687 | // skip array header | |
699 | buf++; |
|
688 | buf++; | |
700 | if (!bunser_int(&buf, end, &nitems)) { |
|
689 | if (!bunser_int(&buf, end, &nitems)) { | |
701 | return 0; |
|
690 | return 0; | |
702 | } |
|
691 | } | |
703 | *ptr = buf; |
|
692 | *ptr = buf; | |
704 |
|
693 | |||
705 | if (mutable) { |
|
694 | if (mutable) { | |
706 | res = PyDict_New(); |
|
695 | res = PyDict_New(); | |
707 | } else { |
|
696 | } else { | |
708 | obj = PyObject_New(bserObject, &bserObjectType); |
|
697 | obj = PyObject_New(bserObject, &bserObjectType); | |
709 | obj->keys = PyTuple_New((Py_ssize_t)nitems); |
|
698 | obj->keys = PyTuple_New((Py_ssize_t)nitems); | |
710 | obj->values = PyTuple_New((Py_ssize_t)nitems); |
|
699 | obj->values = PyTuple_New((Py_ssize_t)nitems); | |
711 | res = (PyObject*)obj; |
|
700 | res = (PyObject*)obj; | |
712 | } |
|
701 | } | |
713 |
|
702 | |||
714 | for (i = 0; i < nitems; i++) { |
|
703 | for (i = 0; i < nitems; i++) { | |
715 | const char* keystr; |
|
704 | const char* keystr; | |
716 | int64_t keylen; |
|
705 | int64_t keylen; | |
717 | PyObject* key; |
|
706 | PyObject* key; | |
718 | PyObject* ele; |
|
707 | PyObject* ele; | |
719 |
|
708 | |||
720 | if (!bunser_bytestring(ptr, end, &keystr, &keylen)) { |
|
709 | if (!bunser_bytestring(ptr, end, &keystr, &keylen)) { | |
721 | Py_DECREF(res); |
|
710 | Py_DECREF(res); | |
722 | return NULL; |
|
711 | return NULL; | |
723 | } |
|
712 | } | |
724 |
|
713 | |||
725 | if (keylen > LONG_MAX) { |
|
714 | if (keylen > LONG_MAX) { | |
726 | PyErr_Format(PyExc_ValueError, "string too big for python"); |
|
715 | PyErr_Format(PyExc_ValueError, "string too big for python"); | |
727 | Py_DECREF(res); |
|
716 | Py_DECREF(res); | |
728 | return NULL; |
|
717 | return NULL; | |
729 | } |
|
718 | } | |
730 |
|
719 | |||
731 | if (mutable) { |
|
720 | if (mutable) { | |
732 | // This will interpret the key as UTF-8. |
|
721 | // This will interpret the key as UTF-8. | |
733 | key = PyUnicode_FromStringAndSize(keystr, (Py_ssize_t)keylen); |
|
722 | key = PyUnicode_FromStringAndSize(keystr, (Py_ssize_t)keylen); | |
734 | } else { |
|
723 | } else { | |
735 | // For immutable objects we'll manage key lookups, so we can avoid going |
|
724 | // For immutable objects we'll manage key lookups, so we can avoid going | |
736 | // through the Unicode APIs. This avoids a potentially expensive and |
|
725 | // through the Unicode APIs. This avoids a potentially expensive and | |
737 | // definitely unnecessary conversion to UTF-16 and back for Python 2. |
|
726 | // definitely unnecessary conversion to UTF-16 and back for Python 2. | |
738 | // TODO: On Python 3 the Unicode APIs are smarter: we might be able to use |
|
727 | // TODO: On Python 3 the Unicode APIs are smarter: we might be able to use | |
739 | // Unicode keys there without an appreciable performance loss. |
|
728 | // Unicode keys there without an appreciable performance loss. | |
740 | key = PyBytes_FromStringAndSize(keystr, (Py_ssize_t)keylen); |
|
729 | key = PyBytes_FromStringAndSize(keystr, (Py_ssize_t)keylen); | |
741 | } |
|
730 | } | |
742 |
|
731 | |||
743 | if (!key) { |
|
732 | if (!key) { | |
744 | Py_DECREF(res); |
|
733 | Py_DECREF(res); | |
745 | return NULL; |
|
734 | return NULL; | |
746 | } |
|
735 | } | |
747 |
|
736 | |||
748 | ele = bser_loads_recursive(ptr, end, ctx); |
|
737 | ele = bser_loads_recursive(ptr, end, ctx); | |
749 |
|
738 | |||
750 | if (!ele) { |
|
739 | if (!ele) { | |
751 | Py_DECREF(key); |
|
740 | Py_DECREF(key); | |
752 | Py_DECREF(res); |
|
741 | Py_DECREF(res); | |
753 | return NULL; |
|
742 | return NULL; | |
754 | } |
|
743 | } | |
755 |
|
744 | |||
756 | if (mutable) { |
|
745 | if (mutable) { | |
757 | PyDict_SetItem(res, key, ele); |
|
746 | PyDict_SetItem(res, key, ele); | |
758 | Py_DECREF(key); |
|
747 | Py_DECREF(key); | |
759 | Py_DECREF(ele); |
|
748 | Py_DECREF(ele); | |
760 | } else { |
|
749 | } else { | |
761 | /* PyTuple_SET_ITEM steals ele, key */ |
|
750 | /* PyTuple_SET_ITEM steals ele, key */ | |
762 | PyTuple_SET_ITEM(obj->values, i, ele); |
|
751 | PyTuple_SET_ITEM(obj->values, i, ele); | |
763 | PyTuple_SET_ITEM(obj->keys, i, key); |
|
752 | PyTuple_SET_ITEM(obj->keys, i, key); | |
764 | } |
|
753 | } | |
765 | } |
|
754 | } | |
766 |
|
755 | |||
767 | return res; |
|
756 | return res; | |
768 | } |
|
757 | } | |
769 |
|
758 | |||
770 | static PyObject* |
|
759 | static PyObject* | |
771 | bunser_template(const char** ptr, const char* end, const unser_ctx_t* ctx) { |
|
760 | bunser_template(const char** ptr, const char* end, const unser_ctx_t* ctx) { | |
772 | const char* buf = *ptr; |
|
761 | const char* buf = *ptr; | |
773 | int64_t nitems, i; |
|
762 | int64_t nitems, i; | |
774 | int mutable = ctx->mutable; |
|
763 | int mutable = ctx->mutable; | |
775 | PyObject* arrval; |
|
764 | PyObject* arrval; | |
776 | PyObject* keys; |
|
765 | PyObject* keys; | |
777 | Py_ssize_t numkeys, keyidx; |
|
766 | Py_ssize_t numkeys, keyidx; | |
778 | unser_ctx_t keys_ctx = {0}; |
|
767 | unser_ctx_t keys_ctx = {0}; | |
779 | if (mutable) { |
|
768 | if (mutable) { | |
780 | keys_ctx.mutable = 1; |
|
769 | keys_ctx.mutable = 1; | |
781 | // Decode keys as UTF-8 in this case. |
|
770 | // Decode keys as UTF-8 in this case. | |
782 | keys_ctx.value_encoding = "utf-8"; |
|
771 | keys_ctx.value_encoding = "utf-8"; | |
783 | keys_ctx.value_errors = "strict"; |
|
772 | keys_ctx.value_errors = "strict"; | |
784 | } else { |
|
773 | } else { | |
785 | // Treat keys as bytestrings in this case -- we'll do Unicode conversions at |
|
774 | // Treat keys as bytestrings in this case -- we'll do Unicode conversions at | |
786 | // lookup time. |
|
775 | // lookup time. | |
787 | } |
|
776 | } | |
788 |
|
777 | |||
789 | if (buf[1] != BSER_ARRAY) { |
|
778 | if (buf[1] != BSER_ARRAY) { | |
790 | PyErr_Format(PyExc_ValueError, "Expect ARRAY to follow TEMPLATE"); |
|
779 | PyErr_Format(PyExc_ValueError, "Expect ARRAY to follow TEMPLATE"); | |
791 | return NULL; |
|
780 | return NULL; | |
792 | } |
|
781 | } | |
793 |
|
782 | |||
794 | // skip header |
|
783 | // skip header | |
795 | buf++; |
|
784 | buf++; | |
796 | *ptr = buf; |
|
785 | *ptr = buf; | |
797 |
|
786 | |||
798 | // Load template keys. |
|
787 | // Load template keys. | |
799 | // For keys we don't want to do any decoding right now. |
|
788 | // For keys we don't want to do any decoding right now. | |
800 | keys = bunser_array(ptr, end, &keys_ctx); |
|
789 | keys = bunser_array(ptr, end, &keys_ctx); | |
801 | if (!keys) { |
|
790 | if (!keys) { | |
802 | return NULL; |
|
791 | return NULL; | |
803 | } |
|
792 | } | |
804 |
|
793 | |||
805 | numkeys = PySequence_Length(keys); |
|
794 | numkeys = PySequence_Length(keys); | |
806 |
|
795 | |||
807 | // Load number of array elements |
|
796 | // Load number of array elements | |
808 | if (!bunser_int(ptr, end, &nitems)) { |
|
797 | if (!bunser_int(ptr, end, &nitems)) { | |
809 | Py_DECREF(keys); |
|
798 | Py_DECREF(keys); | |
810 | return 0; |
|
799 | return 0; | |
811 | } |
|
800 | } | |
812 |
|
801 | |||
813 | if (nitems > LONG_MAX) { |
|
802 | if (nitems > LONG_MAX) { | |
814 | PyErr_Format(PyExc_ValueError, "Too many items for python"); |
|
803 | PyErr_Format(PyExc_ValueError, "Too many items for python"); | |
815 | Py_DECREF(keys); |
|
804 | Py_DECREF(keys); | |
816 | return NULL; |
|
805 | return NULL; | |
817 | } |
|
806 | } | |
818 |
|
807 | |||
819 | arrval = PyList_New((Py_ssize_t)nitems); |
|
808 | arrval = PyList_New((Py_ssize_t)nitems); | |
820 | if (!arrval) { |
|
809 | if (!arrval) { | |
821 | Py_DECREF(keys); |
|
810 | Py_DECREF(keys); | |
822 | return NULL; |
|
811 | return NULL; | |
823 | } |
|
812 | } | |
824 |
|
813 | |||
825 | for (i = 0; i < nitems; i++) { |
|
814 | for (i = 0; i < nitems; i++) { | |
826 | PyObject* dict = NULL; |
|
815 | PyObject* dict = NULL; | |
827 | bserObject* obj = NULL; |
|
816 | bserObject* obj = NULL; | |
828 |
|
817 | |||
829 | if (mutable) { |
|
818 | if (mutable) { | |
830 | dict = PyDict_New(); |
|
819 | dict = PyDict_New(); | |
831 | } else { |
|
820 | } else { | |
832 | obj = PyObject_New(bserObject, &bserObjectType); |
|
821 | obj = PyObject_New(bserObject, &bserObjectType); | |
833 | if (obj) { |
|
822 | if (obj) { | |
834 | obj->keys = keys; |
|
823 | obj->keys = keys; | |
835 | Py_INCREF(obj->keys); |
|
824 | Py_INCREF(obj->keys); | |
836 | obj->values = PyTuple_New(numkeys); |
|
825 | obj->values = PyTuple_New(numkeys); | |
837 | } |
|
826 | } | |
838 | dict = (PyObject*)obj; |
|
827 | dict = (PyObject*)obj; | |
839 | } |
|
828 | } | |
840 | if (!dict) { |
|
829 | if (!dict) { | |
841 | fail: |
|
830 | fail: | |
842 | Py_DECREF(keys); |
|
831 | Py_DECREF(keys); | |
843 | Py_DECREF(arrval); |
|
832 | Py_DECREF(arrval); | |
844 | return NULL; |
|
833 | return NULL; | |
845 | } |
|
834 | } | |
846 |
|
835 | |||
847 | for (keyidx = 0; keyidx < numkeys; keyidx++) { |
|
836 | for (keyidx = 0; keyidx < numkeys; keyidx++) { | |
848 | PyObject* key; |
|
837 | PyObject* key; | |
849 | PyObject* ele; |
|
838 | PyObject* ele; | |
850 |
|
839 | |||
851 | if (**ptr == BSER_SKIP) { |
|
840 | if (**ptr == BSER_SKIP) { | |
852 | *ptr = *ptr + 1; |
|
841 | *ptr = *ptr + 1; | |
853 | ele = Py_None; |
|
842 | ele = Py_None; | |
854 | Py_INCREF(ele); |
|
843 | Py_INCREF(ele); | |
855 | } else { |
|
844 | } else { | |
856 | ele = bser_loads_recursive(ptr, end, ctx); |
|
845 | ele = bser_loads_recursive(ptr, end, ctx); | |
857 | } |
|
846 | } | |
858 |
|
847 | |||
859 | if (!ele) { |
|
848 | if (!ele) { | |
860 | goto fail; |
|
849 | goto fail; | |
861 | } |
|
850 | } | |
862 |
|
851 | |||
863 | if (mutable) { |
|
852 | if (mutable) { | |
864 | key = PyList_GET_ITEM(keys, keyidx); |
|
853 | key = PyList_GET_ITEM(keys, keyidx); | |
865 | PyDict_SetItem(dict, key, ele); |
|
854 | PyDict_SetItem(dict, key, ele); | |
866 | Py_DECREF(ele); |
|
855 | Py_DECREF(ele); | |
867 | } else { |
|
856 | } else { | |
868 | PyTuple_SET_ITEM(obj->values, keyidx, ele); |
|
857 | PyTuple_SET_ITEM(obj->values, keyidx, ele); | |
869 | // DECREF(ele) not required as SET_ITEM steals the ref |
|
858 | // DECREF(ele) not required as SET_ITEM steals the ref | |
870 | } |
|
859 | } | |
871 | } |
|
860 | } | |
872 |
|
861 | |||
873 | PyList_SET_ITEM(arrval, i, dict); |
|
862 | PyList_SET_ITEM(arrval, i, dict); | |
874 | // DECREF(obj) not required as SET_ITEM steals the ref |
|
863 | // DECREF(obj) not required as SET_ITEM steals the ref | |
875 | } |
|
864 | } | |
876 |
|
865 | |||
877 | Py_DECREF(keys); |
|
866 | Py_DECREF(keys); | |
878 |
|
867 | |||
879 | return arrval; |
|
868 | return arrval; | |
880 | } |
|
869 | } | |
881 |
|
870 | |||
882 | static PyObject* bser_loads_recursive( |
|
871 | static PyObject* bser_loads_recursive( | |
883 | const char** ptr, |
|
872 | const char** ptr, | |
884 | const char* end, |
|
873 | const char* end, | |
885 | const unser_ctx_t* ctx) { |
|
874 | const unser_ctx_t* ctx) { | |
886 | const char* buf = *ptr; |
|
875 | const char* buf = *ptr; | |
887 |
|
876 | |||
888 | switch (buf[0]) { |
|
877 | switch (buf[0]) { | |
889 | case BSER_INT8: |
|
878 | case BSER_INT8: | |
890 | case BSER_INT16: |
|
879 | case BSER_INT16: | |
891 | case BSER_INT32: |
|
880 | case BSER_INT32: | |
892 | case BSER_INT64: { |
|
881 | case BSER_INT64: { | |
893 | int64_t ival; |
|
882 | int64_t ival; | |
894 | if (!bunser_int(ptr, end, &ival)) { |
|
883 | if (!bunser_int(ptr, end, &ival)) { | |
895 | return NULL; |
|
884 | return NULL; | |
896 | } |
|
885 | } | |
897 | // Python 3 has one integer type. |
|
886 | // Python 3 has one integer type. | |
898 | #if PY_MAJOR_VERSION >= 3 |
|
887 | #if PY_MAJOR_VERSION >= 3 | |
899 | return PyLong_FromLongLong(ival); |
|
888 | return PyLong_FromLongLong(ival); | |
900 | #else |
|
889 | #else | |
901 | if (ival < LONG_MIN || ival > LONG_MAX) { |
|
890 | if (ival < LONG_MIN || ival > LONG_MAX) { | |
902 | return PyLong_FromLongLong(ival); |
|
891 | return PyLong_FromLongLong(ival); | |
903 | } |
|
892 | } | |
904 | return PyInt_FromSsize_t(Py_SAFE_DOWNCAST(ival, int64_t, Py_ssize_t)); |
|
893 | return PyInt_FromSsize_t(Py_SAFE_DOWNCAST(ival, int64_t, Py_ssize_t)); | |
905 | #endif // PY_MAJOR_VERSION >= 3 |
|
894 | #endif // PY_MAJOR_VERSION >= 3 | |
906 | } |
|
895 | } | |
907 |
|
896 | |||
908 | case BSER_REAL: { |
|
897 | case BSER_REAL: { | |
909 | double dval; |
|
898 | double dval; | |
910 | memcpy(&dval, buf + 1, sizeof(dval)); |
|
899 | memcpy(&dval, buf + 1, sizeof(dval)); | |
911 | *ptr = buf + 1 + sizeof(double); |
|
900 | *ptr = buf + 1 + sizeof(double); | |
912 | return PyFloat_FromDouble(dval); |
|
901 | return PyFloat_FromDouble(dval); | |
913 | } |
|
902 | } | |
914 |
|
903 | |||
915 | case BSER_TRUE: |
|
904 | case BSER_TRUE: | |
916 | *ptr = buf + 1; |
|
905 | *ptr = buf + 1; | |
917 | Py_INCREF(Py_True); |
|
906 | Py_INCREF(Py_True); | |
918 | return Py_True; |
|
907 | return Py_True; | |
919 |
|
908 | |||
920 | case BSER_FALSE: |
|
909 | case BSER_FALSE: | |
921 | *ptr = buf + 1; |
|
910 | *ptr = buf + 1; | |
922 | Py_INCREF(Py_False); |
|
911 | Py_INCREF(Py_False); | |
923 | return Py_False; |
|
912 | return Py_False; | |
924 |
|
913 | |||
925 | case BSER_NULL: |
|
914 | case BSER_NULL: | |
926 | *ptr = buf + 1; |
|
915 | *ptr = buf + 1; | |
927 | Py_INCREF(Py_None); |
|
916 | Py_INCREF(Py_None); | |
928 | return Py_None; |
|
917 | return Py_None; | |
929 |
|
918 | |||
930 | case BSER_BYTESTRING: { |
|
919 | case BSER_BYTESTRING: { | |
931 | const char* start; |
|
920 | const char* start; | |
932 | int64_t len; |
|
921 | int64_t len; | |
933 |
|
922 | |||
934 | if (!bunser_bytestring(ptr, end, &start, &len)) { |
|
923 | if (!bunser_bytestring(ptr, end, &start, &len)) { | |
935 | return NULL; |
|
924 | return NULL; | |
936 | } |
|
925 | } | |
937 |
|
926 | |||
938 | if (len > LONG_MAX) { |
|
927 | if (len > LONG_MAX) { | |
939 | PyErr_Format(PyExc_ValueError, "string too long for python"); |
|
928 | PyErr_Format(PyExc_ValueError, "string too long for python"); | |
940 | return NULL; |
|
929 | return NULL; | |
941 | } |
|
930 | } | |
942 |
|
931 | |||
943 | if (ctx->value_encoding != NULL) { |
|
932 | if (ctx->value_encoding != NULL) { | |
944 | return PyUnicode_Decode( |
|
933 | return PyUnicode_Decode( | |
945 | start, (long)len, ctx->value_encoding, ctx->value_errors); |
|
934 | start, (long)len, ctx->value_encoding, ctx->value_errors); | |
946 | } else { |
|
935 | } else { | |
947 | return PyBytes_FromStringAndSize(start, (long)len); |
|
936 | return PyBytes_FromStringAndSize(start, (long)len); | |
948 | } |
|
937 | } | |
949 | } |
|
938 | } | |
950 |
|
939 | |||
951 | case BSER_UTF8STRING: { |
|
940 | case BSER_UTF8STRING: { | |
952 | const char* start; |
|
941 | const char* start; | |
953 | int64_t len; |
|
942 | int64_t len; | |
954 |
|
943 | |||
955 | if (!bunser_bytestring(ptr, end, &start, &len)) { |
|
944 | if (!bunser_bytestring(ptr, end, &start, &len)) { | |
956 | return NULL; |
|
945 | return NULL; | |
957 | } |
|
946 | } | |
958 |
|
947 | |||
959 | if (len > LONG_MAX) { |
|
948 | if (len > LONG_MAX) { | |
960 | PyErr_Format(PyExc_ValueError, "string too long for python"); |
|
949 | PyErr_Format(PyExc_ValueError, "string too long for python"); | |
961 | return NULL; |
|
950 | return NULL; | |
962 | } |
|
951 | } | |
963 |
|
952 | |||
964 | return PyUnicode_Decode(start, (long)len, "utf-8", "strict"); |
|
953 | return PyUnicode_Decode(start, (long)len, "utf-8", "strict"); | |
965 | } |
|
954 | } | |
966 |
|
955 | |||
967 | case BSER_ARRAY: |
|
956 | case BSER_ARRAY: | |
968 | return bunser_array(ptr, end, ctx); |
|
957 | return bunser_array(ptr, end, ctx); | |
969 |
|
958 | |||
970 | case BSER_OBJECT: |
|
959 | case BSER_OBJECT: | |
971 | return bunser_object(ptr, end, ctx); |
|
960 | return bunser_object(ptr, end, ctx); | |
972 |
|
961 | |||
973 | case BSER_TEMPLATE: |
|
962 | case BSER_TEMPLATE: | |
974 | return bunser_template(ptr, end, ctx); |
|
963 | return bunser_template(ptr, end, ctx); | |
975 |
|
964 | |||
976 | default: |
|
965 | default: | |
977 | PyErr_Format(PyExc_ValueError, "unhandled bser opcode 0x%02x", buf[0]); |
|
966 | PyErr_Format(PyExc_ValueError, "unhandled bser opcode 0x%02x", buf[0]); | |
978 | } |
|
967 | } | |
979 |
|
968 | |||
980 | return NULL; |
|
969 | return NULL; | |
981 | } |
|
970 | } | |
982 |
|
971 | |||
983 | static int _pdu_info_helper( |
|
972 | static int _pdu_info_helper( | |
984 | const char* data, |
|
973 | const char* data, | |
985 | const char* end, |
|
974 | const char* end, | |
986 | uint32_t* bser_version_out, |
|
975 | uint32_t* bser_version_out, | |
987 | uint32_t* bser_capabilities_out, |
|
976 | uint32_t* bser_capabilities_out, | |
988 | int64_t* expected_len_out, |
|
977 | int64_t* expected_len_out, | |
989 | off_t* position_out) { |
|
978 | off_t* position_out) { | |
990 | uint32_t bser_version; |
|
979 | uint32_t bser_version; | |
991 | uint32_t bser_capabilities = 0; |
|
980 | uint32_t bser_capabilities = 0; | |
992 | int64_t expected_len; |
|
981 | int64_t expected_len; | |
993 |
|
982 | |||
994 | const char* start; |
|
983 | const char* start; | |
995 | start = data; |
|
984 | start = data; | |
996 | // Validate the header and length |
|
985 | // Validate the header and length | |
997 | if (memcmp(data, EMPTY_HEADER, 2) == 0) { |
|
986 | if (memcmp(data, EMPTY_HEADER, 2) == 0) { | |
998 | bser_version = 1; |
|
987 | bser_version = 1; | |
999 | } else if (memcmp(data, EMPTY_HEADER_V2, 2) == 0) { |
|
988 | } else if (memcmp(data, EMPTY_HEADER_V2, 2) == 0) { | |
1000 | bser_version = 2; |
|
989 | bser_version = 2; | |
1001 | } else { |
|
990 | } else { | |
1002 | PyErr_SetString(PyExc_ValueError, "invalid bser header"); |
|
991 | PyErr_SetString(PyExc_ValueError, "invalid bser header"); | |
1003 | return 0; |
|
992 | return 0; | |
1004 | } |
|
993 | } | |
1005 |
|
994 | |||
1006 | data += 2; |
|
995 | data += 2; | |
1007 |
|
996 | |||
1008 | if (bser_version == 2) { |
|
997 | if (bser_version == 2) { | |
1009 | // Expect an integer telling us what capabilities are supported by the |
|
998 | // Expect an integer telling us what capabilities are supported by the | |
1010 | // remote server (currently unused). |
|
999 | // remote server (currently unused). | |
1011 | if (!memcpy(&bser_capabilities, &data, sizeof(bser_capabilities))) { |
|
1000 | if (!memcpy(&bser_capabilities, &data, sizeof(bser_capabilities))) { | |
1012 | return 0; |
|
1001 | return 0; | |
1013 | } |
|
1002 | } | |
1014 | data += sizeof(bser_capabilities); |
|
1003 | data += sizeof(bser_capabilities); | |
1015 | } |
|
1004 | } | |
1016 |
|
1005 | |||
1017 | // Expect an integer telling us how big the rest of the data |
|
1006 | // Expect an integer telling us how big the rest of the data | |
1018 | // should be |
|
1007 | // should be | |
1019 | if (!bunser_int(&data, end, &expected_len)) { |
|
1008 | if (!bunser_int(&data, end, &expected_len)) { | |
1020 | return 0; |
|
1009 | return 0; | |
1021 | } |
|
1010 | } | |
1022 |
|
1011 | |||
1023 | *bser_version_out = bser_version; |
|
1012 | *bser_version_out = bser_version; | |
1024 | *bser_capabilities_out = (uint32_t)bser_capabilities; |
|
1013 | *bser_capabilities_out = (uint32_t)bser_capabilities; | |
1025 | *expected_len_out = expected_len; |
|
1014 | *expected_len_out = expected_len; | |
1026 | *position_out = (off_t)(data - start); |
|
1015 | *position_out = (off_t)(data - start); | |
1027 | return 1; |
|
1016 | return 1; | |
1028 | } |
|
1017 | } | |
1029 |
|
1018 | |||
1030 | // This function parses the PDU header and provides info about the packet |
|
1019 | // This function parses the PDU header and provides info about the packet | |
1031 | // Returns false if unsuccessful |
|
1020 | // Returns false if unsuccessful | |
1032 | static int pdu_info_helper( |
|
1021 | static int pdu_info_helper( | |
1033 | PyObject* self, |
|
1022 | PyObject* self, | |
1034 | PyObject* args, |
|
1023 | PyObject* args, | |
1035 | uint32_t* bser_version_out, |
|
1024 | uint32_t* bser_version_out, | |
1036 | uint32_t* bser_capabilities_out, |
|
1025 | uint32_t* bser_capabilities_out, | |
1037 | int64_t* total_len_out) { |
|
1026 | int64_t* total_len_out) { | |
1038 | const char* start = NULL; |
|
1027 | const char* start = NULL; | |
1039 | const char* data = NULL; |
|
1028 | const char* data = NULL; | |
1040 | int datalen = 0; |
|
1029 | int datalen = 0; | |
1041 | const char* end; |
|
1030 | const char* end; | |
1042 | int64_t expected_len; |
|
1031 | int64_t expected_len; | |
1043 | off_t position; |
|
1032 | off_t position; | |
1044 |
|
1033 | |||
1045 | if (!PyArg_ParseTuple(args, "s#", &start, &datalen)) { |
|
1034 | if (!PyArg_ParseTuple(args, "s#", &start, &datalen)) { | |
1046 | return 0; |
|
1035 | return 0; | |
1047 | } |
|
1036 | } | |
1048 | data = start; |
|
1037 | data = start; | |
1049 | end = data + datalen; |
|
1038 | end = data + datalen; | |
1050 |
|
1039 | |||
1051 | if (!_pdu_info_helper( |
|
1040 | if (!_pdu_info_helper( | |
1052 | data, |
|
1041 | data, | |
1053 | end, |
|
1042 | end, | |
1054 | bser_version_out, |
|
1043 | bser_version_out, | |
1055 | bser_capabilities_out, |
|
1044 | bser_capabilities_out, | |
1056 | &expected_len, |
|
1045 | &expected_len, | |
1057 | &position)) { |
|
1046 | &position)) { | |
1058 | return 0; |
|
1047 | return 0; | |
1059 | } |
|
1048 | } | |
1060 | *total_len_out = (int64_t)(expected_len + position); |
|
1049 | *total_len_out = (int64_t)(expected_len + position); | |
1061 | return 1; |
|
1050 | return 1; | |
1062 | } |
|
1051 | } | |
1063 |
|
1052 | |||
1064 | // Expected use case is to read a packet from the socket and then call |
|
1053 | // Expected use case is to read a packet from the socket and then call | |
1065 | // bser.pdu_info on the packet. It returns the BSER version, BSER capabilities, |
|
1054 | // bser.pdu_info on the packet. It returns the BSER version, BSER capabilities, | |
1066 | // and the total length of the entire response that the peer is sending, |
|
1055 | // and the total length of the entire response that the peer is sending, | |
1067 | // including the bytes already received. This allows the client to compute the |
|
1056 | // including the bytes already received. This allows the client to compute the | |
1068 | // data size it needs to read before it can decode the data. |
|
1057 | // data size it needs to read before it can decode the data. | |
1069 | static PyObject* bser_pdu_info(PyObject* self, PyObject* args) { |
|
1058 | static PyObject* bser_pdu_info(PyObject* self, PyObject* args) { | |
1070 | uint32_t version, capabilities; |
|
1059 | uint32_t version, capabilities; | |
1071 | int64_t total_len; |
|
1060 | int64_t total_len; | |
1072 | if (!pdu_info_helper(self, args, &version, &capabilities, &total_len)) { |
|
1061 | if (!pdu_info_helper(self, args, &version, &capabilities, &total_len)) { | |
1073 | return NULL; |
|
1062 | return NULL; | |
1074 | } |
|
1063 | } | |
1075 | return Py_BuildValue("kkL", version, capabilities, total_len); |
|
1064 | return Py_BuildValue("kkL", version, capabilities, total_len); | |
1076 | } |
|
1065 | } | |
1077 |
|
1066 | |||
1078 | static PyObject* bser_pdu_len(PyObject* self, PyObject* args) { |
|
1067 | static PyObject* bser_pdu_len(PyObject* self, PyObject* args) { | |
1079 | uint32_t version, capabilities; |
|
1068 | uint32_t version, capabilities; | |
1080 | int64_t total_len; |
|
1069 | int64_t total_len; | |
1081 | if (!pdu_info_helper(self, args, &version, &capabilities, &total_len)) { |
|
1070 | if (!pdu_info_helper(self, args, &version, &capabilities, &total_len)) { | |
1082 | return NULL; |
|
1071 | return NULL; | |
1083 | } |
|
1072 | } | |
1084 | return Py_BuildValue("L", total_len); |
|
1073 | return Py_BuildValue("L", total_len); | |
1085 | } |
|
1074 | } | |
1086 |
|
1075 | |||
1087 | static PyObject* bser_loads(PyObject* self, PyObject* args, PyObject* kw) { |
|
1076 | static PyObject* bser_loads(PyObject* self, PyObject* args, PyObject* kw) { | |
1088 | const char* data = NULL; |
|
1077 | const char* data = NULL; | |
1089 | int datalen = 0; |
|
1078 | int datalen = 0; | |
1090 | const char* start; |
|
1079 | const char* start; | |
1091 | const char* end; |
|
1080 | const char* end; | |
1092 | int64_t expected_len; |
|
1081 | int64_t expected_len; | |
1093 | off_t position; |
|
1082 | off_t position; | |
1094 | PyObject* mutable_obj = NULL; |
|
1083 | PyObject* mutable_obj = NULL; | |
1095 | const char* value_encoding = NULL; |
|
1084 | const char* value_encoding = NULL; | |
1096 | const char* value_errors = NULL; |
|
1085 | const char* value_errors = NULL; | |
1097 | unser_ctx_t ctx = {1, 0}; |
|
1086 | unser_ctx_t ctx = {1, 0}; | |
1098 |
|
1087 | |||
1099 | static char* kw_list[] = { |
|
1088 | static char* kw_list[] = { | |
1100 | "buf", "mutable", "value_encoding", "value_errors", NULL}; |
|
1089 | "buf", "mutable", "value_encoding", "value_errors", NULL}; | |
1101 |
|
1090 | |||
1102 | if (!PyArg_ParseTupleAndKeywords( |
|
1091 | if (!PyArg_ParseTupleAndKeywords( | |
1103 | args, |
|
1092 | args, | |
1104 | kw, |
|
1093 | kw, | |
1105 | "s#|Ozz:loads", |
|
1094 | "s#|Ozz:loads", | |
1106 | kw_list, |
|
1095 | kw_list, | |
1107 | &start, |
|
1096 | &start, | |
1108 | &datalen, |
|
1097 | &datalen, | |
1109 | &mutable_obj, |
|
1098 | &mutable_obj, | |
1110 | &value_encoding, |
|
1099 | &value_encoding, | |
1111 | &value_errors)) { |
|
1100 | &value_errors)) { | |
1112 | return NULL; |
|
1101 | return NULL; | |
1113 | } |
|
1102 | } | |
1114 |
|
1103 | |||
1115 | if (mutable_obj) { |
|
1104 | if (mutable_obj) { | |
1116 | ctx.mutable = PyObject_IsTrue(mutable_obj) > 0 ? 1 : 0; |
|
1105 | ctx.mutable = PyObject_IsTrue(mutable_obj) > 0 ? 1 : 0; | |
1117 | } |
|
1106 | } | |
1118 | ctx.value_encoding = value_encoding; |
|
1107 | ctx.value_encoding = value_encoding; | |
1119 | if (value_encoding == NULL) { |
|
1108 | if (value_encoding == NULL) { | |
1120 | ctx.value_errors = NULL; |
|
1109 | ctx.value_errors = NULL; | |
1121 | } else if (value_errors == NULL) { |
|
1110 | } else if (value_errors == NULL) { | |
1122 | ctx.value_errors = "strict"; |
|
1111 | ctx.value_errors = "strict"; | |
1123 | } else { |
|
1112 | } else { | |
1124 | ctx.value_errors = value_errors; |
|
1113 | ctx.value_errors = value_errors; | |
1125 | } |
|
1114 | } | |
1126 | data = start; |
|
1115 | data = start; | |
1127 | end = data + datalen; |
|
1116 | end = data + datalen; | |
1128 |
|
1117 | |||
1129 | if (!_pdu_info_helper( |
|
1118 | if (!_pdu_info_helper( | |
1130 | data, |
|
1119 | data, | |
1131 | end, |
|
1120 | end, | |
1132 | &ctx.bser_version, |
|
1121 | &ctx.bser_version, | |
1133 | &ctx.bser_capabilities, |
|
1122 | &ctx.bser_capabilities, | |
1134 | &expected_len, |
|
1123 | &expected_len, | |
1135 | &position)) { |
|
1124 | &position)) { | |
1136 | return NULL; |
|
1125 | return NULL; | |
1137 | } |
|
1126 | } | |
1138 |
|
1127 | |||
1139 | data = start + position; |
|
1128 | data = start + position; | |
1140 | // Verify |
|
1129 | // Verify | |
1141 | if (expected_len + data != end) { |
|
1130 | if (expected_len + data != end) { | |
1142 | PyErr_SetString(PyExc_ValueError, "bser data len != header len"); |
|
1131 | PyErr_SetString(PyExc_ValueError, "bser data len != header len"); | |
1143 | return NULL; |
|
1132 | return NULL; | |
1144 | } |
|
1133 | } | |
1145 |
|
1134 | |||
1146 | return bser_loads_recursive(&data, end, &ctx); |
|
1135 | return bser_loads_recursive(&data, end, &ctx); | |
1147 | } |
|
1136 | } | |
1148 |
|
1137 | |||
1149 | static PyObject* bser_load(PyObject* self, PyObject* args, PyObject* kw) { |
|
1138 | static PyObject* bser_load(PyObject* self, PyObject* args, PyObject* kw) { | |
1150 |
PyObject |
|
1139 | PyObject* load; | |
|
1140 | PyObject* load_method; | |||
|
1141 | PyObject* string; | |||
|
1142 | PyObject* load_method_args; | |||
|
1143 | PyObject* load_method_kwargs; | |||
1151 | PyObject* fp = NULL; |
|
1144 | PyObject* fp = NULL; | |
1152 | PyObject* mutable_obj = NULL; |
|
1145 | PyObject* mutable_obj = NULL; | |
1153 |
c |
|
1146 | PyObject* value_encoding = NULL; | |
1154 |
c |
|
1147 | PyObject* value_errors = NULL; | |
1155 |
|
1148 | |||
1156 | static char* kw_list[] = { |
|
1149 | static char* kw_list[] = { | |
1157 | "fp", "mutable", "value_encoding", "value_errors", NULL}; |
|
1150 | "fp", "mutable", "value_encoding", "value_errors", NULL}; | |
1158 |
|
1151 | |||
1159 | if (!PyArg_ParseTupleAndKeywords( |
|
1152 | if (!PyArg_ParseTupleAndKeywords( | |
1160 | args, |
|
1153 | args, | |
1161 | kw, |
|
1154 | kw, | |
1162 |
"OO |
|
1155 | "O|OOO:load", | |
1163 | kw_list, |
|
1156 | kw_list, | |
1164 | &fp, |
|
1157 | &fp, | |
1165 | &mutable_obj, |
|
1158 | &mutable_obj, | |
1166 | &value_encoding, |
|
1159 | &value_encoding, | |
1167 | &value_errors)) { |
|
1160 | &value_errors)) { | |
1168 | return NULL; |
|
1161 | return NULL; | |
1169 | } |
|
1162 | } | |
1170 |
|
1163 | |||
1171 | load = PyImport_ImportModule("pywatchman.load"); |
|
1164 | load = PyImport_ImportModule("pywatchman.load"); | |
1172 | if (load == NULL) { |
|
1165 | if (load == NULL) { | |
1173 | return NULL; |
|
1166 | return NULL; | |
1174 | } |
|
1167 | } | |
1175 | string = PyObject_CallMethod( |
|
1168 | load_method = PyObject_GetAttrString(load, "load"); | |
1176 | load, "load", "OOzz", fp, mutable_obj, value_encoding, value_errors); |
|
1169 | if (load_method == NULL) { | |
|
1170 | return NULL; | |||
|
1171 | } | |||
|
1172 | // Mandatory method arguments | |||
|
1173 | load_method_args = Py_BuildValue("(O)", fp); | |||
|
1174 | if (load_method_args == NULL) { | |||
|
1175 | return NULL; | |||
|
1176 | } | |||
|
1177 | // Optional method arguments | |||
|
1178 | load_method_kwargs = PyDict_New(); | |||
|
1179 | if (load_method_kwargs == NULL) { | |||
|
1180 | return NULL; | |||
|
1181 | } | |||
|
1182 | if (mutable_obj) { | |||
|
1183 | PyDict_SetItemString(load_method_kwargs, "mutable", mutable_obj); | |||
|
1184 | } | |||
|
1185 | if (value_encoding) { | |||
|
1186 | PyDict_SetItemString(load_method_kwargs, "value_encoding", value_encoding); | |||
|
1187 | } | |||
|
1188 | if (value_errors) { | |||
|
1189 | PyDict_SetItemString(load_method_kwargs, "value_errors", value_errors); | |||
|
1190 | } | |||
|
1191 | string = PyObject_Call(load_method, load_method_args, load_method_kwargs); | |||
|
1192 | Py_DECREF(load_method_kwargs); | |||
|
1193 | Py_DECREF(load_method_args); | |||
|
1194 | Py_DECREF(load_method); | |||
1177 | Py_DECREF(load); |
|
1195 | Py_DECREF(load); | |
1178 | return string; |
|
1196 | return string; | |
1179 | } |
|
1197 | } | |
1180 |
|
1198 | |||
1181 | // clang-format off |
|
1199 | // clang-format off | |
1182 | static PyMethodDef bser_methods[] = { |
|
1200 | static PyMethodDef bser_methods[] = { | |
1183 | {"loads", (PyCFunction)bser_loads, METH_VARARGS | METH_KEYWORDS, |
|
1201 | {"loads", (PyCFunction)bser_loads, METH_VARARGS | METH_KEYWORDS, | |
1184 | "Deserialize string."}, |
|
1202 | "Deserialize string."}, | |
1185 | {"load", (PyCFunction)bser_load, METH_VARARGS | METH_KEYWORDS, |
|
1203 | {"load", (PyCFunction)bser_load, METH_VARARGS | METH_KEYWORDS, | |
1186 | "Deserialize a file object"}, |
|
1204 | "Deserialize a file object"}, | |
1187 | {"pdu_info", (PyCFunction)bser_pdu_info, METH_VARARGS, |
|
1205 | {"pdu_info", (PyCFunction)bser_pdu_info, METH_VARARGS, | |
1188 | "Extract PDU information."}, |
|
1206 | "Extract PDU information."}, | |
1189 | {"pdu_len", (PyCFunction)bser_pdu_len, METH_VARARGS, |
|
1207 | {"pdu_len", (PyCFunction)bser_pdu_len, METH_VARARGS, | |
1190 | "Extract total PDU length."}, |
|
1208 | "Extract total PDU length."}, | |
1191 | {"dumps", (PyCFunction)bser_dumps, METH_VARARGS | METH_KEYWORDS, |
|
1209 | {"dumps", (PyCFunction)bser_dumps, METH_VARARGS | METH_KEYWORDS, | |
1192 | "Serialize string."}, |
|
1210 | "Serialize string."}, | |
1193 | {NULL, NULL, 0, NULL} |
|
1211 | {NULL, NULL, 0, NULL} | |
1194 | }; |
|
1212 | }; | |
1195 |
|
1213 | |||
1196 | #if PY_MAJOR_VERSION >= 3 |
|
1214 | #if PY_MAJOR_VERSION >= 3 | |
1197 | static struct PyModuleDef bser_module = { |
|
1215 | static struct PyModuleDef bser_module = { | |
1198 | PyModuleDef_HEAD_INIT, |
|
1216 | PyModuleDef_HEAD_INIT, | |
1199 | "bser", |
|
1217 | "bser", | |
1200 | "Efficient encoding and decoding of BSER.", |
|
1218 | "Efficient encoding and decoding of BSER.", | |
1201 | -1, |
|
1219 | -1, | |
1202 | bser_methods |
|
1220 | bser_methods | |
1203 | }; |
|
1221 | }; | |
1204 | // clang-format on |
|
1222 | // clang-format on | |
1205 |
|
1223 | |||
1206 | PyMODINIT_FUNC PyInit_bser(void) { |
|
1224 | PyMODINIT_FUNC PyInit_bser(void) { | |
1207 | PyObject* mod; |
|
1225 | PyObject* mod; | |
1208 |
|
1226 | |||
1209 | mod = PyModule_Create(&bser_module); |
|
1227 | mod = PyModule_Create(&bser_module); | |
1210 | PyType_Ready(&bserObjectType); |
|
1228 | PyType_Ready(&bserObjectType); | |
1211 |
|
1229 | |||
1212 | return mod; |
|
1230 | return mod; | |
1213 | } |
|
1231 | } | |
1214 | #else |
|
1232 | #else | |
1215 |
|
1233 | |||
1216 | PyMODINIT_FUNC initbser(void) { |
|
1234 | PyMODINIT_FUNC initbser(void) { | |
1217 | (void)Py_InitModule("bser", bser_methods); |
|
1235 | (void)Py_InitModule("bser", bser_methods); | |
1218 | PyType_Ready(&bserObjectType); |
|
1236 | PyType_Ready(&bserObjectType); | |
1219 | } |
|
1237 | } | |
1220 | #endif // PY_MAJOR_VERSION >= 3 |
|
1238 | #endif // PY_MAJOR_VERSION >= 3 | |
1221 |
|
1239 | |||
1222 | /* vim:ts=2:sw=2:et: |
|
1240 | /* vim:ts=2:sw=2:et: | |
1223 | */ |
|
1241 | */ |
@@ -1,71 +1,77 b'' | |||||
1 | # Copyright 2015 Facebook, Inc. |
|
1 | # Copyright 2015 Facebook, Inc. | |
2 | # All rights reserved. |
|
2 | # All rights reserved. | |
3 | # |
|
3 | # | |
4 | # Redistribution and use in source and binary forms, with or without |
|
4 | # Redistribution and use in source and binary forms, with or without | |
5 | # modification, are permitted provided that the following conditions are met: |
|
5 | # modification, are permitted provided that the following conditions are met: | |
6 | # |
|
6 | # | |
7 | # * Redistributions of source code must retain the above copyright notice, |
|
7 | # * Redistributions of source code must retain the above copyright notice, | |
8 | # this list of conditions and the following disclaimer. |
|
8 | # this list of conditions and the following disclaimer. | |
9 | # |
|
9 | # | |
10 | # * Redistributions in binary form must reproduce the above copyright notice, |
|
10 | # * Redistributions in binary form must reproduce the above copyright notice, | |
11 | # this list of conditions and the following disclaimer in the documentation |
|
11 | # this list of conditions and the following disclaimer in the documentation | |
12 | # and/or other materials provided with the distribution. |
|
12 | # and/or other materials provided with the distribution. | |
13 | # |
|
13 | # | |
14 | # * Neither the name Facebook nor the names of its contributors may be used to |
|
14 | # * Neither the name Facebook nor the names of its contributors may be used to | |
15 | # endorse or promote products derived from this software without specific |
|
15 | # endorse or promote products derived from this software without specific | |
16 | # prior written permission. |
|
16 | # prior written permission. | |
17 | # |
|
17 | # | |
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
|
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" | |
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
|
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
|
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE | |
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
|
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE | |
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
|
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
|
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR | |
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
|
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER | |
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
|
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, | |
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
|
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
|
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
28 |
|
28 | |||
29 | from __future__ import absolute_import |
|
|||
30 | from __future__ import division |
|
|||
31 | from __future__ import print_function |
|
|||
32 | # no unicode literals |
|
29 | # no unicode literals | |
|
30 | from __future__ import absolute_import, division, print_function | |||
33 |
|
31 | |||
34 | import re |
|
32 | import re | |
35 |
|
33 | |||
|
34 | ||||
36 | def parse_version(vstr): |
|
35 | def parse_version(vstr): | |
37 | res = 0 |
|
36 | res = 0 | |
38 |
for n in vstr.split( |
|
37 | for n in vstr.split("."): | |
39 | res = res * 1000 |
|
38 | res = res * 1000 | |
40 | res = res + int(n) |
|
39 | res = res + int(n) | |
41 | return res |
|
40 | return res | |
42 |
|
41 | |||
|
42 | ||||
43 | cap_versions = { |
|
43 | cap_versions = { | |
44 | "cmd-watch-del-all": "3.1.1", |
|
44 | "cmd-watch-del-all": "3.1.1", | |
45 | "cmd-watch-project": "3.1", |
|
45 | "cmd-watch-project": "3.1", | |
46 | "relative_root": "3.3", |
|
46 | "relative_root": "3.3", | |
47 | "term-dirname": "3.1", |
|
47 | "term-dirname": "3.1", | |
48 | "term-idirname": "3.1", |
|
48 | "term-idirname": "3.1", | |
49 | "wildmatch": "3.7", |
|
49 | "wildmatch": "3.7", | |
50 | } |
|
50 | } | |
51 |
|
51 | |||
|
52 | ||||
52 | def check(version, name): |
|
53 | def check(version, name): | |
53 | if name in cap_versions: |
|
54 | if name in cap_versions: | |
54 | return version >= parse_version(cap_versions[name]) |
|
55 | return version >= parse_version(cap_versions[name]) | |
55 | return False |
|
56 | return False | |
56 |
|
57 | |||
|
58 | ||||
57 | def synthesize(vers, opts): |
|
59 | def synthesize(vers, opts): | |
58 | """ Synthesize a capability enabled version response |
|
60 | """ Synthesize a capability enabled version response | |
59 | This is a very limited emulation for relatively recent feature sets |
|
61 | This is a very limited emulation for relatively recent feature sets | |
60 | """ |
|
62 | """ | |
61 |
parsed_version = parse_version(vers[ |
|
63 | parsed_version = parse_version(vers["version"]) | |
62 |
vers[ |
|
64 | vers["capabilities"] = {} | |
63 |
for name in opts[ |
|
65 | for name in opts["optional"]: | |
64 |
vers[ |
|
66 | vers["capabilities"][name] = check(parsed_version, name) | |
65 | for name in opts['required']: |
|
67 | failed = False # noqa: F841 T25377293 Grandfathered in | |
|
68 | for name in opts["required"]: | |||
66 | have = check(parsed_version, name) |
|
69 | have = check(parsed_version, name) | |
67 |
vers[ |
|
70 | vers["capabilities"][name] = have | |
68 | if not have: |
|
71 | if not have: | |
69 | vers['error'] = 'client required capability `' + name + \ |
|
72 | vers["error"] = ( | |
70 | '` is not supported by this server' |
|
73 | "client required capability `" | |
|
74 | + name | |||
|
75 | + "` is not supported by this server" | |||
|
76 | ) | |||
71 | return vers |
|
77 | return vers |
@@ -1,65 +1,71 b'' | |||||
1 | # Copyright 2016-present Facebook, Inc. |
|
1 | # Copyright 2016-present Facebook, Inc. | |
2 | # All rights reserved. |
|
2 | # All rights reserved. | |
3 | # |
|
3 | # | |
4 | # Redistribution and use in source and binary forms, with or without |
|
4 | # Redistribution and use in source and binary forms, with or without | |
5 | # modification, are permitted provided that the following conditions are met: |
|
5 | # modification, are permitted provided that the following conditions are met: | |
6 | # |
|
6 | # | |
7 | # * Redistributions of source code must retain the above copyright notice, |
|
7 | # * Redistributions of source code must retain the above copyright notice, | |
8 | # this list of conditions and the following disclaimer. |
|
8 | # this list of conditions and the following disclaimer. | |
9 | # |
|
9 | # | |
10 | # * Redistributions in binary form must reproduce the above copyright notice, |
|
10 | # * Redistributions in binary form must reproduce the above copyright notice, | |
11 | # this list of conditions and the following disclaimer in the documentation |
|
11 | # this list of conditions and the following disclaimer in the documentation | |
12 | # and/or other materials provided with the distribution. |
|
12 | # and/or other materials provided with the distribution. | |
13 | # |
|
13 | # | |
14 | # * Neither the name Facebook nor the names of its contributors may be used to |
|
14 | # * Neither the name Facebook nor the names of its contributors may be used to | |
15 | # endorse or promote products derived from this software without specific |
|
15 | # endorse or promote products derived from this software without specific | |
16 | # prior written permission. |
|
16 | # prior written permission. | |
17 | # |
|
17 | # | |
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
|
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" | |
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
|
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
|
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE | |
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
|
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE | |
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
|
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
|
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR | |
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
|
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER | |
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
|
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, | |
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
|
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
|
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
28 |
|
28 | |||
29 | from __future__ import absolute_import |
|
|||
30 | from __future__ import division |
|
|||
31 | from __future__ import print_function |
|
|||
32 | # no unicode literals |
|
29 | # no unicode literals | |
33 |
|
30 | from __future__ import absolute_import, division, print_function | ||
34 | '''Compatibility module across Python 2 and 3.''' |
|
|||
35 |
|
31 | |||
36 | import sys |
|
32 | import sys | |
37 |
|
33 | |||
|
34 | ||||
|
35 | """Compatibility module across Python 2 and 3.""" | |||
|
36 | ||||
|
37 | ||||
|
38 | PYTHON2 = sys.version_info < (3, 0) | |||
38 | PYTHON3 = sys.version_info >= (3, 0) |
|
39 | PYTHON3 = sys.version_info >= (3, 0) | |
39 |
|
40 | |||
40 | # This is adapted from https://bitbucket.org/gutworth/six, and used under the |
|
41 | # This is adapted from https://bitbucket.org/gutworth/six, and used under the | |
41 | # MIT license. See LICENSE for a full copyright notice. |
|
42 | # MIT license. See LICENSE for a full copyright notice. | |
42 | if PYTHON3: |
|
43 | if PYTHON3: | |
|
44 | ||||
43 | def reraise(tp, value, tb=None): |
|
45 | def reraise(tp, value, tb=None): | |
44 | try: |
|
46 | try: | |
45 | if value is None: |
|
47 | if value is None: | |
46 | value = tp() |
|
48 | value = tp() | |
47 | if value.__traceback__ is not tb: |
|
49 | if value.__traceback__ is not tb: | |
48 | raise value.with_traceback(tb) |
|
50 | raise value.with_traceback(tb) | |
49 | raise value |
|
51 | raise value | |
50 | finally: |
|
52 | finally: | |
51 | value = None |
|
53 | value = None | |
52 | tb = None |
|
54 | tb = None | |
|
55 | ||||
|
56 | ||||
53 | else: |
|
57 | else: | |
54 |
exec( |
|
58 | exec( | |
|
59 | """ | |||
55 |
|
|
60 | def reraise(tp, value, tb=None): | |
56 |
|
|
61 | try: | |
57 |
|
|
62 | raise tp, value, tb | |
58 |
|
|
63 | finally: | |
59 |
|
|
64 | tb = None | |
60 |
|
|
65 | """.strip() | |
|
66 | ) | |||
61 |
|
67 | |||
62 | if PYTHON3: |
|
68 | if PYTHON3: | |
63 | UNICODE = str |
|
69 | UNICODE = str | |
64 | else: |
|
70 | else: | |
65 | UNICODE = unicode |
|
71 | UNICODE = unicode # noqa: F821 We handled versioning above |
@@ -1,73 +1,75 b'' | |||||
1 | # Copyright 2016-present Facebook, Inc. |
|
1 | # Copyright 2016-present Facebook, Inc. | |
2 | # All rights reserved. |
|
2 | # All rights reserved. | |
3 | # |
|
3 | # | |
4 | # Redistribution and use in source and binary forms, with or without |
|
4 | # Redistribution and use in source and binary forms, with or without | |
5 | # modification, are permitted provided that the following conditions are met: |
|
5 | # modification, are permitted provided that the following conditions are met: | |
6 | # |
|
6 | # | |
7 | # * Redistributions of source code must retain the above copyright notice, |
|
7 | # * Redistributions of source code must retain the above copyright notice, | |
8 | # this list of conditions and the following disclaimer. |
|
8 | # this list of conditions and the following disclaimer. | |
9 | # |
|
9 | # | |
10 | # * Redistributions in binary form must reproduce the above copyright notice, |
|
10 | # * Redistributions in binary form must reproduce the above copyright notice, | |
11 | # this list of conditions and the following disclaimer in the documentation |
|
11 | # this list of conditions and the following disclaimer in the documentation | |
12 | # and/or other materials provided with the distribution. |
|
12 | # and/or other materials provided with the distribution. | |
13 | # |
|
13 | # | |
14 | # * Neither the name Facebook nor the names of its contributors may be used to |
|
14 | # * Neither the name Facebook nor the names of its contributors may be used to | |
15 | # endorse or promote products derived from this software without specific |
|
15 | # endorse or promote products derived from this software without specific | |
16 | # prior written permission. |
|
16 | # prior written permission. | |
17 | # |
|
17 | # | |
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
|
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" | |
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
|
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
|
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE | |
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
|
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE | |
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
|
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
|
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR | |
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
|
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER | |
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
|
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, | |
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
|
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
|
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
28 |
|
28 | |||
29 | from __future__ import absolute_import |
|
|||
30 | from __future__ import division |
|
|||
31 | from __future__ import print_function |
|
|||
32 | # no unicode literals |
|
29 | # no unicode literals | |
33 |
|
30 | from __future__ import absolute_import, division, print_function | ||
34 | '''Module to deal with filename encoding on the local system, as returned by |
|
|||
35 | Watchman.''' |
|
|||
36 |
|
31 | |||
37 | import sys |
|
32 | import sys | |
38 |
|
33 | |||
39 |
from . import |
|
34 | from . import compat | |
40 | compat, |
|
35 | ||
41 | ) |
|
36 | ||
|
37 | """Module to deal with filename encoding on the local system, as returned by | |||
|
38 | Watchman.""" | |||
|
39 | ||||
42 |
|
40 | |||
43 | if compat.PYTHON3: |
|
41 | if compat.PYTHON3: | |
44 |
default_local_errors = |
|
42 | default_local_errors = "surrogateescape" | |
45 |
|
43 | |||
46 | def get_local_encoding(): |
|
44 | def get_local_encoding(): | |
47 |
if sys.platform == |
|
45 | if sys.platform == "win32": | |
48 | # Watchman always returns UTF-8 encoded strings on Windows. |
|
46 | # Watchman always returns UTF-8 encoded strings on Windows. | |
49 |
return |
|
47 | return "utf-8" | |
50 | # On the Python 3 versions we support, sys.getfilesystemencoding never |
|
48 | # On the Python 3 versions we support, sys.getfilesystemencoding never | |
51 | # returns None. |
|
49 | # returns None. | |
52 | return sys.getfilesystemencoding() |
|
50 | return sys.getfilesystemencoding() | |
|
51 | ||||
|
52 | ||||
53 | else: |
|
53 | else: | |
54 | # Python 2 doesn't support surrogateescape, so use 'strict' by |
|
54 | # Python 2 doesn't support surrogateescape, so use 'strict' by | |
55 | # default. Users can register a custom surrogateescape error handler and use |
|
55 | # default. Users can register a custom surrogateescape error handler and use | |
56 | # that if they so desire. |
|
56 | # that if they so desire. | |
57 |
default_local_errors = |
|
57 | default_local_errors = "strict" | |
58 |
|
58 | |||
59 | def get_local_encoding(): |
|
59 | def get_local_encoding(): | |
60 |
if sys.platform == |
|
60 | if sys.platform == "win32": | |
61 | # Watchman always returns UTF-8 encoded strings on Windows. |
|
61 | # Watchman always returns UTF-8 encoded strings on Windows. | |
62 |
return |
|
62 | return "utf-8" | |
63 | fsencoding = sys.getfilesystemencoding() |
|
63 | fsencoding = sys.getfilesystemencoding() | |
64 | if fsencoding is None: |
|
64 | if fsencoding is None: | |
65 | # This is very unlikely to happen, but if it does, just use UTF-8 |
|
65 | # This is very unlikely to happen, but if it does, just use UTF-8 | |
66 |
fsencoding = |
|
66 | fsencoding = "utf-8" | |
67 | return fsencoding |
|
67 | return fsencoding | |
68 |
|
68 | |||
|
69 | ||||
69 | def encode_local(s): |
|
70 | def encode_local(s): | |
70 | return s.encode(get_local_encoding(), default_local_errors) |
|
71 | return s.encode(get_local_encoding(), default_local_errors) | |
71 |
|
72 | |||
|
73 | ||||
72 | def decode_local(bs): |
|
74 | def decode_local(bs): | |
73 | return bs.decode(get_local_encoding(), default_local_errors) |
|
75 | return bs.decode(get_local_encoding(), default_local_errors) |
@@ -1,107 +1,109 b'' | |||||
1 | # Copyright 2016 Facebook, Inc. |
|
1 | # Copyright 2016 Facebook, Inc. | |
2 | # All rights reserved. |
|
2 | # All rights reserved. | |
3 | # |
|
3 | # | |
4 | # Redistribution and use in source and binary forms, with or without |
|
4 | # Redistribution and use in source and binary forms, with or without | |
5 | # modification, are permitted provided that the following conditions are met: |
|
5 | # modification, are permitted provided that the following conditions are met: | |
6 | # |
|
6 | # | |
7 | # * Redistributions of source code must retain the above copyright notice, |
|
7 | # * Redistributions of source code must retain the above copyright notice, | |
8 | # this list of conditions and the following disclaimer. |
|
8 | # this list of conditions and the following disclaimer. | |
9 | # |
|
9 | # | |
10 | # * Redistributions in binary form must reproduce the above copyright notice, |
|
10 | # * Redistributions in binary form must reproduce the above copyright notice, | |
11 | # this list of conditions and the following disclaimer in the documentation |
|
11 | # this list of conditions and the following disclaimer in the documentation | |
12 | # and/or other materials provided with the distribution. |
|
12 | # and/or other materials provided with the distribution. | |
13 | # |
|
13 | # | |
14 | # * Neither the name Facebook nor the names of its contributors may be used to |
|
14 | # * Neither the name Facebook nor the names of its contributors may be used to | |
15 | # endorse or promote products derived from this software without specific |
|
15 | # endorse or promote products derived from this software without specific | |
16 | # prior written permission. |
|
16 | # prior written permission. | |
17 | # |
|
17 | # | |
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
|
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" | |
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
|
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
|
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE | |
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
|
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE | |
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
|
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
|
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR | |
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
|
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER | |
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
|
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, | |
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
|
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
|
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
28 |
|
28 | |||
29 | from __future__ import absolute_import |
|
|||
30 | from __future__ import division |
|
|||
31 | from __future__ import print_function |
|
|||
32 | # no unicode literals |
|
29 | # no unicode literals | |
|
30 | from __future__ import absolute_import, division, print_function | |||
|
31 | ||||
|
32 | import ctypes | |||
|
33 | ||||
33 |
|
34 | |||
34 | try: |
|
35 | try: | |
35 | from . import bser |
|
36 | from . import bser | |
36 | except ImportError: |
|
37 | except ImportError: | |
37 | from . import pybser as bser |
|
38 | from . import pybser as bser | |
38 |
|
39 | |||
39 | import ctypes |
|
|||
40 |
|
40 | |||
41 | EMPTY_HEADER = b"\x00\x01\x05\x00\x00\x00\x00" |
|
41 | EMPTY_HEADER = b"\x00\x01\x05\x00\x00\x00\x00" | |
42 |
|
42 | |||
43 |
|
43 | |||
44 | def _read_bytes(fp, buf): |
|
44 | def _read_bytes(fp, buf): | |
45 | """Read bytes from a file-like object |
|
45 | """Read bytes from a file-like object | |
46 |
|
46 | |||
47 | @param fp: File-like object that implements read(int) |
|
47 | @param fp: File-like object that implements read(int) | |
48 | @type fp: file |
|
48 | @type fp: file | |
49 |
|
49 | |||
50 | @param buf: Buffer to read into |
|
50 | @param buf: Buffer to read into | |
51 | @type buf: bytes |
|
51 | @type buf: bytes | |
52 |
|
52 | |||
53 | @return: buf |
|
53 | @return: buf | |
54 | """ |
|
54 | """ | |
55 |
|
55 | |||
56 | # Do the first read without resizing the input buffer |
|
56 | # Do the first read without resizing the input buffer | |
57 | offset = 0 |
|
57 | offset = 0 | |
58 | remaining = len(buf) |
|
58 | remaining = len(buf) | |
59 | while remaining > 0: |
|
59 | while remaining > 0: | |
60 | l = fp.readinto((ctypes.c_char * remaining).from_buffer(buf, offset)) |
|
60 | l = fp.readinto((ctypes.c_char * remaining).from_buffer(buf, offset)) | |
61 | if l is None or l == 0: |
|
61 | if l is None or l == 0: | |
62 | return offset |
|
62 | return offset | |
63 | offset += l |
|
63 | offset += l | |
64 | remaining -= l |
|
64 | remaining -= l | |
65 | return offset |
|
65 | return offset | |
66 |
|
66 | |||
67 |
|
67 | |||
68 | def load(fp, mutable=True, value_encoding=None, value_errors=None): |
|
68 | def load(fp, mutable=True, value_encoding=None, value_errors=None): | |
69 | """Deserialize a BSER-encoded blob. |
|
69 | """Deserialize a BSER-encoded blob. | |
70 |
|
70 | |||
71 | @param fp: The file-object to deserialize. |
|
71 | @param fp: The file-object to deserialize. | |
72 | @type file: |
|
72 | @type file: | |
73 |
|
73 | |||
74 | @param mutable: Whether to return mutable results. |
|
74 | @param mutable: Whether to return mutable results. | |
75 | @type mutable: bool |
|
75 | @type mutable: bool | |
76 |
|
76 | |||
77 | @param value_encoding: Optional codec to use to decode values. If |
|
77 | @param value_encoding: Optional codec to use to decode values. If | |
78 | unspecified or None, return values as bytestrings. |
|
78 | unspecified or None, return values as bytestrings. | |
79 | @type value_encoding: str |
|
79 | @type value_encoding: str | |
80 |
|
80 | |||
81 | @param value_errors: Optional error handler for codec. 'strict' by default. |
|
81 | @param value_errors: Optional error handler for codec. 'strict' by default. | |
82 | The other most common argument is 'surrogateescape' on |
|
82 | The other most common argument is 'surrogateescape' on | |
83 | Python 3. If value_encoding is None, this is ignored. |
|
83 | Python 3. If value_encoding is None, this is ignored. | |
84 | @type value_errors: str |
|
84 | @type value_errors: str | |
85 | """ |
|
85 | """ | |
86 | buf = ctypes.create_string_buffer(8192) |
|
86 | buf = ctypes.create_string_buffer(8192) | |
87 | SNIFF_BUFFER_SIZE = len(EMPTY_HEADER) |
|
87 | SNIFF_BUFFER_SIZE = len(EMPTY_HEADER) | |
88 | header = (ctypes.c_char * SNIFF_BUFFER_SIZE).from_buffer(buf) |
|
88 | header = (ctypes.c_char * SNIFF_BUFFER_SIZE).from_buffer(buf) | |
89 | read_len = _read_bytes(fp, header) |
|
89 | read_len = _read_bytes(fp, header) | |
90 | if read_len < len(header): |
|
90 | if read_len < len(header): | |
91 | return None |
|
91 | return None | |
92 |
|
92 | |||
93 | total_len = bser.pdu_len(buf) |
|
93 | total_len = bser.pdu_len(buf) | |
94 | if total_len > len(buf): |
|
94 | if total_len > len(buf): | |
95 | ctypes.resize(buf, total_len) |
|
95 | ctypes.resize(buf, total_len) | |
96 |
|
96 | |||
97 | body = (ctypes.c_char * (total_len - len(header))).from_buffer( |
|
97 | body = (ctypes.c_char * (total_len - len(header))).from_buffer( | |
98 |
buf, len(header) |
|
98 | buf, len(header) | |
|
99 | ) | |||
99 | read_len = _read_bytes(fp, body) |
|
100 | read_len = _read_bytes(fp, body) | |
100 | if read_len < len(body): |
|
101 | if read_len < len(body): | |
101 |
raise RuntimeError( |
|
102 | raise RuntimeError("bser data ended early") | |
102 |
|
103 | |||
103 | return bser.loads( |
|
104 | return bser.loads( | |
104 | (ctypes.c_char * total_len).from_buffer(buf, 0), |
|
105 | (ctypes.c_char * total_len).from_buffer(buf, 0), | |
105 | mutable, |
|
106 | mutable, | |
106 | value_encoding, |
|
107 | value_encoding, | |
107 |
value_errors |
|
108 | value_errors, | |
|
109 | ) |
@@ -1,483 +1,576 b'' | |||||
1 | # Copyright 2015 Facebook, Inc. |
|
1 | # Copyright 2015 Facebook, Inc. | |
2 | # All rights reserved. |
|
2 | # All rights reserved. | |
3 | # |
|
3 | # | |
4 | # Redistribution and use in source and binary forms, with or without |
|
4 | # Redistribution and use in source and binary forms, with or without | |
5 | # modification, are permitted provided that the following conditions are met: |
|
5 | # modification, are permitted provided that the following conditions are met: | |
6 | # |
|
6 | # | |
7 | # * Redistributions of source code must retain the above copyright notice, |
|
7 | # * Redistributions of source code must retain the above copyright notice, | |
8 | # this list of conditions and the following disclaimer. |
|
8 | # this list of conditions and the following disclaimer. | |
9 | # |
|
9 | # | |
10 | # * Redistributions in binary form must reproduce the above copyright notice, |
|
10 | # * Redistributions in binary form must reproduce the above copyright notice, | |
11 | # this list of conditions and the following disclaimer in the documentation |
|
11 | # this list of conditions and the following disclaimer in the documentation | |
12 | # and/or other materials provided with the distribution. |
|
12 | # and/or other materials provided with the distribution. | |
13 | # |
|
13 | # | |
14 | # * Neither the name Facebook nor the names of its contributors may be used to |
|
14 | # * Neither the name Facebook nor the names of its contributors may be used to | |
15 | # endorse or promote products derived from this software without specific |
|
15 | # endorse or promote products derived from this software without specific | |
16 | # prior written permission. |
|
16 | # prior written permission. | |
17 | # |
|
17 | # | |
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
|
18 | # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" | |
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
|
19 | # AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
|
20 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE | |
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
|
21 | # DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE | |
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
|
22 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
|
23 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR | |
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
|
24 | # SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER | |
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
|
25 | # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, | |
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
|
26 | # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
|
27 | # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
28 |
|
28 | |||
29 | from __future__ import absolute_import |
|
|||
30 | from __future__ import division |
|
|||
31 | from __future__ import print_function |
|
|||
32 | # no unicode literals |
|
29 | # no unicode literals | |
|
30 | from __future__ import absolute_import, division, print_function | |||
33 |
|
31 | |||
34 | import binascii |
|
32 | import binascii | |
35 | import collections |
|
33 | import collections | |
36 | import ctypes |
|
34 | import ctypes | |
37 | import struct |
|
35 | import struct | |
38 | import sys |
|
36 | import sys | |
39 |
|
37 | |||
40 |
from . import |
|
38 | from . import compat | |
41 | compat, |
|
39 | ||
42 | ) |
|
|||
43 |
|
40 | |||
44 |
BSER_ARRAY = b |
|
41 | BSER_ARRAY = b"\x00" | |
45 |
BSER_OBJECT = b |
|
42 | BSER_OBJECT = b"\x01" | |
46 |
BSER_BYTESTRING = b |
|
43 | BSER_BYTESTRING = b"\x02" | |
47 |
BSER_INT8 = b |
|
44 | BSER_INT8 = b"\x03" | |
48 |
BSER_INT16 = b |
|
45 | BSER_INT16 = b"\x04" | |
49 |
BSER_INT32 = b |
|
46 | BSER_INT32 = b"\x05" | |
50 |
BSER_INT64 = b |
|
47 | BSER_INT64 = b"\x06" | |
51 |
BSER_REAL = b |
|
48 | BSER_REAL = b"\x07" | |
52 |
BSER_TRUE = b |
|
49 | BSER_TRUE = b"\x08" | |
53 |
BSER_FALSE = b |
|
50 | BSER_FALSE = b"\x09" | |
54 |
BSER_NULL = b |
|
51 | BSER_NULL = b"\x0a" | |
55 |
BSER_TEMPLATE = b |
|
52 | BSER_TEMPLATE = b"\x0b" | |
56 |
BSER_SKIP = b |
|
53 | BSER_SKIP = b"\x0c" | |
57 |
BSER_UTF8STRING = b |
|
54 | BSER_UTF8STRING = b"\x0d" | |
58 |
|
55 | |||
59 | if compat.PYTHON3: |
|
56 | if compat.PYTHON3: | |
60 | STRING_TYPES = (str, bytes) |
|
57 | STRING_TYPES = (str, bytes) | |
61 | unicode = str |
|
58 | unicode = str | |
|
59 | ||||
62 | def tobytes(i): |
|
60 | def tobytes(i): | |
63 |
return str(i).encode( |
|
61 | return str(i).encode("ascii") | |
|
62 | ||||
64 | long = int |
|
63 | long = int | |
65 | else: |
|
64 | else: | |
66 | STRING_TYPES = (unicode, str) |
|
65 | STRING_TYPES = (unicode, str) | |
67 | tobytes = bytes |
|
66 | tobytes = bytes | |
68 |
|
67 | |||
69 | # Leave room for the serialization header, which includes |
|
68 | # Leave room for the serialization header, which includes | |
70 | # our overall length. To make things simpler, we'll use an |
|
69 | # our overall length. To make things simpler, we'll use an | |
71 | # int32 for the header |
|
70 | # int32 for the header | |
72 | EMPTY_HEADER = b"\x00\x01\x05\x00\x00\x00\x00" |
|
71 | EMPTY_HEADER = b"\x00\x01\x05\x00\x00\x00\x00" | |
73 | EMPTY_HEADER_V2 = b"\x00\x02\x00\x00\x00\x00\x05\x00\x00\x00\x00" |
|
72 | EMPTY_HEADER_V2 = b"\x00\x02\x00\x00\x00\x00\x05\x00\x00\x00\x00" | |
74 |
|
73 | |||
|
74 | ||||
75 | def _int_size(x): |
|
75 | def _int_size(x): | |
76 | """Return the smallest size int that can store the value""" |
|
76 | """Return the smallest size int that can store the value""" | |
77 | if -0x80 <= x <= 0x7F: |
|
77 | if -0x80 <= x <= 0x7F: | |
78 | return 1 |
|
78 | return 1 | |
79 | elif -0x8000 <= x <= 0x7FFF: |
|
79 | elif -0x8000 <= x <= 0x7FFF: | |
80 | return 2 |
|
80 | return 2 | |
81 | elif -0x80000000 <= x <= 0x7FFFFFFF: |
|
81 | elif -0x80000000 <= x <= 0x7FFFFFFF: | |
82 | return 4 |
|
82 | return 4 | |
83 | elif long(-0x8000000000000000) <= x <= long(0x7FFFFFFFFFFFFFFF): |
|
83 | elif long(-0x8000000000000000) <= x <= long(0x7FFFFFFFFFFFFFFF): | |
84 | return 8 |
|
84 | return 8 | |
85 | else: |
|
85 | else: | |
86 |
raise RuntimeError( |
|
86 | raise RuntimeError("Cannot represent value: " + str(x)) | |
|
87 | ||||
87 |
|
88 | |||
88 | def _buf_pos(buf, pos): |
|
89 | def _buf_pos(buf, pos): | |
89 | ret = buf[pos] |
|
90 | ret = buf[pos] | |
90 | # In Python 2, buf is a str array so buf[pos] is a string. In Python 3, buf |
|
91 | # Normalize the return type to bytes | |
91 | # is a bytes array and buf[pos] is an integer. |
|
92 | if compat.PYTHON3 and not isinstance(ret, bytes): | |
92 | if compat.PYTHON3: |
|
|||
93 | ret = bytes((ret,)) |
|
93 | ret = bytes((ret,)) | |
94 | return ret |
|
94 | return ret | |
95 |
|
95 | |||
|
96 | ||||
96 | class _bser_buffer(object): |
|
97 | class _bser_buffer(object): | |
97 |
|
||||
98 | def __init__(self, version): |
|
98 | def __init__(self, version): | |
99 | self.bser_version = version |
|
99 | self.bser_version = version | |
100 | self.buf = ctypes.create_string_buffer(8192) |
|
100 | self.buf = ctypes.create_string_buffer(8192) | |
101 | if self.bser_version == 1: |
|
101 | if self.bser_version == 1: | |
102 | struct.pack_into(tobytes(len(EMPTY_HEADER)) + b's', self.buf, 0, |
|
102 | struct.pack_into( | |
103 | EMPTY_HEADER) |
|
103 | tobytes(len(EMPTY_HEADER)) + b"s", self.buf, 0, EMPTY_HEADER | |
|
104 | ) | |||
104 | self.wpos = len(EMPTY_HEADER) |
|
105 | self.wpos = len(EMPTY_HEADER) | |
105 | else: |
|
106 | else: | |
106 | assert self.bser_version == 2 |
|
107 | assert self.bser_version == 2 | |
107 | struct.pack_into(tobytes(len(EMPTY_HEADER_V2)) + b's', self.buf, 0, |
|
108 | struct.pack_into( | |
108 |
|
|
109 | tobytes(len(EMPTY_HEADER_V2)) + b"s", | |
|
110 | self.buf, | |||
|
111 | 0, | |||
|
112 | EMPTY_HEADER_V2, | |||
|
113 | ) | |||
109 | self.wpos = len(EMPTY_HEADER_V2) |
|
114 | self.wpos = len(EMPTY_HEADER_V2) | |
110 |
|
115 | |||
111 | def ensure_size(self, size): |
|
116 | def ensure_size(self, size): | |
112 | while ctypes.sizeof(self.buf) - self.wpos < size: |
|
117 | while ctypes.sizeof(self.buf) - self.wpos < size: | |
113 | ctypes.resize(self.buf, ctypes.sizeof(self.buf) * 2) |
|
118 | ctypes.resize(self.buf, ctypes.sizeof(self.buf) * 2) | |
114 |
|
119 | |||
115 | def append_long(self, val): |
|
120 | def append_long(self, val): | |
116 | size = _int_size(val) |
|
121 | size = _int_size(val) | |
117 | to_write = size + 1 |
|
122 | to_write = size + 1 | |
118 | self.ensure_size(to_write) |
|
123 | self.ensure_size(to_write) | |
119 | if size == 1: |
|
124 | if size == 1: | |
120 |
struct.pack_into(b |
|
125 | struct.pack_into(b"=cb", self.buf, self.wpos, BSER_INT8, val) | |
121 | elif size == 2: |
|
126 | elif size == 2: | |
122 |
struct.pack_into(b |
|
127 | struct.pack_into(b"=ch", self.buf, self.wpos, BSER_INT16, val) | |
123 | elif size == 4: |
|
128 | elif size == 4: | |
124 |
struct.pack_into(b |
|
129 | struct.pack_into(b"=ci", self.buf, self.wpos, BSER_INT32, val) | |
125 | elif size == 8: |
|
130 | elif size == 8: | |
126 |
struct.pack_into(b |
|
131 | struct.pack_into(b"=cq", self.buf, self.wpos, BSER_INT64, val) | |
127 | else: |
|
132 | else: | |
128 |
raise RuntimeError( |
|
133 | raise RuntimeError("Cannot represent this long value") | |
129 | self.wpos += to_write |
|
134 | self.wpos += to_write | |
130 |
|
135 | |||
131 |
|
||||
132 | def append_string(self, s): |
|
136 | def append_string(self, s): | |
133 | if isinstance(s, unicode): |
|
137 | if isinstance(s, unicode): | |
134 |
s = s.encode( |
|
138 | s = s.encode("utf-8") | |
135 | s_len = len(s) |
|
139 | s_len = len(s) | |
136 | size = _int_size(s_len) |
|
140 | size = _int_size(s_len) | |
137 | to_write = 2 + size + s_len |
|
141 | to_write = 2 + size + s_len | |
138 | self.ensure_size(to_write) |
|
142 | self.ensure_size(to_write) | |
139 | if size == 1: |
|
143 | if size == 1: | |
140 | struct.pack_into(b'=ccb' + tobytes(s_len) + b's', self.buf, |
|
144 | struct.pack_into( | |
141 | self.wpos, BSER_BYTESTRING, BSER_INT8, s_len, s) |
|
145 | b"=ccb" + tobytes(s_len) + b"s", | |
|
146 | self.buf, | |||
|
147 | self.wpos, | |||
|
148 | BSER_BYTESTRING, | |||
|
149 | BSER_INT8, | |||
|
150 | s_len, | |||
|
151 | s, | |||
|
152 | ) | |||
142 | elif size == 2: |
|
153 | elif size == 2: | |
143 | struct.pack_into(b'=cch' + tobytes(s_len) + b's', self.buf, |
|
154 | struct.pack_into( | |
144 | self.wpos, BSER_BYTESTRING, BSER_INT16, s_len, s) |
|
155 | b"=cch" + tobytes(s_len) + b"s", | |
|
156 | self.buf, | |||
|
157 | self.wpos, | |||
|
158 | BSER_BYTESTRING, | |||
|
159 | BSER_INT16, | |||
|
160 | s_len, | |||
|
161 | s, | |||
|
162 | ) | |||
145 | elif size == 4: |
|
163 | elif size == 4: | |
146 | struct.pack_into(b'=cci' + tobytes(s_len) + b's', self.buf, |
|
164 | struct.pack_into( | |
147 | self.wpos, BSER_BYTESTRING, BSER_INT32, s_len, s) |
|
165 | b"=cci" + tobytes(s_len) + b"s", | |
|
166 | self.buf, | |||
|
167 | self.wpos, | |||
|
168 | BSER_BYTESTRING, | |||
|
169 | BSER_INT32, | |||
|
170 | s_len, | |||
|
171 | s, | |||
|
172 | ) | |||
148 | elif size == 8: |
|
173 | elif size == 8: | |
149 | struct.pack_into(b'=ccq' + tobytes(s_len) + b's', self.buf, |
|
174 | struct.pack_into( | |
150 | self.wpos, BSER_BYTESTRING, BSER_INT64, s_len, s) |
|
175 | b"=ccq" + tobytes(s_len) + b"s", | |
|
176 | self.buf, | |||
|
177 | self.wpos, | |||
|
178 | BSER_BYTESTRING, | |||
|
179 | BSER_INT64, | |||
|
180 | s_len, | |||
|
181 | s, | |||
|
182 | ) | |||
151 | else: |
|
183 | else: | |
152 |
raise RuntimeError( |
|
184 | raise RuntimeError("Cannot represent this string value") | |
153 | self.wpos += to_write |
|
185 | self.wpos += to_write | |
154 |
|
186 | |||
155 |
|
||||
156 | def append_recursive(self, val): |
|
187 | def append_recursive(self, val): | |
157 | if isinstance(val, bool): |
|
188 | if isinstance(val, bool): | |
158 | needed = 1 |
|
189 | needed = 1 | |
159 | self.ensure_size(needed) |
|
190 | self.ensure_size(needed) | |
160 | if val: |
|
191 | if val: | |
161 | to_encode = BSER_TRUE |
|
192 | to_encode = BSER_TRUE | |
162 | else: |
|
193 | else: | |
163 | to_encode = BSER_FALSE |
|
194 | to_encode = BSER_FALSE | |
164 |
struct.pack_into(b |
|
195 | struct.pack_into(b"=c", self.buf, self.wpos, to_encode) | |
165 | self.wpos += needed |
|
196 | self.wpos += needed | |
166 | elif val is None: |
|
197 | elif val is None: | |
167 | needed = 1 |
|
198 | needed = 1 | |
168 | self.ensure_size(needed) |
|
199 | self.ensure_size(needed) | |
169 |
struct.pack_into(b |
|
200 | struct.pack_into(b"=c", self.buf, self.wpos, BSER_NULL) | |
170 | self.wpos += needed |
|
201 | self.wpos += needed | |
171 | elif isinstance(val, (int, long)): |
|
202 | elif isinstance(val, (int, long)): | |
172 | self.append_long(val) |
|
203 | self.append_long(val) | |
173 | elif isinstance(val, STRING_TYPES): |
|
204 | elif isinstance(val, STRING_TYPES): | |
174 | self.append_string(val) |
|
205 | self.append_string(val) | |
175 | elif isinstance(val, float): |
|
206 | elif isinstance(val, float): | |
176 | needed = 9 |
|
207 | needed = 9 | |
177 | self.ensure_size(needed) |
|
208 | self.ensure_size(needed) | |
178 |
struct.pack_into(b |
|
209 | struct.pack_into(b"=cd", self.buf, self.wpos, BSER_REAL, val) | |
179 | self.wpos += needed |
|
210 | self.wpos += needed | |
180 |
elif isinstance(val, collections.Mapping) and |
|
211 | elif isinstance(val, collections.Mapping) and isinstance( | |
181 |
|
|
212 | val, collections.Sized | |
|
213 | ): | |||
182 | val_len = len(val) |
|
214 | val_len = len(val) | |
183 | size = _int_size(val_len) |
|
215 | size = _int_size(val_len) | |
184 | needed = 2 + size |
|
216 | needed = 2 + size | |
185 | self.ensure_size(needed) |
|
217 | self.ensure_size(needed) | |
186 | if size == 1: |
|
218 | if size == 1: | |
187 |
struct.pack_into( |
|
219 | struct.pack_into( | |
188 |
|
|
220 | b"=ccb", | |
|
221 | self.buf, | |||
|
222 | self.wpos, | |||
|
223 | BSER_OBJECT, | |||
|
224 | BSER_INT8, | |||
|
225 | val_len, | |||
|
226 | ) | |||
189 | elif size == 2: |
|
227 | elif size == 2: | |
190 |
struct.pack_into( |
|
228 | struct.pack_into( | |
191 |
|
|
229 | b"=cch", | |
|
230 | self.buf, | |||
|
231 | self.wpos, | |||
|
232 | BSER_OBJECT, | |||
|
233 | BSER_INT16, | |||
|
234 | val_len, | |||
|
235 | ) | |||
192 | elif size == 4: |
|
236 | elif size == 4: | |
193 |
struct.pack_into( |
|
237 | struct.pack_into( | |
194 |
|
|
238 | b"=cci", | |
|
239 | self.buf, | |||
|
240 | self.wpos, | |||
|
241 | BSER_OBJECT, | |||
|
242 | BSER_INT32, | |||
|
243 | val_len, | |||
|
244 | ) | |||
195 | elif size == 8: |
|
245 | elif size == 8: | |
196 |
struct.pack_into( |
|
246 | struct.pack_into( | |
197 |
|
|
247 | b"=ccq", | |
|
248 | self.buf, | |||
|
249 | self.wpos, | |||
|
250 | BSER_OBJECT, | |||
|
251 | BSER_INT64, | |||
|
252 | val_len, | |||
|
253 | ) | |||
198 | else: |
|
254 | else: | |
199 |
raise RuntimeError( |
|
255 | raise RuntimeError("Cannot represent this mapping value") | |
200 | self.wpos += needed |
|
256 | self.wpos += needed | |
201 | if compat.PYTHON3: |
|
257 | if compat.PYTHON3: | |
202 | iteritems = val.items() |
|
258 | iteritems = val.items() | |
203 | else: |
|
259 | else: | |
204 | iteritems = val.iteritems() |
|
260 | iteritems = val.iteritems() # noqa: B301 Checked version above | |
205 | for k, v in iteritems: |
|
261 | for k, v in iteritems: | |
206 | self.append_string(k) |
|
262 | self.append_string(k) | |
207 | self.append_recursive(v) |
|
263 | self.append_recursive(v) | |
208 |
elif isinstance(val, collections.Iterable) and |
|
264 | elif isinstance(val, collections.Iterable) and isinstance( | |
209 |
|
|
265 | val, collections.Sized | |
|
266 | ): | |||
210 | val_len = len(val) |
|
267 | val_len = len(val) | |
211 | size = _int_size(val_len) |
|
268 | size = _int_size(val_len) | |
212 | needed = 2 + size |
|
269 | needed = 2 + size | |
213 | self.ensure_size(needed) |
|
270 | self.ensure_size(needed) | |
214 | if size == 1: |
|
271 | if size == 1: | |
215 |
struct.pack_into( |
|
272 | struct.pack_into( | |
216 |
BSER_INT8, val_len |
|
273 | b"=ccb", self.buf, self.wpos, BSER_ARRAY, BSER_INT8, val_len | |
|
274 | ) | |||
217 | elif size == 2: |
|
275 | elif size == 2: | |
218 |
struct.pack_into( |
|
276 | struct.pack_into( | |
219 |
|
|
277 | b"=cch", | |
|
278 | self.buf, | |||
|
279 | self.wpos, | |||
|
280 | BSER_ARRAY, | |||
|
281 | BSER_INT16, | |||
|
282 | val_len, | |||
|
283 | ) | |||
220 | elif size == 4: |
|
284 | elif size == 4: | |
221 |
struct.pack_into( |
|
285 | struct.pack_into( | |
222 |
|
|
286 | b"=cci", | |
|
287 | self.buf, | |||
|
288 | self.wpos, | |||
|
289 | BSER_ARRAY, | |||
|
290 | BSER_INT32, | |||
|
291 | val_len, | |||
|
292 | ) | |||
223 | elif size == 8: |
|
293 | elif size == 8: | |
224 |
struct.pack_into( |
|
294 | struct.pack_into( | |
225 |
|
|
295 | b"=ccq", | |
|
296 | self.buf, | |||
|
297 | self.wpos, | |||
|
298 | BSER_ARRAY, | |||
|
299 | BSER_INT64, | |||
|
300 | val_len, | |||
|
301 | ) | |||
226 | else: |
|
302 | else: | |
227 |
raise RuntimeError( |
|
303 | raise RuntimeError("Cannot represent this sequence value") | |
228 | self.wpos += needed |
|
304 | self.wpos += needed | |
229 | for v in val: |
|
305 | for v in val: | |
230 | self.append_recursive(v) |
|
306 | self.append_recursive(v) | |
231 | else: |
|
307 | else: | |
232 |
raise RuntimeError( |
|
308 | raise RuntimeError("Cannot represent unknown value type") | |
233 |
|
309 | |||
234 |
|
310 | |||
235 | def dumps(obj, version=1, capabilities=0): |
|
311 | def dumps(obj, version=1, capabilities=0): | |
236 | bser_buf = _bser_buffer(version=version) |
|
312 | bser_buf = _bser_buffer(version=version) | |
237 | bser_buf.append_recursive(obj) |
|
313 | bser_buf.append_recursive(obj) | |
238 | # Now fill in the overall length |
|
314 | # Now fill in the overall length | |
239 | if version == 1: |
|
315 | if version == 1: | |
240 | obj_len = bser_buf.wpos - len(EMPTY_HEADER) |
|
316 | obj_len = bser_buf.wpos - len(EMPTY_HEADER) | |
241 |
struct.pack_into(b |
|
317 | struct.pack_into(b"=i", bser_buf.buf, 3, obj_len) | |
242 | else: |
|
318 | else: | |
243 | obj_len = bser_buf.wpos - len(EMPTY_HEADER_V2) |
|
319 | obj_len = bser_buf.wpos - len(EMPTY_HEADER_V2) | |
244 |
struct.pack_into(b |
|
320 | struct.pack_into(b"=i", bser_buf.buf, 2, capabilities) | |
245 |
struct.pack_into(b |
|
321 | struct.pack_into(b"=i", bser_buf.buf, 7, obj_len) | |
246 | return bser_buf.buf.raw[:bser_buf.wpos] |
|
322 | return bser_buf.buf.raw[: bser_buf.wpos] | |
|
323 | ||||
247 |
|
324 | |||
248 | # This is a quack-alike with the bserObjectType in bser.c |
|
325 | # This is a quack-alike with the bserObjectType in bser.c | |
249 | # It provides by getattr accessors and getitem for both index |
|
326 | # It provides by getattr accessors and getitem for both index | |
250 | # and name. |
|
327 | # and name. | |
251 | class _BunserDict(object): |
|
328 | class _BunserDict(object): | |
252 |
__slots__ = ( |
|
329 | __slots__ = ("_keys", "_values") | |
253 |
|
330 | |||
254 | def __init__(self, keys, values): |
|
331 | def __init__(self, keys, values): | |
255 | self._keys = keys |
|
332 | self._keys = keys | |
256 | self._values = values |
|
333 | self._values = values | |
257 |
|
334 | |||
258 | def __getattr__(self, name): |
|
335 | def __getattr__(self, name): | |
259 | return self.__getitem__(name) |
|
336 | return self.__getitem__(name) | |
260 |
|
337 | |||
261 | def __getitem__(self, key): |
|
338 | def __getitem__(self, key): | |
262 | if isinstance(key, (int, long)): |
|
339 | if isinstance(key, (int, long)): | |
263 | return self._values[key] |
|
340 | return self._values[key] | |
264 |
elif key.startswith( |
|
341 | elif key.startswith("st_"): | |
265 | # hack^Wfeature to allow mercurial to use "st_size" to |
|
342 | # hack^Wfeature to allow mercurial to use "st_size" to | |
266 | # reference "size" |
|
343 | # reference "size" | |
267 | key = key[3:] |
|
344 | key = key[3:] | |
268 | try: |
|
345 | try: | |
269 | return self._values[self._keys.index(key)] |
|
346 | return self._values[self._keys.index(key)] | |
270 | except ValueError: |
|
347 | except ValueError: | |
271 |
raise KeyError( |
|
348 | raise KeyError("_BunserDict has no key %s" % key) | |
272 |
|
349 | |||
273 | def __len__(self): |
|
350 | def __len__(self): | |
274 | return len(self._keys) |
|
351 | return len(self._keys) | |
275 |
|
352 | |||
|
353 | ||||
276 | class Bunser(object): |
|
354 | class Bunser(object): | |
277 | def __init__(self, mutable=True, value_encoding=None, value_errors=None): |
|
355 | def __init__(self, mutable=True, value_encoding=None, value_errors=None): | |
278 | self.mutable = mutable |
|
356 | self.mutable = mutable | |
279 | self.value_encoding = value_encoding |
|
357 | self.value_encoding = value_encoding | |
280 |
|
358 | |||
281 | if value_encoding is None: |
|
359 | if value_encoding is None: | |
282 | self.value_errors = None |
|
360 | self.value_errors = None | |
283 | elif value_errors is None: |
|
361 | elif value_errors is None: | |
284 |
self.value_errors = |
|
362 | self.value_errors = "strict" | |
285 | else: |
|
363 | else: | |
286 | self.value_errors = value_errors |
|
364 | self.value_errors = value_errors | |
287 |
|
365 | |||
288 | @staticmethod |
|
366 | @staticmethod | |
289 | def unser_int(buf, pos): |
|
367 | def unser_int(buf, pos): | |
290 | try: |
|
368 | try: | |
291 | int_type = _buf_pos(buf, pos) |
|
369 | int_type = _buf_pos(buf, pos) | |
292 | except IndexError: |
|
370 | except IndexError: | |
293 |
raise ValueError( |
|
371 | raise ValueError("Invalid bser int encoding, pos out of range") | |
294 | if int_type == BSER_INT8: |
|
372 | if int_type == BSER_INT8: | |
295 | needed = 2 |
|
373 | needed = 2 | |
296 |
fmt = b |
|
374 | fmt = b"=b" | |
297 | elif int_type == BSER_INT16: |
|
375 | elif int_type == BSER_INT16: | |
298 | needed = 3 |
|
376 | needed = 3 | |
299 |
fmt = b |
|
377 | fmt = b"=h" | |
300 | elif int_type == BSER_INT32: |
|
378 | elif int_type == BSER_INT32: | |
301 | needed = 5 |
|
379 | needed = 5 | |
302 |
fmt = b |
|
380 | fmt = b"=i" | |
303 | elif int_type == BSER_INT64: |
|
381 | elif int_type == BSER_INT64: | |
304 | needed = 9 |
|
382 | needed = 9 | |
305 |
fmt = b |
|
383 | fmt = b"=q" | |
306 | else: |
|
384 | else: | |
307 |
raise ValueError( |
|
385 | raise ValueError( | |
308 | binascii.hexlify(int_type).decode('ascii')) |
|
386 | "Invalid bser int encoding 0x%s at position %s" | |
|
387 | % (binascii.hexlify(int_type).decode("ascii"), pos) | |||
|
388 | ) | |||
309 | int_val = struct.unpack_from(fmt, buf, pos + 1)[0] |
|
389 | int_val = struct.unpack_from(fmt, buf, pos + 1)[0] | |
310 | return (int_val, pos + needed) |
|
390 | return (int_val, pos + needed) | |
311 |
|
391 | |||
312 | def unser_utf8_string(self, buf, pos): |
|
392 | def unser_utf8_string(self, buf, pos): | |
313 | str_len, pos = self.unser_int(buf, pos + 1) |
|
393 | str_len, pos = self.unser_int(buf, pos + 1) | |
314 |
str_val = struct.unpack_from(tobytes(str_len) + b |
|
394 | str_val = struct.unpack_from(tobytes(str_len) + b"s", buf, pos)[0] | |
315 |
return (str_val.decode( |
|
395 | return (str_val.decode("utf-8"), pos + str_len) | |
316 |
|
396 | |||
317 | def unser_bytestring(self, buf, pos): |
|
397 | def unser_bytestring(self, buf, pos): | |
318 | str_len, pos = self.unser_int(buf, pos + 1) |
|
398 | str_len, pos = self.unser_int(buf, pos + 1) | |
319 |
str_val = struct.unpack_from(tobytes(str_len) + b |
|
399 | str_val = struct.unpack_from(tobytes(str_len) + b"s", buf, pos)[0] | |
320 | if self.value_encoding is not None: |
|
400 | if self.value_encoding is not None: | |
321 | str_val = str_val.decode(self.value_encoding, self.value_errors) |
|
401 | str_val = str_val.decode(self.value_encoding, self.value_errors) | |
322 | # str_len stays the same because that's the length in bytes |
|
402 | # str_len stays the same because that's the length in bytes | |
323 | return (str_val, pos + str_len) |
|
403 | return (str_val, pos + str_len) | |
324 |
|
404 | |||
325 | def unser_array(self, buf, pos): |
|
405 | def unser_array(self, buf, pos): | |
326 | arr_len, pos = self.unser_int(buf, pos + 1) |
|
406 | arr_len, pos = self.unser_int(buf, pos + 1) | |
327 | arr = [] |
|
407 | arr = [] | |
328 |
for |
|
408 | for _ in range(arr_len): | |
329 | arr_item, pos = self.loads_recursive(buf, pos) |
|
409 | arr_item, pos = self.loads_recursive(buf, pos) | |
330 | arr.append(arr_item) |
|
410 | arr.append(arr_item) | |
331 |
|
411 | |||
332 | if not self.mutable: |
|
412 | if not self.mutable: | |
333 | arr = tuple(arr) |
|
413 | arr = tuple(arr) | |
334 |
|
414 | |||
335 | return arr, pos |
|
415 | return arr, pos | |
336 |
|
416 | |||
337 | def unser_object(self, buf, pos): |
|
417 | def unser_object(self, buf, pos): | |
338 | obj_len, pos = self.unser_int(buf, pos + 1) |
|
418 | obj_len, pos = self.unser_int(buf, pos + 1) | |
339 | if self.mutable: |
|
419 | if self.mutable: | |
340 | obj = {} |
|
420 | obj = {} | |
341 | else: |
|
421 | else: | |
342 | keys = [] |
|
422 | keys = [] | |
343 | vals = [] |
|
423 | vals = [] | |
344 |
|
424 | |||
345 |
for |
|
425 | for _ in range(obj_len): | |
346 | key, pos = self.unser_utf8_string(buf, pos) |
|
426 | key, pos = self.unser_utf8_string(buf, pos) | |
347 | val, pos = self.loads_recursive(buf, pos) |
|
427 | val, pos = self.loads_recursive(buf, pos) | |
348 | if self.mutable: |
|
428 | if self.mutable: | |
349 | obj[key] = val |
|
429 | obj[key] = val | |
350 | else: |
|
430 | else: | |
351 | keys.append(key) |
|
431 | keys.append(key) | |
352 | vals.append(val) |
|
432 | vals.append(val) | |
353 |
|
433 | |||
354 | if not self.mutable: |
|
434 | if not self.mutable: | |
355 | obj = _BunserDict(keys, vals) |
|
435 | obj = _BunserDict(keys, vals) | |
356 |
|
436 | |||
357 | return obj, pos |
|
437 | return obj, pos | |
358 |
|
438 | |||
359 | def unser_template(self, buf, pos): |
|
439 | def unser_template(self, buf, pos): | |
360 | val_type = _buf_pos(buf, pos + 1) |
|
440 | val_type = _buf_pos(buf, pos + 1) | |
361 | if val_type != BSER_ARRAY: |
|
441 | if val_type != BSER_ARRAY: | |
362 |
raise RuntimeError( |
|
442 | raise RuntimeError("Expect ARRAY to follow TEMPLATE") | |
363 | # force UTF-8 on keys |
|
443 | # force UTF-8 on keys | |
364 |
keys_bunser = Bunser(mutable=self.mutable, value_encoding= |
|
444 | keys_bunser = Bunser(mutable=self.mutable, value_encoding="utf-8") | |
365 | keys, pos = keys_bunser.unser_array(buf, pos + 1) |
|
445 | keys, pos = keys_bunser.unser_array(buf, pos + 1) | |
366 | nitems, pos = self.unser_int(buf, pos) |
|
446 | nitems, pos = self.unser_int(buf, pos) | |
367 | arr = [] |
|
447 | arr = [] | |
368 |
for |
|
448 | for _ in range(nitems): | |
369 | if self.mutable: |
|
449 | if self.mutable: | |
370 | obj = {} |
|
450 | obj = {} | |
371 | else: |
|
451 | else: | |
372 | vals = [] |
|
452 | vals = [] | |
373 |
|
453 | |||
374 | for keyidx in range(len(keys)): |
|
454 | for keyidx in range(len(keys)): | |
375 | if _buf_pos(buf, pos) == BSER_SKIP: |
|
455 | if _buf_pos(buf, pos) == BSER_SKIP: | |
376 | pos += 1 |
|
456 | pos += 1 | |
377 | ele = None |
|
457 | ele = None | |
378 | else: |
|
458 | else: | |
379 | ele, pos = self.loads_recursive(buf, pos) |
|
459 | ele, pos = self.loads_recursive(buf, pos) | |
380 |
|
460 | |||
381 | if self.mutable: |
|
461 | if self.mutable: | |
382 | key = keys[keyidx] |
|
462 | key = keys[keyidx] | |
383 | obj[key] = ele |
|
463 | obj[key] = ele | |
384 | else: |
|
464 | else: | |
385 | vals.append(ele) |
|
465 | vals.append(ele) | |
386 |
|
466 | |||
387 | if not self.mutable: |
|
467 | if not self.mutable: | |
388 | obj = _BunserDict(keys, vals) |
|
468 | obj = _BunserDict(keys, vals) | |
389 |
|
469 | |||
390 | arr.append(obj) |
|
470 | arr.append(obj) | |
391 | return arr, pos |
|
471 | return arr, pos | |
392 |
|
472 | |||
393 | def loads_recursive(self, buf, pos): |
|
473 | def loads_recursive(self, buf, pos): | |
394 | val_type = _buf_pos(buf, pos) |
|
474 | val_type = _buf_pos(buf, pos) | |
395 | if (val_type == BSER_INT8 or val_type == BSER_INT16 or |
|
475 | if ( | |
396 |
val_type == BSER_INT |
|
476 | val_type == BSER_INT8 | |
|
477 | or val_type == BSER_INT16 | |||
|
478 | or val_type == BSER_INT32 | |||
|
479 | or val_type == BSER_INT64 | |||
|
480 | ): | |||
397 | return self.unser_int(buf, pos) |
|
481 | return self.unser_int(buf, pos) | |
398 | elif val_type == BSER_REAL: |
|
482 | elif val_type == BSER_REAL: | |
399 |
val = struct.unpack_from(b |
|
483 | val = struct.unpack_from(b"=d", buf, pos + 1)[0] | |
400 | return (val, pos + 9) |
|
484 | return (val, pos + 9) | |
401 | elif val_type == BSER_TRUE: |
|
485 | elif val_type == BSER_TRUE: | |
402 | return (True, pos + 1) |
|
486 | return (True, pos + 1) | |
403 | elif val_type == BSER_FALSE: |
|
487 | elif val_type == BSER_FALSE: | |
404 | return (False, pos + 1) |
|
488 | return (False, pos + 1) | |
405 | elif val_type == BSER_NULL: |
|
489 | elif val_type == BSER_NULL: | |
406 | return (None, pos + 1) |
|
490 | return (None, pos + 1) | |
407 | elif val_type == BSER_BYTESTRING: |
|
491 | elif val_type == BSER_BYTESTRING: | |
408 | return self.unser_bytestring(buf, pos) |
|
492 | return self.unser_bytestring(buf, pos) | |
409 | elif val_type == BSER_UTF8STRING: |
|
493 | elif val_type == BSER_UTF8STRING: | |
410 | return self.unser_utf8_string(buf, pos) |
|
494 | return self.unser_utf8_string(buf, pos) | |
411 | elif val_type == BSER_ARRAY: |
|
495 | elif val_type == BSER_ARRAY: | |
412 | return self.unser_array(buf, pos) |
|
496 | return self.unser_array(buf, pos) | |
413 | elif val_type == BSER_OBJECT: |
|
497 | elif val_type == BSER_OBJECT: | |
414 | return self.unser_object(buf, pos) |
|
498 | return self.unser_object(buf, pos) | |
415 | elif val_type == BSER_TEMPLATE: |
|
499 | elif val_type == BSER_TEMPLATE: | |
416 | return self.unser_template(buf, pos) |
|
500 | return self.unser_template(buf, pos) | |
417 | else: |
|
501 | else: | |
418 |
raise ValueError( |
|
502 | raise ValueError( | |
419 | binascii.hexlify(val_type).decode('ascii')) |
|
503 | "unhandled bser opcode 0x%s" | |
|
504 | % binascii.hexlify(val_type).decode("ascii") | |||
|
505 | ) | |||
420 |
|
506 | |||
421 |
|
507 | |||
422 | def _pdu_info_helper(buf): |
|
508 | def _pdu_info_helper(buf): | |
|
509 | bser_version = -1 | |||
423 | if buf[0:2] == EMPTY_HEADER[0:2]: |
|
510 | if buf[0:2] == EMPTY_HEADER[0:2]: | |
424 | bser_version = 1 |
|
511 | bser_version = 1 | |
425 | bser_capabilities = 0 |
|
512 | bser_capabilities = 0 | |
426 | expected_len, pos2 = Bunser.unser_int(buf, 2) |
|
513 | expected_len, pos2 = Bunser.unser_int(buf, 2) | |
427 | elif buf[0:2] == EMPTY_HEADER_V2[0:2]: |
|
514 | elif buf[0:2] == EMPTY_HEADER_V2[0:2]: | |
428 | if len(buf) < 8: |
|
515 | if len(buf) < 8: | |
429 |
raise ValueError( |
|
516 | raise ValueError("Invalid BSER header") | |
430 | bser_version = 2 |
|
517 | bser_version = 2 | |
431 | bser_capabilities = struct.unpack_from("I", buf, 2)[0] |
|
518 | bser_capabilities = struct.unpack_from("I", buf, 2)[0] | |
432 | expected_len, pos2 = Bunser.unser_int(buf, 6) |
|
519 | expected_len, pos2 = Bunser.unser_int(buf, 6) | |
433 | else: |
|
520 | else: | |
434 |
raise ValueError( |
|
521 | raise ValueError("Invalid BSER header") | |
435 |
|
522 | |||
436 | return bser_version, bser_capabilities, expected_len, pos2 |
|
523 | return bser_version, bser_capabilities, expected_len, pos2 | |
437 |
|
524 | |||
438 |
|
525 | |||
439 | def pdu_info(buf): |
|
526 | def pdu_info(buf): | |
440 | info = _pdu_info_helper(buf) |
|
527 | info = _pdu_info_helper(buf) | |
441 | return info[0], info[1], info[2] + info[3] |
|
528 | return info[0], info[1], info[2] + info[3] | |
442 |
|
529 | |||
443 |
|
530 | |||
444 | def pdu_len(buf): |
|
531 | def pdu_len(buf): | |
445 | info = _pdu_info_helper(buf) |
|
532 | info = _pdu_info_helper(buf) | |
446 | return info[2] + info[3] |
|
533 | return info[2] + info[3] | |
447 |
|
534 | |||
448 |
|
535 | |||
449 | def loads(buf, mutable=True, value_encoding=None, value_errors=None): |
|
536 | def loads(buf, mutable=True, value_encoding=None, value_errors=None): | |
450 | """Deserialize a BSER-encoded blob. |
|
537 | """Deserialize a BSER-encoded blob. | |
451 |
|
538 | |||
452 | @param buf: The buffer to deserialize. |
|
539 | @param buf: The buffer to deserialize. | |
453 | @type buf: bytes |
|
540 | @type buf: bytes | |
454 |
|
541 | |||
455 | @param mutable: Whether to return mutable results. |
|
542 | @param mutable: Whether to return mutable results. | |
456 | @type mutable: bool |
|
543 | @type mutable: bool | |
457 |
|
544 | |||
458 | @param value_encoding: Optional codec to use to decode values. If |
|
545 | @param value_encoding: Optional codec to use to decode values. If | |
459 | unspecified or None, return values as bytestrings. |
|
546 | unspecified or None, return values as bytestrings. | |
460 | @type value_encoding: str |
|
547 | @type value_encoding: str | |
461 |
|
548 | |||
462 | @param value_errors: Optional error handler for codec. 'strict' by default. |
|
549 | @param value_errors: Optional error handler for codec. 'strict' by default. | |
463 | The other most common argument is 'surrogateescape' on |
|
550 | The other most common argument is 'surrogateescape' on | |
464 | Python 3. If value_encoding is None, this is ignored. |
|
551 | Python 3. If value_encoding is None, this is ignored. | |
465 | @type value_errors: str |
|
552 | @type value_errors: str | |
466 | """ |
|
553 | """ | |
467 |
|
554 | |||
468 | info = _pdu_info_helper(buf) |
|
555 | info = _pdu_info_helper(buf) | |
469 | expected_len = info[2] |
|
556 | expected_len = info[2] | |
470 | pos = info[3] |
|
557 | pos = info[3] | |
471 |
|
558 | |||
472 | if len(buf) != expected_len + pos: |
|
559 | if len(buf) != expected_len + pos: | |
473 |
raise ValueError( |
|
560 | raise ValueError( | |
|
561 | "bser data len %d != header len %d" % (expected_len + pos, len(buf)) | |||
|
562 | ) | |||
474 |
|
563 | |||
475 | bunser = Bunser(mutable=mutable, value_encoding=value_encoding, |
|
564 | bunser = Bunser( | |
476 | value_errors=value_errors) |
|
565 | mutable=mutable, | |
|
566 | value_encoding=value_encoding, | |||
|
567 | value_errors=value_errors, | |||
|
568 | ) | |||
477 |
|
569 | |||
478 | return bunser.loads_recursive(buf, pos)[0] |
|
570 | return bunser.loads_recursive(buf, pos)[0] | |
479 |
|
571 | |||
480 |
|
572 | |||
481 | def load(fp, mutable=True, value_encoding=None, value_errors=None): |
|
573 | def load(fp, mutable=True, value_encoding=None, value_errors=None): | |
482 | from . import load |
|
574 | from . import load | |
|
575 | ||||
483 | return load.load(fp, mutable, value_encoding, value_errors) |
|
576 | return load.load(fp, mutable, value_encoding, value_errors) |
@@ -1,118 +1,119 b'' | |||||
1 | # watchmanclient.py - Watchman client for the fsmonitor extension |
|
1 | # watchmanclient.py - Watchman client for the fsmonitor extension | |
2 | # |
|
2 | # | |
3 | # Copyright 2013-2016 Facebook, Inc. |
|
3 | # Copyright 2013-2016 Facebook, Inc. | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import getpass |
|
10 | import getpass | |
11 |
|
11 | |||
12 | from mercurial import util |
|
12 | from mercurial import util | |
|
13 | from mercurial.utils import procutil | |||
13 |
|
14 | |||
14 | from . import pywatchman |
|
15 | from . import pywatchman | |
15 |
|
16 | |||
16 |
|
17 | |||
17 | class Unavailable(Exception): |
|
18 | class Unavailable(Exception): | |
18 | def __init__(self, msg, warn=True, invalidate=False): |
|
19 | def __init__(self, msg, warn=True, invalidate=False): | |
19 | self.msg = msg |
|
20 | self.msg = msg | |
20 | self.warn = warn |
|
21 | self.warn = warn | |
21 | if self.msg == b'timed out waiting for response': |
|
22 | if self.msg == b'timed out waiting for response': | |
22 | self.warn = False |
|
23 | self.warn = False | |
23 | self.invalidate = invalidate |
|
24 | self.invalidate = invalidate | |
24 |
|
25 | |||
25 | def __str__(self): |
|
26 | def __str__(self): | |
26 | if self.warn: |
|
27 | if self.warn: | |
27 | return b'warning: Watchman unavailable: %s' % self.msg |
|
28 | return b'warning: Watchman unavailable: %s' % self.msg | |
28 | else: |
|
29 | else: | |
29 | return b'Watchman unavailable: %s' % self.msg |
|
30 | return b'Watchman unavailable: %s' % self.msg | |
30 |
|
31 | |||
31 |
|
32 | |||
32 | class WatchmanNoRoot(Unavailable): |
|
33 | class WatchmanNoRoot(Unavailable): | |
33 | def __init__(self, root, msg): |
|
34 | def __init__(self, root, msg): | |
34 | self.root = root |
|
35 | self.root = root | |
35 | super(WatchmanNoRoot, self).__init__(msg) |
|
36 | super(WatchmanNoRoot, self).__init__(msg) | |
36 |
|
37 | |||
37 |
|
38 | |||
38 | class client(object): |
|
39 | class client(object): | |
39 | def __init__(self, ui, root, timeout=1.0): |
|
40 | def __init__(self, ui, root, timeout=1.0): | |
40 | err = None |
|
41 | err = None | |
41 | if not self._user: |
|
42 | if not self._user: | |
42 | err = b"couldn't get user" |
|
43 | err = b"couldn't get user" | |
43 | warn = True |
|
44 | warn = True | |
44 | if self._user in ui.configlist(b'fsmonitor', b'blacklistusers'): |
|
45 | if self._user in ui.configlist(b'fsmonitor', b'blacklistusers'): | |
45 | err = b'user %s in blacklist' % self._user |
|
46 | err = b'user %s in blacklist' % self._user | |
46 | warn = False |
|
47 | warn = False | |
47 |
|
48 | |||
48 | if err: |
|
49 | if err: | |
49 | raise Unavailable(err, warn) |
|
50 | raise Unavailable(err, warn) | |
50 |
|
51 | |||
51 | self._timeout = timeout |
|
52 | self._timeout = timeout | |
52 | self._watchmanclient = None |
|
53 | self._watchmanclient = None | |
53 | self._root = root |
|
54 | self._root = root | |
54 | self._ui = ui |
|
55 | self._ui = ui | |
55 | self._firsttime = True |
|
56 | self._firsttime = True | |
56 |
|
57 | |||
57 | def settimeout(self, timeout): |
|
58 | def settimeout(self, timeout): | |
58 | self._timeout = timeout |
|
59 | self._timeout = timeout | |
59 | if self._watchmanclient is not None: |
|
60 | if self._watchmanclient is not None: | |
60 | self._watchmanclient.setTimeout(timeout) |
|
61 | self._watchmanclient.setTimeout(timeout) | |
61 |
|
62 | |||
62 | def getcurrentclock(self): |
|
63 | def getcurrentclock(self): | |
63 | result = self.command(b'clock') |
|
64 | result = self.command(b'clock') | |
64 | if not util.safehasattr(result, 'clock'): |
|
65 | if not util.safehasattr(result, 'clock'): | |
65 | raise Unavailable( |
|
66 | raise Unavailable( | |
66 | b'clock result is missing clock value', invalidate=True |
|
67 | b'clock result is missing clock value', invalidate=True | |
67 | ) |
|
68 | ) | |
68 | return result.clock |
|
69 | return result.clock | |
69 |
|
70 | |||
70 | def clearconnection(self): |
|
71 | def clearconnection(self): | |
71 | self._watchmanclient = None |
|
72 | self._watchmanclient = None | |
72 |
|
73 | |||
73 | def available(self): |
|
74 | def available(self): | |
74 | return self._watchmanclient is not None or self._firsttime |
|
75 | return self._watchmanclient is not None or self._firsttime | |
75 |
|
76 | |||
76 | @util.propertycache |
|
77 | @util.propertycache | |
77 | def _user(self): |
|
78 | def _user(self): | |
78 | try: |
|
79 | try: | |
79 | return getpass.getuser() |
|
80 | return getpass.getuser() | |
80 | except KeyError: |
|
81 | except KeyError: | |
81 | # couldn't figure out our user |
|
82 | # couldn't figure out our user | |
82 | return None |
|
83 | return None | |
83 |
|
84 | |||
84 | def _command(self, *args): |
|
85 | def _command(self, *args): | |
85 | watchmanargs = (args[0], self._root) + args[1:] |
|
86 | watchmanargs = (args[0], self._root) + args[1:] | |
86 | try: |
|
87 | try: | |
87 | if self._watchmanclient is None: |
|
88 | if self._watchmanclient is None: | |
88 | self._firsttime = False |
|
89 | self._firsttime = False | |
89 | watchman_exe = self._ui.configpath( |
|
90 | watchman_exe = self._ui.configpath( | |
90 | b'fsmonitor', b'watchman_exe' |
|
91 | b'fsmonitor', b'watchman_exe' | |
91 | ) |
|
92 | ) | |
92 | self._watchmanclient = pywatchman.client( |
|
93 | self._watchmanclient = pywatchman.client( | |
93 | timeout=self._timeout, |
|
94 | timeout=self._timeout, | |
94 | useImmutableBser=True, |
|
95 | useImmutableBser=True, | |
95 |
watchman_exe |
|
96 | binpath=procutil.tonativestr(watchman_exe), | |
96 | ) |
|
97 | ) | |
97 | return self._watchmanclient.query(*watchmanargs) |
|
98 | return self._watchmanclient.query(*watchmanargs) | |
98 | except pywatchman.CommandError as ex: |
|
99 | except pywatchman.CommandError as ex: | |
99 | if b'unable to resolve root' in ex.msg: |
|
100 | if b'unable to resolve root' in ex.msg: | |
100 | raise WatchmanNoRoot(self._root, ex.msg) |
|
101 | raise WatchmanNoRoot(self._root, ex.msg) | |
101 | raise Unavailable(ex.msg) |
|
102 | raise Unavailable(ex.msg) | |
102 | except pywatchman.WatchmanError as ex: |
|
103 | except pywatchman.WatchmanError as ex: | |
103 | raise Unavailable(str(ex)) |
|
104 | raise Unavailable(str(ex)) | |
104 |
|
105 | |||
105 | def command(self, *args): |
|
106 | def command(self, *args): | |
106 | try: |
|
107 | try: | |
107 | try: |
|
108 | try: | |
108 | return self._command(*args) |
|
109 | return self._command(*args) | |
109 | except WatchmanNoRoot: |
|
110 | except WatchmanNoRoot: | |
110 | # this 'watch' command can also raise a WatchmanNoRoot if |
|
111 | # this 'watch' command can also raise a WatchmanNoRoot if | |
111 | # watchman refuses to accept this root |
|
112 | # watchman refuses to accept this root | |
112 | self._command(b'watch') |
|
113 | self._command(b'watch') | |
113 | return self._command(*args) |
|
114 | return self._command(*args) | |
114 | except Unavailable: |
|
115 | except Unavailable: | |
115 | # this is in an outer scope to catch Unavailable form any of the |
|
116 | # this is in an outer scope to catch Unavailable form any of the | |
116 | # above _command calls |
|
117 | # above _command calls | |
117 | self._watchmanclient = None |
|
118 | self._watchmanclient = None | |
118 | raise |
|
119 | raise |
General Comments 0
You need to be logged in to leave comments.
Login now