Olivier Deprez | f4ef2d0 | 2021-04-20 13:36:24 +0200 | [diff] [blame^] | 1 | #! /usr/local/bin/python |
| 2 | |
| 3 | # NOTE: the above "/usr/local/bin/python" is NOT a mistake. It is |
| 4 | # intentionally NOT "/usr/bin/env python". On many systems |
| 5 | # (e.g. Solaris), /usr/local/bin is not in $PATH as passed to CGI |
| 6 | # scripts, and /usr/local/bin is the default directory where Python is |
| 7 | # installed, so /usr/bin/env would be unable to find python. Granted, |
| 8 | # binary installations by Linux vendors often install Python in |
| 9 | # /usr/bin. So let those vendors patch cgi.py to match their choice |
| 10 | # of installation. |
| 11 | |
| 12 | """Support module for CGI (Common Gateway Interface) scripts. |
| 13 | |
| 14 | This module defines a number of utilities for use by CGI scripts |
| 15 | written in Python. |
| 16 | """ |
| 17 | |
| 18 | # History |
| 19 | # ------- |
| 20 | # |
| 21 | # Michael McLay started this module. Steve Majewski changed the |
| 22 | # interface to SvFormContentDict and FormContentDict. The multipart |
| 23 | # parsing was inspired by code submitted by Andreas Paepcke. Guido van |
| 24 | # Rossum rewrote, reformatted and documented the module and is currently |
| 25 | # responsible for its maintenance. |
| 26 | # |
| 27 | |
| 28 | __version__ = "2.6" |
| 29 | |
| 30 | |
| 31 | # Imports |
| 32 | # ======= |
| 33 | |
| 34 | from io import StringIO, BytesIO, TextIOWrapper |
| 35 | from collections.abc import Mapping |
| 36 | import sys |
| 37 | import os |
| 38 | import urllib.parse |
| 39 | from email.parser import FeedParser |
| 40 | from email.message import Message |
| 41 | import html |
| 42 | import locale |
| 43 | import tempfile |
| 44 | |
| 45 | __all__ = ["MiniFieldStorage", "FieldStorage", "parse", "parse_multipart", |
| 46 | "parse_header", "test", "print_exception", "print_environ", |
| 47 | "print_form", "print_directory", "print_arguments", |
| 48 | "print_environ_usage"] |
| 49 | |
| 50 | # Logging support |
| 51 | # =============== |
| 52 | |
| 53 | logfile = "" # Filename to log to, if not empty |
| 54 | logfp = None # File object to log to, if not None |
| 55 | |
| 56 | def initlog(*allargs): |
| 57 | """Write a log message, if there is a log file. |
| 58 | |
| 59 | Even though this function is called initlog(), you should always |
| 60 | use log(); log is a variable that is set either to initlog |
| 61 | (initially), to dolog (once the log file has been opened), or to |
| 62 | nolog (when logging is disabled). |
| 63 | |
| 64 | The first argument is a format string; the remaining arguments (if |
| 65 | any) are arguments to the % operator, so e.g. |
| 66 | log("%s: %s", "a", "b") |
| 67 | will write "a: b" to the log file, followed by a newline. |
| 68 | |
| 69 | If the global logfp is not None, it should be a file object to |
| 70 | which log data is written. |
| 71 | |
| 72 | If the global logfp is None, the global logfile may be a string |
| 73 | giving a filename to open, in append mode. This file should be |
| 74 | world writable!!! If the file can't be opened, logging is |
| 75 | silently disabled (since there is no safe place where we could |
| 76 | send an error message). |
| 77 | |
| 78 | """ |
| 79 | global log, logfile, logfp |
| 80 | if logfile and not logfp: |
| 81 | try: |
| 82 | logfp = open(logfile, "a") |
| 83 | except OSError: |
| 84 | pass |
| 85 | if not logfp: |
| 86 | log = nolog |
| 87 | else: |
| 88 | log = dolog |
| 89 | log(*allargs) |
| 90 | |
| 91 | def dolog(fmt, *args): |
| 92 | """Write a log message to the log file. See initlog() for docs.""" |
| 93 | logfp.write(fmt%args + "\n") |
| 94 | |
| 95 | def nolog(*allargs): |
| 96 | """Dummy function, assigned to log when logging is disabled.""" |
| 97 | pass |
| 98 | |
| 99 | def closelog(): |
| 100 | """Close the log file.""" |
| 101 | global log, logfile, logfp |
| 102 | logfile = '' |
| 103 | if logfp: |
| 104 | logfp.close() |
| 105 | logfp = None |
| 106 | log = initlog |
| 107 | |
| 108 | log = initlog # The current logging function |
| 109 | |
| 110 | |
| 111 | # Parsing functions |
| 112 | # ================= |
| 113 | |
| 114 | # Maximum input we will accept when REQUEST_METHOD is POST |
| 115 | # 0 ==> unlimited input |
| 116 | maxlen = 0 |
| 117 | |
| 118 | def parse(fp=None, environ=os.environ, keep_blank_values=0, strict_parsing=0): |
| 119 | """Parse a query in the environment or from a file (default stdin) |
| 120 | |
| 121 | Arguments, all optional: |
| 122 | |
| 123 | fp : file pointer; default: sys.stdin.buffer |
| 124 | |
| 125 | environ : environment dictionary; default: os.environ |
| 126 | |
| 127 | keep_blank_values: flag indicating whether blank values in |
| 128 | percent-encoded forms should be treated as blank strings. |
| 129 | A true value indicates that blanks should be retained as |
| 130 | blank strings. The default false value indicates that |
| 131 | blank values are to be ignored and treated as if they were |
| 132 | not included. |
| 133 | |
| 134 | strict_parsing: flag indicating what to do with parsing errors. |
| 135 | If false (the default), errors are silently ignored. |
| 136 | If true, errors raise a ValueError exception. |
| 137 | """ |
| 138 | if fp is None: |
| 139 | fp = sys.stdin |
| 140 | |
| 141 | # field keys and values (except for files) are returned as strings |
| 142 | # an encoding is required to decode the bytes read from self.fp |
| 143 | if hasattr(fp,'encoding'): |
| 144 | encoding = fp.encoding |
| 145 | else: |
| 146 | encoding = 'latin-1' |
| 147 | |
| 148 | # fp.read() must return bytes |
| 149 | if isinstance(fp, TextIOWrapper): |
| 150 | fp = fp.buffer |
| 151 | |
| 152 | if not 'REQUEST_METHOD' in environ: |
| 153 | environ['REQUEST_METHOD'] = 'GET' # For testing stand-alone |
| 154 | if environ['REQUEST_METHOD'] == 'POST': |
| 155 | ctype, pdict = parse_header(environ['CONTENT_TYPE']) |
| 156 | if ctype == 'multipart/form-data': |
| 157 | return parse_multipart(fp, pdict) |
| 158 | elif ctype == 'application/x-www-form-urlencoded': |
| 159 | clength = int(environ['CONTENT_LENGTH']) |
| 160 | if maxlen and clength > maxlen: |
| 161 | raise ValueError('Maximum content length exceeded') |
| 162 | qs = fp.read(clength).decode(encoding) |
| 163 | else: |
| 164 | qs = '' # Unknown content-type |
| 165 | if 'QUERY_STRING' in environ: |
| 166 | if qs: qs = qs + '&' |
| 167 | qs = qs + environ['QUERY_STRING'] |
| 168 | elif sys.argv[1:]: |
| 169 | if qs: qs = qs + '&' |
| 170 | qs = qs + sys.argv[1] |
| 171 | environ['QUERY_STRING'] = qs # XXX Shouldn't, really |
| 172 | elif 'QUERY_STRING' in environ: |
| 173 | qs = environ['QUERY_STRING'] |
| 174 | else: |
| 175 | if sys.argv[1:]: |
| 176 | qs = sys.argv[1] |
| 177 | else: |
| 178 | qs = "" |
| 179 | environ['QUERY_STRING'] = qs # XXX Shouldn't, really |
| 180 | return urllib.parse.parse_qs(qs, keep_blank_values, strict_parsing, |
| 181 | encoding=encoding) |
| 182 | |
| 183 | |
| 184 | def parse_multipart(fp, pdict, encoding="utf-8", errors="replace"): |
| 185 | """Parse multipart input. |
| 186 | |
| 187 | Arguments: |
| 188 | fp : input file |
| 189 | pdict: dictionary containing other parameters of content-type header |
| 190 | encoding, errors: request encoding and error handler, passed to |
| 191 | FieldStorage |
| 192 | |
| 193 | Returns a dictionary just like parse_qs(): keys are the field names, each |
| 194 | value is a list of values for that field. For non-file fields, the value |
| 195 | is a list of strings. |
| 196 | """ |
| 197 | # RFC 2026, Section 5.1 : The "multipart" boundary delimiters are always |
| 198 | # represented as 7bit US-ASCII. |
| 199 | boundary = pdict['boundary'].decode('ascii') |
| 200 | ctype = "multipart/form-data; boundary={}".format(boundary) |
| 201 | headers = Message() |
| 202 | headers.set_type(ctype) |
| 203 | try: |
| 204 | headers['Content-Length'] = pdict['CONTENT-LENGTH'] |
| 205 | except KeyError: |
| 206 | pass |
| 207 | fs = FieldStorage(fp, headers=headers, encoding=encoding, errors=errors, |
| 208 | environ={'REQUEST_METHOD': 'POST'}) |
| 209 | return {k: fs.getlist(k) for k in fs} |
| 210 | |
| 211 | def _parseparam(s): |
| 212 | while s[:1] == ';': |
| 213 | s = s[1:] |
| 214 | end = s.find(';') |
| 215 | while end > 0 and (s.count('"', 0, end) - s.count('\\"', 0, end)) % 2: |
| 216 | end = s.find(';', end + 1) |
| 217 | if end < 0: |
| 218 | end = len(s) |
| 219 | f = s[:end] |
| 220 | yield f.strip() |
| 221 | s = s[end:] |
| 222 | |
| 223 | def parse_header(line): |
| 224 | """Parse a Content-type like header. |
| 225 | |
| 226 | Return the main content-type and a dictionary of options. |
| 227 | |
| 228 | """ |
| 229 | parts = _parseparam(';' + line) |
| 230 | key = parts.__next__() |
| 231 | pdict = {} |
| 232 | for p in parts: |
| 233 | i = p.find('=') |
| 234 | if i >= 0: |
| 235 | name = p[:i].strip().lower() |
| 236 | value = p[i+1:].strip() |
| 237 | if len(value) >= 2 and value[0] == value[-1] == '"': |
| 238 | value = value[1:-1] |
| 239 | value = value.replace('\\\\', '\\').replace('\\"', '"') |
| 240 | pdict[name] = value |
| 241 | return key, pdict |
| 242 | |
| 243 | |
| 244 | # Classes for field storage |
| 245 | # ========================= |
| 246 | |
| 247 | class MiniFieldStorage: |
| 248 | |
| 249 | """Like FieldStorage, for use when no file uploads are possible.""" |
| 250 | |
| 251 | # Dummy attributes |
| 252 | filename = None |
| 253 | list = None |
| 254 | type = None |
| 255 | file = None |
| 256 | type_options = {} |
| 257 | disposition = None |
| 258 | disposition_options = {} |
| 259 | headers = {} |
| 260 | |
| 261 | def __init__(self, name, value): |
| 262 | """Constructor from field name and value.""" |
| 263 | self.name = name |
| 264 | self.value = value |
| 265 | # self.file = StringIO(value) |
| 266 | |
| 267 | def __repr__(self): |
| 268 | """Return printable representation.""" |
| 269 | return "MiniFieldStorage(%r, %r)" % (self.name, self.value) |
| 270 | |
| 271 | |
| 272 | class FieldStorage: |
| 273 | |
| 274 | """Store a sequence of fields, reading multipart/form-data. |
| 275 | |
| 276 | This class provides naming, typing, files stored on disk, and |
| 277 | more. At the top level, it is accessible like a dictionary, whose |
| 278 | keys are the field names. (Note: None can occur as a field name.) |
| 279 | The items are either a Python list (if there's multiple values) or |
| 280 | another FieldStorage or MiniFieldStorage object. If it's a single |
| 281 | object, it has the following attributes: |
| 282 | |
| 283 | name: the field name, if specified; otherwise None |
| 284 | |
| 285 | filename: the filename, if specified; otherwise None; this is the |
| 286 | client side filename, *not* the file name on which it is |
| 287 | stored (that's a temporary file you don't deal with) |
| 288 | |
| 289 | value: the value as a *string*; for file uploads, this |
| 290 | transparently reads the file every time you request the value |
| 291 | and returns *bytes* |
| 292 | |
| 293 | file: the file(-like) object from which you can read the data *as |
| 294 | bytes* ; None if the data is stored a simple string |
| 295 | |
| 296 | type: the content-type, or None if not specified |
| 297 | |
| 298 | type_options: dictionary of options specified on the content-type |
| 299 | line |
| 300 | |
| 301 | disposition: content-disposition, or None if not specified |
| 302 | |
| 303 | disposition_options: dictionary of corresponding options |
| 304 | |
| 305 | headers: a dictionary(-like) object (sometimes email.message.Message or a |
| 306 | subclass thereof) containing *all* headers |
| 307 | |
| 308 | The class is subclassable, mostly for the purpose of overriding |
| 309 | the make_file() method, which is called internally to come up with |
| 310 | a file open for reading and writing. This makes it possible to |
| 311 | override the default choice of storing all files in a temporary |
| 312 | directory and unlinking them as soon as they have been opened. |
| 313 | |
| 314 | """ |
| 315 | def __init__(self, fp=None, headers=None, outerboundary=b'', |
| 316 | environ=os.environ, keep_blank_values=0, strict_parsing=0, |
| 317 | limit=None, encoding='utf-8', errors='replace', |
| 318 | max_num_fields=None): |
| 319 | """Constructor. Read multipart/* until last part. |
| 320 | |
| 321 | Arguments, all optional: |
| 322 | |
| 323 | fp : file pointer; default: sys.stdin.buffer |
| 324 | (not used when the request method is GET) |
| 325 | Can be : |
| 326 | 1. a TextIOWrapper object |
| 327 | 2. an object whose read() and readline() methods return bytes |
| 328 | |
| 329 | headers : header dictionary-like object; default: |
| 330 | taken from environ as per CGI spec |
| 331 | |
| 332 | outerboundary : terminating multipart boundary |
| 333 | (for internal use only) |
| 334 | |
| 335 | environ : environment dictionary; default: os.environ |
| 336 | |
| 337 | keep_blank_values: flag indicating whether blank values in |
| 338 | percent-encoded forms should be treated as blank strings. |
| 339 | A true value indicates that blanks should be retained as |
| 340 | blank strings. The default false value indicates that |
| 341 | blank values are to be ignored and treated as if they were |
| 342 | not included. |
| 343 | |
| 344 | strict_parsing: flag indicating what to do with parsing errors. |
| 345 | If false (the default), errors are silently ignored. |
| 346 | If true, errors raise a ValueError exception. |
| 347 | |
| 348 | limit : used internally to read parts of multipart/form-data forms, |
| 349 | to exit from the reading loop when reached. It is the difference |
| 350 | between the form content-length and the number of bytes already |
| 351 | read |
| 352 | |
| 353 | encoding, errors : the encoding and error handler used to decode the |
| 354 | binary stream to strings. Must be the same as the charset defined |
| 355 | for the page sending the form (content-type : meta http-equiv or |
| 356 | header) |
| 357 | |
| 358 | max_num_fields: int. If set, then __init__ throws a ValueError |
| 359 | if there are more than n fields read by parse_qsl(). |
| 360 | |
| 361 | """ |
| 362 | method = 'GET' |
| 363 | self.keep_blank_values = keep_blank_values |
| 364 | self.strict_parsing = strict_parsing |
| 365 | self.max_num_fields = max_num_fields |
| 366 | if 'REQUEST_METHOD' in environ: |
| 367 | method = environ['REQUEST_METHOD'].upper() |
| 368 | self.qs_on_post = None |
| 369 | if method == 'GET' or method == 'HEAD': |
| 370 | if 'QUERY_STRING' in environ: |
| 371 | qs = environ['QUERY_STRING'] |
| 372 | elif sys.argv[1:]: |
| 373 | qs = sys.argv[1] |
| 374 | else: |
| 375 | qs = "" |
| 376 | qs = qs.encode(locale.getpreferredencoding(), 'surrogateescape') |
| 377 | fp = BytesIO(qs) |
| 378 | if headers is None: |
| 379 | headers = {'content-type': |
| 380 | "application/x-www-form-urlencoded"} |
| 381 | if headers is None: |
| 382 | headers = {} |
| 383 | if method == 'POST': |
| 384 | # Set default content-type for POST to what's traditional |
| 385 | headers['content-type'] = "application/x-www-form-urlencoded" |
| 386 | if 'CONTENT_TYPE' in environ: |
| 387 | headers['content-type'] = environ['CONTENT_TYPE'] |
| 388 | if 'QUERY_STRING' in environ: |
| 389 | self.qs_on_post = environ['QUERY_STRING'] |
| 390 | if 'CONTENT_LENGTH' in environ: |
| 391 | headers['content-length'] = environ['CONTENT_LENGTH'] |
| 392 | else: |
| 393 | if not (isinstance(headers, (Mapping, Message))): |
| 394 | raise TypeError("headers must be mapping or an instance of " |
| 395 | "email.message.Message") |
| 396 | self.headers = headers |
| 397 | if fp is None: |
| 398 | self.fp = sys.stdin.buffer |
| 399 | # self.fp.read() must return bytes |
| 400 | elif isinstance(fp, TextIOWrapper): |
| 401 | self.fp = fp.buffer |
| 402 | else: |
| 403 | if not (hasattr(fp, 'read') and hasattr(fp, 'readline')): |
| 404 | raise TypeError("fp must be file pointer") |
| 405 | self.fp = fp |
| 406 | |
| 407 | self.encoding = encoding |
| 408 | self.errors = errors |
| 409 | |
| 410 | if not isinstance(outerboundary, bytes): |
| 411 | raise TypeError('outerboundary must be bytes, not %s' |
| 412 | % type(outerboundary).__name__) |
| 413 | self.outerboundary = outerboundary |
| 414 | |
| 415 | self.bytes_read = 0 |
| 416 | self.limit = limit |
| 417 | |
| 418 | # Process content-disposition header |
| 419 | cdisp, pdict = "", {} |
| 420 | if 'content-disposition' in self.headers: |
| 421 | cdisp, pdict = parse_header(self.headers['content-disposition']) |
| 422 | self.disposition = cdisp |
| 423 | self.disposition_options = pdict |
| 424 | self.name = None |
| 425 | if 'name' in pdict: |
| 426 | self.name = pdict['name'] |
| 427 | self.filename = None |
| 428 | if 'filename' in pdict: |
| 429 | self.filename = pdict['filename'] |
| 430 | self._binary_file = self.filename is not None |
| 431 | |
| 432 | # Process content-type header |
| 433 | # |
| 434 | # Honor any existing content-type header. But if there is no |
| 435 | # content-type header, use some sensible defaults. Assume |
| 436 | # outerboundary is "" at the outer level, but something non-false |
| 437 | # inside a multi-part. The default for an inner part is text/plain, |
| 438 | # but for an outer part it should be urlencoded. This should catch |
| 439 | # bogus clients which erroneously forget to include a content-type |
| 440 | # header. |
| 441 | # |
| 442 | # See below for what we do if there does exist a content-type header, |
| 443 | # but it happens to be something we don't understand. |
| 444 | if 'content-type' in self.headers: |
| 445 | ctype, pdict = parse_header(self.headers['content-type']) |
| 446 | elif self.outerboundary or method != 'POST': |
| 447 | ctype, pdict = "text/plain", {} |
| 448 | else: |
| 449 | ctype, pdict = 'application/x-www-form-urlencoded', {} |
| 450 | self.type = ctype |
| 451 | self.type_options = pdict |
| 452 | if 'boundary' in pdict: |
| 453 | self.innerboundary = pdict['boundary'].encode(self.encoding, |
| 454 | self.errors) |
| 455 | else: |
| 456 | self.innerboundary = b"" |
| 457 | |
| 458 | clen = -1 |
| 459 | if 'content-length' in self.headers: |
| 460 | try: |
| 461 | clen = int(self.headers['content-length']) |
| 462 | except ValueError: |
| 463 | pass |
| 464 | if maxlen and clen > maxlen: |
| 465 | raise ValueError('Maximum content length exceeded') |
| 466 | self.length = clen |
| 467 | if self.limit is None and clen >= 0: |
| 468 | self.limit = clen |
| 469 | |
| 470 | self.list = self.file = None |
| 471 | self.done = 0 |
| 472 | if ctype == 'application/x-www-form-urlencoded': |
| 473 | self.read_urlencoded() |
| 474 | elif ctype[:10] == 'multipart/': |
| 475 | self.read_multi(environ, keep_blank_values, strict_parsing) |
| 476 | else: |
| 477 | self.read_single() |
| 478 | |
| 479 | def __del__(self): |
| 480 | try: |
| 481 | self.file.close() |
| 482 | except AttributeError: |
| 483 | pass |
| 484 | |
| 485 | def __enter__(self): |
| 486 | return self |
| 487 | |
| 488 | def __exit__(self, *args): |
| 489 | self.file.close() |
| 490 | |
| 491 | def __repr__(self): |
| 492 | """Return a printable representation.""" |
| 493 | return "FieldStorage(%r, %r, %r)" % ( |
| 494 | self.name, self.filename, self.value) |
| 495 | |
| 496 | def __iter__(self): |
| 497 | return iter(self.keys()) |
| 498 | |
| 499 | def __getattr__(self, name): |
| 500 | if name != 'value': |
| 501 | raise AttributeError(name) |
| 502 | if self.file: |
| 503 | self.file.seek(0) |
| 504 | value = self.file.read() |
| 505 | self.file.seek(0) |
| 506 | elif self.list is not None: |
| 507 | value = self.list |
| 508 | else: |
| 509 | value = None |
| 510 | return value |
| 511 | |
| 512 | def __getitem__(self, key): |
| 513 | """Dictionary style indexing.""" |
| 514 | if self.list is None: |
| 515 | raise TypeError("not indexable") |
| 516 | found = [] |
| 517 | for item in self.list: |
| 518 | if item.name == key: found.append(item) |
| 519 | if not found: |
| 520 | raise KeyError(key) |
| 521 | if len(found) == 1: |
| 522 | return found[0] |
| 523 | else: |
| 524 | return found |
| 525 | |
| 526 | def getvalue(self, key, default=None): |
| 527 | """Dictionary style get() method, including 'value' lookup.""" |
| 528 | if key in self: |
| 529 | value = self[key] |
| 530 | if isinstance(value, list): |
| 531 | return [x.value for x in value] |
| 532 | else: |
| 533 | return value.value |
| 534 | else: |
| 535 | return default |
| 536 | |
| 537 | def getfirst(self, key, default=None): |
| 538 | """ Return the first value received.""" |
| 539 | if key in self: |
| 540 | value = self[key] |
| 541 | if isinstance(value, list): |
| 542 | return value[0].value |
| 543 | else: |
| 544 | return value.value |
| 545 | else: |
| 546 | return default |
| 547 | |
| 548 | def getlist(self, key): |
| 549 | """ Return list of received values.""" |
| 550 | if key in self: |
| 551 | value = self[key] |
| 552 | if isinstance(value, list): |
| 553 | return [x.value for x in value] |
| 554 | else: |
| 555 | return [value.value] |
| 556 | else: |
| 557 | return [] |
| 558 | |
| 559 | def keys(self): |
| 560 | """Dictionary style keys() method.""" |
| 561 | if self.list is None: |
| 562 | raise TypeError("not indexable") |
| 563 | return list(set(item.name for item in self.list)) |
| 564 | |
| 565 | def __contains__(self, key): |
| 566 | """Dictionary style __contains__ method.""" |
| 567 | if self.list is None: |
| 568 | raise TypeError("not indexable") |
| 569 | return any(item.name == key for item in self.list) |
| 570 | |
| 571 | def __len__(self): |
| 572 | """Dictionary style len(x) support.""" |
| 573 | return len(self.keys()) |
| 574 | |
| 575 | def __bool__(self): |
| 576 | if self.list is None: |
| 577 | raise TypeError("Cannot be converted to bool.") |
| 578 | return bool(self.list) |
| 579 | |
| 580 | def read_urlencoded(self): |
| 581 | """Internal: read data in query string format.""" |
| 582 | qs = self.fp.read(self.length) |
| 583 | if not isinstance(qs, bytes): |
| 584 | raise ValueError("%s should return bytes, got %s" \ |
| 585 | % (self.fp, type(qs).__name__)) |
| 586 | qs = qs.decode(self.encoding, self.errors) |
| 587 | if self.qs_on_post: |
| 588 | qs += '&' + self.qs_on_post |
| 589 | query = urllib.parse.parse_qsl( |
| 590 | qs, self.keep_blank_values, self.strict_parsing, |
| 591 | encoding=self.encoding, errors=self.errors, |
| 592 | max_num_fields=self.max_num_fields) |
| 593 | self.list = [MiniFieldStorage(key, value) for key, value in query] |
| 594 | self.skip_lines() |
| 595 | |
| 596 | FieldStorageClass = None |
| 597 | |
| 598 | def read_multi(self, environ, keep_blank_values, strict_parsing): |
| 599 | """Internal: read a part that is itself multipart.""" |
| 600 | ib = self.innerboundary |
| 601 | if not valid_boundary(ib): |
| 602 | raise ValueError('Invalid boundary in multipart form: %r' % (ib,)) |
| 603 | self.list = [] |
| 604 | if self.qs_on_post: |
| 605 | query = urllib.parse.parse_qsl( |
| 606 | self.qs_on_post, self.keep_blank_values, self.strict_parsing, |
| 607 | encoding=self.encoding, errors=self.errors, |
| 608 | max_num_fields=self.max_num_fields) |
| 609 | self.list.extend(MiniFieldStorage(key, value) for key, value in query) |
| 610 | |
| 611 | klass = self.FieldStorageClass or self.__class__ |
| 612 | first_line = self.fp.readline() # bytes |
| 613 | if not isinstance(first_line, bytes): |
| 614 | raise ValueError("%s should return bytes, got %s" \ |
| 615 | % (self.fp, type(first_line).__name__)) |
| 616 | self.bytes_read += len(first_line) |
| 617 | |
| 618 | # Ensure that we consume the file until we've hit our inner boundary |
| 619 | while (first_line.strip() != (b"--" + self.innerboundary) and |
| 620 | first_line): |
| 621 | first_line = self.fp.readline() |
| 622 | self.bytes_read += len(first_line) |
| 623 | |
| 624 | # Propagate max_num_fields into the sub class appropriately |
| 625 | max_num_fields = self.max_num_fields |
| 626 | if max_num_fields is not None: |
| 627 | max_num_fields -= len(self.list) |
| 628 | |
| 629 | while True: |
| 630 | parser = FeedParser() |
| 631 | hdr_text = b"" |
| 632 | while True: |
| 633 | data = self.fp.readline() |
| 634 | hdr_text += data |
| 635 | if not data.strip(): |
| 636 | break |
| 637 | if not hdr_text: |
| 638 | break |
| 639 | # parser takes strings, not bytes |
| 640 | self.bytes_read += len(hdr_text) |
| 641 | parser.feed(hdr_text.decode(self.encoding, self.errors)) |
| 642 | headers = parser.close() |
| 643 | |
| 644 | # Some clients add Content-Length for part headers, ignore them |
| 645 | if 'content-length' in headers: |
| 646 | del headers['content-length'] |
| 647 | |
| 648 | limit = None if self.limit is None \ |
| 649 | else self.limit - self.bytes_read |
| 650 | part = klass(self.fp, headers, ib, environ, keep_blank_values, |
| 651 | strict_parsing, limit, |
| 652 | self.encoding, self.errors, max_num_fields) |
| 653 | |
| 654 | if max_num_fields is not None: |
| 655 | max_num_fields -= 1 |
| 656 | if part.list: |
| 657 | max_num_fields -= len(part.list) |
| 658 | if max_num_fields < 0: |
| 659 | raise ValueError('Max number of fields exceeded') |
| 660 | |
| 661 | self.bytes_read += part.bytes_read |
| 662 | self.list.append(part) |
| 663 | if part.done or self.bytes_read >= self.length > 0: |
| 664 | break |
| 665 | self.skip_lines() |
| 666 | |
| 667 | def read_single(self): |
| 668 | """Internal: read an atomic part.""" |
| 669 | if self.length >= 0: |
| 670 | self.read_binary() |
| 671 | self.skip_lines() |
| 672 | else: |
| 673 | self.read_lines() |
| 674 | self.file.seek(0) |
| 675 | |
| 676 | bufsize = 8*1024 # I/O buffering size for copy to file |
| 677 | |
| 678 | def read_binary(self): |
| 679 | """Internal: read binary data.""" |
| 680 | self.file = self.make_file() |
| 681 | todo = self.length |
| 682 | if todo >= 0: |
| 683 | while todo > 0: |
| 684 | data = self.fp.read(min(todo, self.bufsize)) # bytes |
| 685 | if not isinstance(data, bytes): |
| 686 | raise ValueError("%s should return bytes, got %s" |
| 687 | % (self.fp, type(data).__name__)) |
| 688 | self.bytes_read += len(data) |
| 689 | if not data: |
| 690 | self.done = -1 |
| 691 | break |
| 692 | self.file.write(data) |
| 693 | todo = todo - len(data) |
| 694 | |
| 695 | def read_lines(self): |
| 696 | """Internal: read lines until EOF or outerboundary.""" |
| 697 | if self._binary_file: |
| 698 | self.file = self.__file = BytesIO() # store data as bytes for files |
| 699 | else: |
| 700 | self.file = self.__file = StringIO() # as strings for other fields |
| 701 | if self.outerboundary: |
| 702 | self.read_lines_to_outerboundary() |
| 703 | else: |
| 704 | self.read_lines_to_eof() |
| 705 | |
| 706 | def __write(self, line): |
| 707 | """line is always bytes, not string""" |
| 708 | if self.__file is not None: |
| 709 | if self.__file.tell() + len(line) > 1000: |
| 710 | self.file = self.make_file() |
| 711 | data = self.__file.getvalue() |
| 712 | self.file.write(data) |
| 713 | self.__file = None |
| 714 | if self._binary_file: |
| 715 | # keep bytes |
| 716 | self.file.write(line) |
| 717 | else: |
| 718 | # decode to string |
| 719 | self.file.write(line.decode(self.encoding, self.errors)) |
| 720 | |
| 721 | def read_lines_to_eof(self): |
| 722 | """Internal: read lines until EOF.""" |
| 723 | while 1: |
| 724 | line = self.fp.readline(1<<16) # bytes |
| 725 | self.bytes_read += len(line) |
| 726 | if not line: |
| 727 | self.done = -1 |
| 728 | break |
| 729 | self.__write(line) |
| 730 | |
| 731 | def read_lines_to_outerboundary(self): |
| 732 | """Internal: read lines until outerboundary. |
| 733 | Data is read as bytes: boundaries and line ends must be converted |
| 734 | to bytes for comparisons. |
| 735 | """ |
| 736 | next_boundary = b"--" + self.outerboundary |
| 737 | last_boundary = next_boundary + b"--" |
| 738 | delim = b"" |
| 739 | last_line_lfend = True |
| 740 | _read = 0 |
| 741 | while 1: |
| 742 | |
| 743 | if self.limit is not None and 0 <= self.limit <= _read: |
| 744 | break |
| 745 | line = self.fp.readline(1<<16) # bytes |
| 746 | self.bytes_read += len(line) |
| 747 | _read += len(line) |
| 748 | if not line: |
| 749 | self.done = -1 |
| 750 | break |
| 751 | if delim == b"\r": |
| 752 | line = delim + line |
| 753 | delim = b"" |
| 754 | if line.startswith(b"--") and last_line_lfend: |
| 755 | strippedline = line.rstrip() |
| 756 | if strippedline == next_boundary: |
| 757 | break |
| 758 | if strippedline == last_boundary: |
| 759 | self.done = 1 |
| 760 | break |
| 761 | odelim = delim |
| 762 | if line.endswith(b"\r\n"): |
| 763 | delim = b"\r\n" |
| 764 | line = line[:-2] |
| 765 | last_line_lfend = True |
| 766 | elif line.endswith(b"\n"): |
| 767 | delim = b"\n" |
| 768 | line = line[:-1] |
| 769 | last_line_lfend = True |
| 770 | elif line.endswith(b"\r"): |
| 771 | # We may interrupt \r\n sequences if they span the 2**16 |
| 772 | # byte boundary |
| 773 | delim = b"\r" |
| 774 | line = line[:-1] |
| 775 | last_line_lfend = False |
| 776 | else: |
| 777 | delim = b"" |
| 778 | last_line_lfend = False |
| 779 | self.__write(odelim + line) |
| 780 | |
| 781 | def skip_lines(self): |
| 782 | """Internal: skip lines until outer boundary if defined.""" |
| 783 | if not self.outerboundary or self.done: |
| 784 | return |
| 785 | next_boundary = b"--" + self.outerboundary |
| 786 | last_boundary = next_boundary + b"--" |
| 787 | last_line_lfend = True |
| 788 | while True: |
| 789 | line = self.fp.readline(1<<16) |
| 790 | self.bytes_read += len(line) |
| 791 | if not line: |
| 792 | self.done = -1 |
| 793 | break |
| 794 | if line.endswith(b"--") and last_line_lfend: |
| 795 | strippedline = line.strip() |
| 796 | if strippedline == next_boundary: |
| 797 | break |
| 798 | if strippedline == last_boundary: |
| 799 | self.done = 1 |
| 800 | break |
| 801 | last_line_lfend = line.endswith(b'\n') |
| 802 | |
| 803 | def make_file(self): |
| 804 | """Overridable: return a readable & writable file. |
| 805 | |
| 806 | The file will be used as follows: |
| 807 | - data is written to it |
| 808 | - seek(0) |
| 809 | - data is read from it |
| 810 | |
| 811 | The file is opened in binary mode for files, in text mode |
| 812 | for other fields |
| 813 | |
| 814 | This version opens a temporary file for reading and writing, |
| 815 | and immediately deletes (unlinks) it. The trick (on Unix!) is |
| 816 | that the file can still be used, but it can't be opened by |
| 817 | another process, and it will automatically be deleted when it |
| 818 | is closed or when the current process terminates. |
| 819 | |
| 820 | If you want a more permanent file, you derive a class which |
| 821 | overrides this method. If you want a visible temporary file |
| 822 | that is nevertheless automatically deleted when the script |
| 823 | terminates, try defining a __del__ method in a derived class |
| 824 | which unlinks the temporary files you have created. |
| 825 | |
| 826 | """ |
| 827 | if self._binary_file: |
| 828 | return tempfile.TemporaryFile("wb+") |
| 829 | else: |
| 830 | return tempfile.TemporaryFile("w+", |
| 831 | encoding=self.encoding, newline = '\n') |
| 832 | |
| 833 | |
| 834 | # Test/debug code |
| 835 | # =============== |
| 836 | |
| 837 | def test(environ=os.environ): |
| 838 | """Robust test CGI script, usable as main program. |
| 839 | |
| 840 | Write minimal HTTP headers and dump all information provided to |
| 841 | the script in HTML form. |
| 842 | |
| 843 | """ |
| 844 | print("Content-type: text/html") |
| 845 | print() |
| 846 | sys.stderr = sys.stdout |
| 847 | try: |
| 848 | form = FieldStorage() # Replace with other classes to test those |
| 849 | print_directory() |
| 850 | print_arguments() |
| 851 | print_form(form) |
| 852 | print_environ(environ) |
| 853 | print_environ_usage() |
| 854 | def f(): |
| 855 | exec("testing print_exception() -- <I>italics?</I>") |
| 856 | def g(f=f): |
| 857 | f() |
| 858 | print("<H3>What follows is a test, not an actual exception:</H3>") |
| 859 | g() |
| 860 | except: |
| 861 | print_exception() |
| 862 | |
| 863 | print("<H1>Second try with a small maxlen...</H1>") |
| 864 | |
| 865 | global maxlen |
| 866 | maxlen = 50 |
| 867 | try: |
| 868 | form = FieldStorage() # Replace with other classes to test those |
| 869 | print_directory() |
| 870 | print_arguments() |
| 871 | print_form(form) |
| 872 | print_environ(environ) |
| 873 | except: |
| 874 | print_exception() |
| 875 | |
| 876 | def print_exception(type=None, value=None, tb=None, limit=None): |
| 877 | if type is None: |
| 878 | type, value, tb = sys.exc_info() |
| 879 | import traceback |
| 880 | print() |
| 881 | print("<H3>Traceback (most recent call last):</H3>") |
| 882 | list = traceback.format_tb(tb, limit) + \ |
| 883 | traceback.format_exception_only(type, value) |
| 884 | print("<PRE>%s<B>%s</B></PRE>" % ( |
| 885 | html.escape("".join(list[:-1])), |
| 886 | html.escape(list[-1]), |
| 887 | )) |
| 888 | del tb |
| 889 | |
| 890 | def print_environ(environ=os.environ): |
| 891 | """Dump the shell environment as HTML.""" |
| 892 | keys = sorted(environ.keys()) |
| 893 | print() |
| 894 | print("<H3>Shell Environment:</H3>") |
| 895 | print("<DL>") |
| 896 | for key in keys: |
| 897 | print("<DT>", html.escape(key), "<DD>", html.escape(environ[key])) |
| 898 | print("</DL>") |
| 899 | print() |
| 900 | |
| 901 | def print_form(form): |
| 902 | """Dump the contents of a form as HTML.""" |
| 903 | keys = sorted(form.keys()) |
| 904 | print() |
| 905 | print("<H3>Form Contents:</H3>") |
| 906 | if not keys: |
| 907 | print("<P>No form fields.") |
| 908 | print("<DL>") |
| 909 | for key in keys: |
| 910 | print("<DT>" + html.escape(key) + ":", end=' ') |
| 911 | value = form[key] |
| 912 | print("<i>" + html.escape(repr(type(value))) + "</i>") |
| 913 | print("<DD>" + html.escape(repr(value))) |
| 914 | print("</DL>") |
| 915 | print() |
| 916 | |
| 917 | def print_directory(): |
| 918 | """Dump the current directory as HTML.""" |
| 919 | print() |
| 920 | print("<H3>Current Working Directory:</H3>") |
| 921 | try: |
| 922 | pwd = os.getcwd() |
| 923 | except OSError as msg: |
| 924 | print("OSError:", html.escape(str(msg))) |
| 925 | else: |
| 926 | print(html.escape(pwd)) |
| 927 | print() |
| 928 | |
| 929 | def print_arguments(): |
| 930 | print() |
| 931 | print("<H3>Command Line Arguments:</H3>") |
| 932 | print() |
| 933 | print(sys.argv) |
| 934 | print() |
| 935 | |
| 936 | def print_environ_usage(): |
| 937 | """Dump a list of environment variables used by CGI as HTML.""" |
| 938 | print(""" |
| 939 | <H3>These environment variables could have been set:</H3> |
| 940 | <UL> |
| 941 | <LI>AUTH_TYPE |
| 942 | <LI>CONTENT_LENGTH |
| 943 | <LI>CONTENT_TYPE |
| 944 | <LI>DATE_GMT |
| 945 | <LI>DATE_LOCAL |
| 946 | <LI>DOCUMENT_NAME |
| 947 | <LI>DOCUMENT_ROOT |
| 948 | <LI>DOCUMENT_URI |
| 949 | <LI>GATEWAY_INTERFACE |
| 950 | <LI>LAST_MODIFIED |
| 951 | <LI>PATH |
| 952 | <LI>PATH_INFO |
| 953 | <LI>PATH_TRANSLATED |
| 954 | <LI>QUERY_STRING |
| 955 | <LI>REMOTE_ADDR |
| 956 | <LI>REMOTE_HOST |
| 957 | <LI>REMOTE_IDENT |
| 958 | <LI>REMOTE_USER |
| 959 | <LI>REQUEST_METHOD |
| 960 | <LI>SCRIPT_NAME |
| 961 | <LI>SERVER_NAME |
| 962 | <LI>SERVER_PORT |
| 963 | <LI>SERVER_PROTOCOL |
| 964 | <LI>SERVER_ROOT |
| 965 | <LI>SERVER_SOFTWARE |
| 966 | </UL> |
| 967 | In addition, HTTP headers sent by the server may be passed in the |
| 968 | environment as well. Here are some common variable names: |
| 969 | <UL> |
| 970 | <LI>HTTP_ACCEPT |
| 971 | <LI>HTTP_CONNECTION |
| 972 | <LI>HTTP_HOST |
| 973 | <LI>HTTP_PRAGMA |
| 974 | <LI>HTTP_REFERER |
| 975 | <LI>HTTP_USER_AGENT |
| 976 | </UL> |
| 977 | """) |
| 978 | |
| 979 | |
| 980 | # Utilities |
| 981 | # ========= |
| 982 | |
| 983 | def valid_boundary(s): |
| 984 | import re |
| 985 | if isinstance(s, bytes): |
| 986 | _vb_pattern = b"^[ -~]{0,200}[!-~]$" |
| 987 | else: |
| 988 | _vb_pattern = "^[ -~]{0,200}[!-~]$" |
| 989 | return re.match(_vb_pattern, s) |
| 990 | |
| 991 | # Invoke mainline |
| 992 | # =============== |
| 993 | |
| 994 | # Call test() when this file is run as a script (not imported as a module) |
| 995 | if __name__ == '__main__': |
| 996 | test() |