Blame - Lib/http/server.py - platform/external/python/cpython3

2008-05-26 16:32:26 +0000

[diff] [blame]

1

"""HTTP server classes.

2

3

Note: BaseHTTPRequestHandler doesn't implement any HTTP request; see

4

SimpleHTTPRequestHandler for simple implementations of GET, HEAD and POST,

5

and CGIHTTPRequestHandler for CGI scripts.

6

7

It does, however, optionally implement HTTP/1.1 persistent connections,

8

as of version 0.3.

9

10

Notes on CGIHTTPRequestHandler

11

------------------------------

12

13

This class implements GET and POST requests to cgi-bin scripts.

14

15

If the os.fork() function is not present (e.g. on Windows),

Amaury Forgeot d'Arc

2008-06-18 22:19:22 +0000

[diff] [blame]

16

subprocess.Popen() is used as a fallback, with slightly altered semantics.

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

17

18

In all cases, the implementation is intentionally naive -- all

19

requests are executed synchronously.

20

21

SECURITY WARNING: DON'T USE THIS CODE UNLESS YOU ARE INSIDE A FIREWALL

22

-- it may execute arbitrary Python code or external programs.

23

24

Note that status code 200 is sent prior to execution of a CGI script, so

25

scripts cannot send other status codes such as 302 (redirect).

XXX To do:

- log requests even later (to capture byte count)

30

- log user-agent header and other interesting goodies

31

- send error log to separate file

"""

# See also:

#

# HTTP Working Group T. Berners-Lee

38

# INTERNET-DRAFT R. T. Fielding

39

# <draft-ietf-http-v10-spec-00.txt> H. Frystyk Nielsen

40

# Expires September 8, 1995 March 8, 1995

41

#

42

# URL: http://www.ics.uci.edu/pub/ietf/http/draft-ietf-http-v10-spec-00.txt

#

# and

#

# Network Working Group R. Fielding

47

# Request for Comments: 2616 et al

48

# Obsoletes: 2068 June 1999

49

# Category: Standards Track

50

#

51

# URL: http://www.faqs.org/rfcs/rfc2616.html

# Log files

# ---------

#

# Here's a quote from the NCSA httpd docs about log file format.

57

#

58

# | The logfile format is as follows. Each line consists of:

59

# |

60

# | host rfc931 authuser [DD/Mon/YYYY:hh:mm:ss] "request" ddd bbbb

61

# |

62

# | host: Either the DNS name or the IP number of the remote client

63

# | rfc931: Any information returned by identd for this person,

64

# | - otherwise.

65

# | authuser: If user sent a userid for authentication, the user name,

66

# | - otherwise.

67

# | DD: Day

68

# | Mon: Month (calendar name)

69

# | YYYY: Year

70

# | hh: hour (24-hour format, the machine's timezone)

71

# | mm: minutes

72

# | ss: seconds

73

# | request: The first line of the HTTP request as sent by the client.

74

# | ddd: the status code returned by the server, - if not available.

75

# | bbbb: the total number of bytes sent,

76

# | *not including the HTTP/1.0 header*, - if not available

77

# |

78

# | You can determine the name of the file accessed through request.

79

#

80

# (Actually, the latter is only true if you know the server configuration

81

# at the time the request was made!)

__version__ = "0.6"

Berker Peksag

2015-02-13 20:48:15 +0200

[diff] [blame]

85

__all__ = [

86

"HTTPServer", "BaseHTTPRequestHandler",

87

"SimpleHTTPRequestHandler", "CGIHTTPRequestHandler",

88

]

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

89

Georg Brandl

1f7fffb

2010-10-15 15:57:45 +0000

[diff] [blame]

90

import html

Jeremy Hylton

914ab45

2009-03-27 17:16:06 +0000

[diff] [blame]

import http.client

import io

import mimetypes

import os

import posixpath

import select

import shutil

import socket # For gethostbyaddr()

import socketserver

import sys

import time

import urllib.parse

Senthil Kumaran

2010-10-03 17:55:45 +0000

[diff] [blame]

103

import copy

Senthil Kumaran

2012-06-03 16:15:54 +0800

[diff] [blame]

104

import argparse

105

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

106

from http import HTTPStatus

107

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

108

109

# Default error message template

110

DEFAULT_ERROR_MESSAGE = """\

Senthil Kumaran

1b407fe

2011-03-20 10:44:30 +0800

[diff] [blame]

111

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"

112

"http://www.w3.org/TR/html4/strict.dtd">

Ezio Melotti

2011-11-02 19:33:29 +0200

[diff] [blame]

113

<html>

Senthil Kumaran

b253c9f

2011-03-17 16:43:22 +0800

[diff] [blame]

114

<head>

Senthil Kumaran

1b407fe

2011-03-20 10:44:30 +0800

[diff] [blame]

115

Senthil Kumaran

b253c9f

2011-03-17 16:43:22 +0800

[diff] [blame]

116

<title>Error response</title>

117

</head>

118

<body>

119

<h1>Error response</h1>

120

<p>Error code: %(code)d</p>

121

<p>Message: %(message)s.</p>

122

<p>Error code explanation: %(code)s - %(explain)s.</p>

123

</body>

124

</html>

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

125

"""

126

127

DEFAULT_ERROR_CONTENT_TYPE = "text/html;charset=utf-8"

128

129

def _quote_html(html):

130

return html.replace("&", "&").replace("<", "<").replace(">", ">")

131

132

class HTTPServer(socketserver.TCPServer):

133

134

allow_reuse_address = 1 # Seems to make sense in testing environment

135

136

def server_bind(self):

137

"""Override server_bind to store the server name."""

138

socketserver.TCPServer.server_bind(self)

139

host, port = self.socket.getsockname()[:2]

140

self.server_name = socket.getfqdn(host)

141

self.server_port = port

142

143

144

class BaseHTTPRequestHandler(socketserver.StreamRequestHandler):

145

146

"""HTTP request handler base class.

147

148

The following explanation of HTTP serves to guide you through the

149

code as well as to expose any misunderstandings I may have about

150

HTTP (so you don't need to read the code to figure out I'm wrong

151

:-).

152

153

HTTP (HyperText Transfer Protocol) is an extensible protocol on

154

top of a reliable stream transport (e.g. TCP/IP). The protocol

155

recognizes three parts to a request:

156

157

1. One line identifying the request type and path

158

2. An optional set of RFC-822-style headers

159

3. An optional data part

160

161

The headers and data are separated by a blank line.

162

163

The first line of the request has the form

where <command> is a (case-sensitive) keyword such as GET or POST,

168

<path> is a string containing path information for the request,

169

and <version> should be the string "HTTP/1.0" or "HTTP/1.1".

170

<path> is encoded using the URL encoding scheme (using %xx to signify

171

the ASCII character with hex code xx).

172

173

The specification specifies that lines are separated by CRLF but

174

for compatibility with the widest range of clients recommends

175

servers also handle LF. Similarly, whitespace in the request line

176

is treated sensibly (allowing multiple spaces between components

177

and allowing trailing whitespace).

178

179

Similarly, for output, lines ought to be separated by CRLF pairs

180

but most clients grok LF characters just fine.

181

182

If the first line of the request has the form

(i.e. <version> is left out) then this is assumed to be an HTTP

187

0.9 request; this form has no optional headers and data part and

188

the reply consists of just the data.

189

190

The reply form of the HTTP 1.x protocol again has three parts:

191

192

1. One line giving the response code

193

2. An optional set of RFC-822-style headers

194

3. The data

195

196

Again, the headers and data are separated by a blank line.

197

198

The response code line has the form

where <version> is the protocol version ("HTTP/1.0" or "HTTP/1.1"),

203

<responsecode> is a 3-digit response code indicating success or

204

failure of the request, and <responsestring> is an optional

205

human-readable string explaining what the response code means.

206

207

This server parses the request and the headers, and then calls a

208

function specific to the request type (<command>). Specifically,

209

a request SPAM will be handled by a method do_SPAM(). If no

210

such method exists the server sends an error response to the

211

client. If it exists, it is called with no arguments:

do_SPAM()

Note that the request name is case sensitive (i.e. SPAM and spam

216

are different requests).

217

218

The various request details are stored in instance variables:

219

220

- client_address is the client IP address in the form (host,

221

port);

222

223

- command, path and version are the broken-down request line;

224

Barry Warsaw

2008-06-12 04:06:45 +0000

[diff] [blame]

225

- headers is an instance of email.message.Message (or a derived

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

226

class) containing the header information;

227

228

- rfile is a file object open for reading positioned at the

229

start of the optional input data part;

230

231

- wfile is a file object open for writing.

232

233

IT IS IMPORTANT TO ADHERE TO THE PROTOCOL FOR WRITING!

234

235

The first thing to be written must be the response line. Then

236

follow 0 or more header lines, then a blank line, and then the

237

actual data (if any). The meaning of the header lines depends on

238

the command executed by the server; in most cases, when data is

239

returned, there should be at least one header line of the form

240

241

Content-type: <type>/<subtype>

242

243

where <type> and <subtype> should be registered MIME types,

244

e.g. "text/html" or "text/plain".

"""

# The Python system version, truncated to its first component.

249

sys_version = "Python/" + sys.version.split()[0]

250

251

# The server software version. You may want to override this.

252

# The format is multiple whitespace-separated strings,

253

# where each string is of the form name[/version].

254

server_version = "BaseHTTP/" + __version__

255

256

error_message_format = DEFAULT_ERROR_MESSAGE

257

error_content_type = DEFAULT_ERROR_CONTENT_TYPE

258

259

# The default request version. This only affects responses up until

260

# the point where the request line is parsed, so it mainly decides what

261

# the client gets back when sending a malformed request line.

262

# Most web servers default to HTTP 0.9, i.e. don't send a status line.

263

default_request_version = "HTTP/0.9"

264

265

def parse_request(self):

266

"""Parse a request (internal).

267

268

The request should be stored in self.raw_requestline; the results

269

are in self.command, self.path, self.request_version and

270

self.headers.

271

272

Return True for success, False for failure; on failure, an

error is sent back.

"""

self.command = None # set in case of error on the first line

277

self.request_version = version = self.default_request_version

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

278

self.close_connection = True

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

279

requestline = str(self.raw_requestline, 'iso-8859-1')

Senthil Kumaran

3075549

2011-12-23 17:03:41 +0800

[diff] [blame]

280

requestline = requestline.rstrip('\r\n')

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

281

self.requestline = requestline

282

words = requestline.split()

283

if len(words) == 3:

Senthil Kumaran

3075549

2011-12-23 17:03:41 +0800

[diff] [blame]

284

command, path, version = words

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

285

if version[:5] != 'HTTP/':

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

286

self.send_error(

287

HTTPStatus.BAD_REQUEST,

288

"Bad request version (%r)" % version)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

289

return False

290

try:

291

base_version_number = version.split('/', 1)[1]

292

version_number = base_version_number.split(".")

293

# RFC 2145 section 3.1 says there can be only one "." and

294

# - major and minor numbers MUST be treated as

295

# separate integers;

296

# - HTTP/2.4 is a lower version than HTTP/2.13, which in

297

# turn is lower than HTTP/12.3;

298

# - Leading zeros MUST be ignored by recipients.

299

if len(version_number) != 2:

300

raise ValueError

301

version_number = int(version_number[0]), int(version_number[1])

302

except (ValueError, IndexError):

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

303

self.send_error(

304

HTTPStatus.BAD_REQUEST,

305

"Bad request version (%r)" % version)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

306

return False

307

if version_number >= (1, 1) and self.protocol_version >= "HTTP/1.1":

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

308

self.close_connection = False

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

309

if version_number >= (2, 0):

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

310

self.send_error(

311

HTTPStatus.HTTP_VERSION_NOT_SUPPORTED,

312

"Invalid HTTP Version (%s)" % base_version_number)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

313

return False

314

elif len(words) == 2:

Senthil Kumaran

3075549

2011-12-23 17:03:41 +0800

[diff] [blame]

315

command, path = words

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

316

self.close_connection = True

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

317

if command != 'GET':

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

318

self.send_error(

319

HTTPStatus.BAD_REQUEST,

320

"Bad HTTP/0.9 request type (%r)" % command)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

return False

elif not words:

return False

else:

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

325

self.send_error(

326

HTTPStatus.BAD_REQUEST,

327

"Bad request syntax (%r)" % requestline)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

328

return False

329

self.command, self.path, self.request_version = command, path, version

330

331

# Examine the headers and look for a Connection directive.

Senthil Kumaran

5466bf1

2010-12-18 16:55:23 +0000

[diff] [blame]

332

try:

333

self.headers = http.client.parse_headers(self.rfile,

334

_class=self.MessageClass)

335

except http.client.LineTooLong:

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

336

self.send_error(

337

HTTPStatus.BAD_REQUEST,

338

"Line too long")

Senthil Kumaran

5466bf1

2010-12-18 16:55:23 +0000

[diff] [blame]

339

return False

Martin Panter

acc0319

2016-04-03 00:45:46 +0000

[diff] [blame]

340

except http.client.HTTPException as err:

341

self.send_error(

342

HTTPStatus.REQUEST_HEADER_FIELDS_TOO_LARGE,

"Too many headers",

str(err)

)

return False

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

347

348

conntype = self.headers.get('Connection', "")

349

if conntype.lower() == 'close':

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

350

self.close_connection = True

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

351

elif (conntype.lower() == 'keep-alive' and

352

self.protocol_version >= "HTTP/1.1"):

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

353

self.close_connection = False

Senthil Kumaran

0f476d4

2010-09-30 06:09:18 +0000

[diff] [blame]

354

# Examine the headers and look for an Expect directive

355

expect = self.headers.get('Expect', "")

356

if (expect.lower() == "100-continue" and

357

self.protocol_version >= "HTTP/1.1" and

358

self.request_version >= "HTTP/1.1"):

359

if not self.handle_expect_100():

return False

return True

def handle_expect_100(self):

364

"""Decide what to do with an "Expect: 100-continue" header.

365

366

If the client is expecting a 100 Continue response, we must

367

respond with either a 100 Continue or a final response before

368

waiting for the request body. The default is to always respond

369

with a 100 Continue. You can behave differently (for example,

370

reject unauthorized requests) by overriding this method.

371

372

This method should either return True (possibly after sending

373

a 100 Continue response) or send an error response and return

374

False.

375

376

"""

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

377

self.send_response_only(HTTPStatus.CONTINUE)

Benjamin Peterson

0442423

2014-01-18 21:50:18 -0500

[diff] [blame]

378

self.end_headers()

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

379

return True

380

381

def handle_one_request(self):

382

"""Handle a single HTTP request.

383

384

You normally don't need to override this method; see the class

385

__doc__ string for information on how to handle specific HTTP

386

commands such as GET and POST.

387

388

"""

Kristján Valur Jónsson

2009-07-01 10:01:31 +0000

[diff] [blame]

389

try:

Antoine Pitrou

c492437

2010-12-16 16:48:36 +0000

[diff] [blame]

390

self.raw_requestline = self.rfile.readline(65537)

391

if len(self.raw_requestline) > 65536:

392

self.requestline = ''

393

self.request_version = ''

394

self.command = ''

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

395

self.send_error(HTTPStatus.REQUEST_URI_TOO_LONG)

Antoine Pitrou

c492437

2010-12-16 16:48:36 +0000

[diff] [blame]

396

return

Kristján Valur Jónsson

2009-07-01 10:01:31 +0000

[diff] [blame]

397

if not self.raw_requestline:

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

398

self.close_connection = True

Kristján Valur Jónsson

2009-07-01 10:01:31 +0000

[diff] [blame]

399

return

400

if not self.parse_request():

401

# An error code has been sent, just exit

402

return

403

mname = 'do_' + self.command

404

if not hasattr(self, mname):

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

405

self.send_error(

406

HTTPStatus.NOT_IMPLEMENTED,

407

"Unsupported method (%r)" % self.command)

Kristján Valur Jónsson

2009-07-01 10:01:31 +0000

[diff] [blame]

408

return

409

method = getattr(self, mname)

410

method()

411

self.wfile.flush() #actually send the response if not already done.

412

except socket.timeout as e:

413

#a read or a write timed out. Discard this connection

414

self.log_error("Request timed out: %r", e)

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

415

self.close_connection = True

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

416

return

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

417

418

def handle(self):

419

"""Handle multiple requests if necessary."""

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

420

self.close_connection = True

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

421

422

self.handle_one_request()

423

while not self.close_connection:

424

self.handle_one_request()

425

Senthil Kumaran

2013-03-15 07:53:21 -0700

[diff] [blame]

426

def send_error(self, code, message=None, explain=None):

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

427

"""Send and log an error reply.

428

Senthil Kumaran

2013-03-15 07:53:21 -0700

[diff] [blame]

429

Arguments are

430

* code: an HTTP error code

431

3 digits

432

* message: a simple optional 1 line reason phrase.

433

*( HTAB / SP / VCHAR / %x80-FF )

434

defaults to short entry matching the response code

435

* explain: a detailed message defaults to the long entry

436

matching the response code.

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

437

438

This sends an error response (so it must be called before any

439

output has been generated), logs the error, and finally sends

440

a piece of HTML explaining the error to the user.

"""

try:

shortmsg, longmsg = self.responses[code]

446

except KeyError:

447

shortmsg, longmsg = '???', '???'

448

if message is None:

449

message = shortmsg

Senthil Kumaran

2013-03-15 07:53:21 -0700

[diff] [blame]

450

if explain is None:

451

explain = longmsg

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

452

self.log_error("code %d, message %s", code, message)

453

# using _quote_html to prevent Cross Site Scripting attacks (see bug #1100201)

454

content = (self.error_message_format %

Senthil Kumaran

2013-03-15 07:53:21 -0700

[diff] [blame]

455

{'code': code, 'message': _quote_html(message), 'explain': _quote_html(explain)})

Senthil Kumaran

52d2720

2012-10-10 23:16:21 -0700

[diff] [blame]

456

body = content.encode('UTF-8', 'replace')

Senthil Kumaran

1e7551d

2013-03-05 02:25:58 -0800

[diff] [blame]

457

self.send_response(code, message)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

458

self.send_header("Content-Type", self.error_content_type)

459

self.send_header('Connection', 'close')

Senthil Kumaran

52d2720

2012-10-10 23:16:21 -0700

[diff] [blame]

460

self.send_header('Content-Length', int(len(body)))

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

461

self.end_headers()

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

462

463

if (self.command != 'HEAD' and

464

code >= 200 and

465

code not in (

466

HTTPStatus.NO_CONTENT, HTTPStatus.NOT_MODIFIED)):

Senthil Kumaran

52d2720

2012-10-10 23:16:21 -0700

[diff] [blame]

467

self.wfile.write(body)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

468

469

def send_response(self, code, message=None):

Senthil Kumaran

2011-05-09 23:25:02 +0800

[diff] [blame]

470

"""Add the response header to the headers buffer and log the

471

response code.

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

472

473

Also send two standard headers with the server software

474

version and the current date.

475

476

"""

477

self.log_request(code)

Senthil Kumaran

0f476d4

2010-09-30 06:09:18 +0000

[diff] [blame]

478

self.send_response_only(code, message)

479

self.send_header('Server', self.version_string())

480

self.send_header('Date', self.date_time_string())

481

482

def send_response_only(self, code, message=None):

483

"""Send the response header only."""

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

484

if message is None:

485

if code in self.responses:

486

message = self.responses[code][0]

487

else:

488

message = ''

489

if self.request_version != 'HTTP/0.9':

Senthil Kumaran

2011-05-09 23:25:02 +0800

[diff] [blame]

490

if not hasattr(self, '_headers_buffer'):

491

self._headers_buffer = []

492

self._headers_buffer.append(("%s %d %s\r\n" %

493

(self.protocol_version, code, message)).encode(

494

'latin-1', 'strict'))

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

495

496

def send_header(self, keyword, value):

Senthil Kumaran

2011-05-09 23:25:02 +0800

[diff] [blame]

497

"""Send a MIME header to the headers buffer."""

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

498

if self.request_version != 'HTTP/0.9':

Senthil Kumaran

e4dad4f

2010-11-21 14:36:14 +0000

[diff] [blame]

499

if not hasattr(self, '_headers_buffer'):

500

self._headers_buffer = []

501

self._headers_buffer.append(

Marc-André Lemburg

8f36af7

2011-02-25 15:42:01 +0000

[diff] [blame]

502

("%s: %s\r\n" % (keyword, value)).encode('latin-1', 'strict'))

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

503

504

if keyword.lower() == 'connection':

505

if value.lower() == 'close':

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

506

self.close_connection = True

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

507

elif value.lower() == 'keep-alive':

Benjamin Peterson

2015-02-17 21:11:10 -0500

[diff] [blame]

508

self.close_connection = False

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

509

510

def end_headers(self):

511

"""Send the blank line ending the MIME headers."""

512

if self.request_version != 'HTTP/0.9':

Senthil Kumaran

e4dad4f

2010-11-21 14:36:14 +0000

[diff] [blame]

513

self._headers_buffer.append(b"\r\n")

Senthil Kumaran

2011-05-09 23:25:02 +0800

[diff] [blame]

514

self.flush_headers()

515

516

def flush_headers(self):

517

if hasattr(self, '_headers_buffer'):

Senthil Kumaran

e4dad4f

2010-11-21 14:36:14 +0000

[diff] [blame]

518

self.wfile.write(b"".join(self._headers_buffer))

519

self._headers_buffer = []

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

520

521

def log_request(self, code='-', size='-'):

522

"""Log an accepted request.

523

524

This is called by send_response().

525

526

"""

Serhiy Storchaka

c0a23e6

2015-03-07 11:51:37 +0200

[diff] [blame]

527

if isinstance(code, HTTPStatus):

528

code = code.value

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

529

self.log_message('"%s" %s %s',

530

self.requestline, str(code), str(size))

531

532

def log_error(self, format, *args):

533

"""Log an error.

534

535

This is called when a request cannot be fulfilled. By

536

default it passes the message on to log_message().

537

538

Arguments are the same as for log_message().

539

540

XXX This should go to the separate error log.

"""

self.log_message(format, *args)

545

546

def log_message(self, format, *args):

547

"""Log an arbitrary message.

548

549

This is used by all other logging functions. Override

550

it if you have specific logging wishes.

551

552

The first argument, FORMAT, is a format string for the

553

message to be logged. If the format string contains

554

any % escapes requiring parameters, they should be

555

specified as subsequent arguments (it's just like

556

printf!).

557

Senthil Kumaran

db727b4

2012-04-29 13:41:03 +0800

[diff] [blame]

558

The client ip and current date/time are prefixed to

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

every message.

"""

sys.stderr.write("%s - - [%s] %s\n" %

564

(self.address_string(),

565

self.log_date_time_string(),

566

format%args))

567

568

def version_string(self):

569

"""Return the server software version string."""

570

return self.server_version + ' ' + self.sys_version

571

572

def date_time_string(self, timestamp=None):

573

"""Return the current date and time formatted for a message header."""

574

if timestamp is None:

575

timestamp = time.time()

576

year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)

577

s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (

578

self.weekdayname[wd],

579

day, self.monthname[month], year,

hh, mm, ss)

return s

def log_date_time_string(self):

584

"""Return the current time formatted for logging."""

585

now = time.time()

586

year, month, day, hh, mm, ss, x, y, z = time.localtime(now)

587

s = "%02d/%3s/%04d %02d:%02d:%02d" % (

588

day, self.monthname[month], year, hh, mm, ss)

589

return s

590

591

weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']

592

593

monthname = [None,

594

'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',

595

'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']

596

597

def address_string(self):

Senthil Kumaran

1aacba4

2012-04-29 12:51:54 +0800

[diff] [blame]

598

"""Return the client address."""

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

599

Senthil Kumaran

1aacba4

2012-04-29 12:51:54 +0800

[diff] [blame]

600

return self.client_address[0]

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

601

602

# Essentially static class variables

603

604

# The version of the HTTP protocol we support.

605

# Set this to HTTP/1.1 to enable automatic keepalive

606

protocol_version = "HTTP/1.0"

607

Barry Warsaw

2008-06-12 04:06:45 +0000

[diff] [blame]

608

# MessageClass used to parse headers

Barry Warsaw

2008-06-12 04:06:45 +0000

[diff] [blame]

609

MessageClass = http.client.HTTPMessage

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

610

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

611

# hack to maintain backwards compatibility

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

612

responses = {

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

613

v: (v.phrase, v.description)

614

for v in HTTPStatus.__members__.values()

615

}

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

616

617

618

class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):

619

620

"""Simple HTTP request handler with GET and HEAD commands.

621

622

This serves files from the current directory and any of its

623

subdirectories. The MIME type for files is determined by

624

calling the .guess_type() method.

625

626

The GET and HEAD requests are identical except that the HEAD

627

request omits the actual contents of the file.

"""

server_version = "SimpleHTTP/" + __version__

632

633

def do_GET(self):

634

"""Serve a GET request."""

635

f = self.send_head()

636

if f:

Serhiy Storchaka

91b0bc2

2014-01-25 19:43:02 +0200

[diff] [blame]

637

try:

638

self.copyfile(f, self.wfile)

639

finally:

640

f.close()

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

641

642

def do_HEAD(self):

643

"""Serve a HEAD request."""

f = self.send_head()

if f:

f.close()

def send_head(self):

"""Common code for GET and HEAD commands.

650

651

This sends the response code and MIME headers.

652

653

Return value is either a file object (which has to be copied

654

to the outputfile by the caller unless the command was HEAD,

655

and must be closed by the caller under all circumstances), or

656

None, in which case the caller has nothing further to do.

657

658

"""

659

path = self.translate_path(self.path)

660

f = None

661

if os.path.isdir(path):

Benjamin Peterson

94cb7a2

2014-12-26 10:53:43 -0600

[diff] [blame]

662

parts = urllib.parse.urlsplit(self.path)

663

if not parts.path.endswith('/'):

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

664

# redirect browser - doing basically what apache does

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

665

self.send_response(HTTPStatus.MOVED_PERMANENTLY)

Benjamin Peterson

94cb7a2

2014-12-26 10:53:43 -0600

[diff] [blame]

666

new_parts = (parts[0], parts[1], parts[2] + '/',

667

parts[3], parts[4])

668

new_url = urllib.parse.urlunsplit(new_parts)

669

self.send_header("Location", new_url)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

670

self.end_headers()

671

return None

672

for index in "index.html", "index.htm":

673

index = os.path.join(path, index)

674

if os.path.exists(index):

path = index

break

else:

return self.list_directory(path)

679

ctype = self.guess_type(path)

680

try:

681

f = open(path, 'rb')

Andrew Svetlov

f7a17b4

2012-12-25 16:47:37 +0200

[diff] [blame]

682

except OSError:

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

683

self.send_error(HTTPStatus.NOT_FOUND, "File not found")

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

684

return None

Serhiy Storchaka

91b0bc2

2014-01-25 19:43:02 +0200

[diff] [blame]

685

try:

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

686

self.send_response(HTTPStatus.OK)

Serhiy Storchaka

91b0bc2

2014-01-25 19:43:02 +0200

[diff] [blame]

687

self.send_header("Content-type", ctype)

688

fs = os.fstat(f.fileno())

689

self.send_header("Content-Length", str(fs[6]))

690

self.send_header("Last-Modified", self.date_time_string(fs.st_mtime))

self.end_headers()

return f

except:

f.close()

raise

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

696

697

def list_directory(self, path):

698

"""Helper to produce a directory listing (absent index.html).

699

700

Return value is either a file object, or None (indicating an

701

error). In either case, the headers are sent, making the

702

interface the same as for send_head().

"""

try:

list = os.listdir(path)

Andrew Svetlov

ad28c7f

2012-12-18 22:02:39 +0200

[diff] [blame]

707

except OSError:

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

708

self.send_error(

709

HTTPStatus.NOT_FOUND,

710

"No permission to list directory")

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

711

return None

712

list.sort(key=lambda a: a.lower())

713

r = []

Serhiy Storchaka

2014-08-17 08:22:11 +0300

[diff] [blame]

714

try:

715

displaypath = urllib.parse.unquote(self.path,

716

errors='surrogatepass')

717

except UnicodeDecodeError:

718

displaypath = urllib.parse.unquote(path)

719

displaypath = html.escape(displaypath)

Ezio Melotti

2011-11-02 19:33:29 +0200

[diff] [blame]

720

enc = sys.getfilesystemencoding()

721

title = 'Directory listing for %s' % displaypath

722

r.append('<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" '

723

'"http://www.w3.org/TR/html4/strict.dtd">')

724

r.append('<html>\n<head>')

725

r.append('<meta http-equiv="Content-Type" '

726

'content="text/html; charset=%s">' % enc)

727

r.append('<title>%s</title>\n</head>' % title)

728

r.append('<body>\n<h1>%s</h1>' % title)

729

r.append('<hr>\n<ul>')

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

730

for name in list:

731

fullname = os.path.join(path, name)

732

displayname = linkname = name

733

# Append / for directories or @ for symbolic links

734

if os.path.isdir(fullname):

735

displayname = name + "/"

736

linkname = name + "/"

737

if os.path.islink(fullname):

738

displayname = name + "@"

739

# Note: a link to a directory displays with @ and links with /

Ezio Melotti

2011-11-02 19:33:29 +0200

[diff] [blame]

740

r.append('<li><a href="%s">%s</a></li>'

Serhiy Storchaka

2014-08-17 08:22:11 +0300

[diff] [blame]

741

% (urllib.parse.quote(linkname,

742

errors='surrogatepass'),

743

html.escape(displayname)))

Ezio Melotti

2011-11-02 19:33:29 +0200

[diff] [blame]

744

r.append('</ul>\n<hr>\n</body>\n</html>\n')

Serhiy Storchaka

2014-08-17 08:22:11 +0300

[diff] [blame]

745

encoded = '\n'.join(r).encode(enc, 'surrogateescape')

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

746

f = io.BytesIO()

747

f.write(encoded)

748

f.seek(0)

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

749

self.send_response(HTTPStatus.OK)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

750

self.send_header("Content-type", "text/html; charset=%s" % enc)

751

self.send_header("Content-Length", str(len(encoded)))

self.end_headers()

return f

def translate_path(self, path):

756

"""Translate a /-separated PATH to the local filename syntax.

757

758

Components that mean special things to the local file system

759

(e.g. drive or directory names) are ignored. (XXX They should

760

probably be diagnosed.)

761

762

"""

763

# abandon query parameters

764

path = path.split('?',1)[0]

765

path = path.split('#',1)[0]

Senthil Kumaran

72c238e

2013-09-13 00:21:18 -0700

[diff] [blame]

766

# Don't forget explicit trailing slash when normalizing. Issue17324

Senthil Kumaran

600b735

2013-09-29 18:59:04 -0700

[diff] [blame]

767

trailing_slash = path.rstrip().endswith('/')

Serhiy Storchaka

2014-08-17 08:22:11 +0300

[diff] [blame]

768

try:

769

path = urllib.parse.unquote(path, errors='surrogatepass')

770

except UnicodeDecodeError:

771

path = urllib.parse.unquote(path)

772

path = posixpath.normpath(path)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

773

words = path.split('/')

774

words = filter(None, words)

775

path = os.getcwd()

776

for word in words:

Martin Panter

d274b3f

2016-04-18 03:45:18 +0000

[diff] [blame]

777

if os.path.dirname(word) or word in (os.curdir, os.pardir):

778

# Ignore components that are not a simple file/directory name

779

continue

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

780

path = os.path.join(path, word)

Senthil Kumaran

72c238e

2013-09-13 00:21:18 -0700

[diff] [blame]

781

if trailing_slash:

782

path += '/'

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

783

return path

784

785

def copyfile(self, source, outputfile):

786

"""Copy all data between two file objects.

787

788

The SOURCE argument is a file object open for reading

789

(or anything with a read() method) and the DESTINATION

790

argument is a file object open for writing (or

791

anything with a write() method).

792

793

The only reason for overriding this would be to change

794

the block size or perhaps to replace newlines by CRLF

795

-- note however that this the default server uses this

796

to copy binary data as well.

797

798

"""

799

shutil.copyfileobj(source, outputfile)

800

801

def guess_type(self, path):

802

"""Guess the type of a file.

803

804

Argument is a PATH (a filename).

805

806

Return value is a string of the form type/subtype,

807

usable for a MIME Content-type header.

808

809

The default implementation looks the file's extension

810

up in the table self.extensions_map, using application/octet-stream

811

as a default; however it would be permissible (if

812

slow) to look inside the data to make a better guess.

"""

base, ext = posixpath.splitext(path)

817

if ext in self.extensions_map:

818

return self.extensions_map[ext]

819

ext = ext.lower()

820

if ext in self.extensions_map:

821

return self.extensions_map[ext]

822

else:

823

return self.extensions_map['']

824

825

if not mimetypes.inited:

826

mimetypes.init() # try to read system mime.types

827

extensions_map = mimetypes.types_map.copy()

828

extensions_map.update({

829

'': 'application/octet-stream', # Default

'.py': 'text/plain',

'.c': 'text/plain',

'.h': 'text/plain',

})

# Utilities for CGIHTTPRequestHandler

837

Senthil Kumaran

2012-04-12 02:34:32 +0800

[diff] [blame]

838

def _url_collapse_path(path):

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

839

"""

840

Given a URL path, remove extra '/'s and '.' path elements and collapse

Martin Panter

9955a37

2015-10-07 10:26:23 +0000

[diff] [blame]

841

any '..' references and returns a collapsed path.

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

842

843

Implements something akin to RFC-2396 5.2 step 6 to parse relative paths.

Senthil Kumaran

2012-04-12 02:34:32 +0800

[diff] [blame]

844

The utility of this function is limited to is_cgi method and helps

845

preventing some security attacks.

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

846

Martin Panter

2015-10-03 05:55:46 +0000

[diff] [blame]

847

Returns: The reconstituted URL, which will always start with a '/'.

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

848

849

Raises: IndexError if too many '..' occur within the path.

Senthil Kumaran

2012-04-12 02:34:32 +0800

[diff] [blame]

850

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

851

"""

Martin Panter

2015-10-03 05:55:46 +0000

[diff] [blame]

852

# Query component should not be involved.

853

path, _, query = path.partition('?')

854

path = urllib.parse.unquote(path)

855

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

856

# Similar to os.path.split(os.path.normpath(path)) but specific to URL

857

# path semantics rather than local operating system semantics.

Senthil Kumaran

2012-04-12 02:34:32 +0800

[diff] [blame]

858

path_parts = path.split('/')

859

head_parts = []

860

for part in path_parts[:-1]:

861

if part == '..':

862

head_parts.pop() # IndexError if more '..' than prior parts

863

elif part and part != '.':

864

head_parts.append( part )

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

865

if path_parts:

Senthil Kumaran

dbb369d

2012-04-11 03:15:28 +0800

[diff] [blame]

866

tail_part = path_parts.pop()

Senthil Kumaran

2012-04-12 02:34:32 +0800

[diff] [blame]

867

if tail_part:

868

if tail_part == '..':

869

head_parts.pop()

870

tail_part = ''

871

elif tail_part == '.':

872

tail_part = ''

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

873

else:

874

tail_part = ''

Senthil Kumaran

2012-04-12 02:34:32 +0800

[diff] [blame]

875

Martin Panter

2015-10-03 05:55:46 +0000

[diff] [blame]

876

if query:

877

tail_part = '?'.join((tail_part, query))

878

Senthil Kumaran

2012-04-12 02:34:32 +0800

[diff] [blame]

879

splitpath = ('/' + '/'.join(head_parts), tail_part)

880

collapsed_path = "/".join(splitpath)

881

882

return collapsed_path

883

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

884

885

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

nobody = None

def nobody_uid():

"""Internal routine to get nobody's uid"""

global nobody

if nobody:

return nobody

try:

import pwd

Brett Cannon

cd171c8

2013-07-04 17:43:24 -0400

[diff] [blame]

895

except ImportError:

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

896

return -1

897

try:

898

nobody = pwd.getpwnam('nobody')[2]

899

except KeyError:

Georg Brandl

cbd2ab1

2010-12-04 10:39:14 +0000

[diff] [blame]

900

nobody = 1 + max(x[2] for x in pwd.getpwall())

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

return nobody

def executable(path):

905

"""Test for executable file."""

Victor Stinner

fb25ba9

2011-06-20 17:45:54 +0200

[diff] [blame]

906

return os.access(path, os.X_OK)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

907

908

909

class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):

910

911

"""Complete HTTP server with GET, HEAD and POST commands.

912

913

GET and HEAD also support running CGI scripts.

914

915

The POST command is *only* implemented for CGI scripts.

"""

# Determine platform specifics

920

have_fork = hasattr(os, 'fork')

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

921

922

# Make rfile unbuffered -- we need to read one line and then pass

923

# the rest to a subprocess, so we can't use buffered input.

rbufsize = 0

def do_POST(self):

"""Serve a POST request.

928

929

This is only implemented for CGI scripts.

"""

if self.is_cgi():

self.run_cgi()

else:

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

936

self.send_error(

937

HTTPStatus.NOT_IMPLEMENTED,

938

"Can only POST to CGI scripts")

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

939

940

def send_head(self):

941

"""Version of send_head that support CGI scripts"""

942

if self.is_cgi():

943

return self.run_cgi()

944

else:

945

return SimpleHTTPRequestHandler.send_head(self)

946

947

def is_cgi(self):

948

"""Test whether self.path corresponds to a CGI script.

949

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

950

Returns True and updates the cgi_info attribute to the tuple

951

(dir, rest) if self.path requires running a CGI script.

952

Returns False otherwise.

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

953

Benjamin Peterson

a7deeee

2009-05-08 20:54:42 +0000

[diff] [blame]

954

If any exception is raised, the caller should assume that

955

self.path was rejected as invalid and act accordingly.

956

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

957

The default implementation tests whether the normalized url

958

path begins with one of the strings in self.cgi_directories

959

(and the next character is a '/' or the end of the string).

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

960

961

"""

Martin Panter

2015-10-03 05:55:46 +0000

[diff] [blame]

962

collapsed_path = _url_collapse_path(self.path)

Senthil Kumaran

2012-04-12 02:34:32 +0800

[diff] [blame]

963

dir_sep = collapsed_path.find('/', 1)

964

head, tail = collapsed_path[:dir_sep], collapsed_path[dir_sep+1:]

Senthil Kumaran

dbb369d

2012-04-11 03:15:28 +0800

[diff] [blame]

965

if head in self.cgi_directories:

966

self.cgi_info = head, tail

Benjamin Peterson

2009-04-11 20:12:10 +0000

[diff] [blame]

967

return True

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

968

return False

969

Senthil Kumaran

2012-04-12 02:34:32 +0800

[diff] [blame]

970

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

971

cgi_directories = ['/cgi-bin', '/htbin']

972

973

def is_executable(self, path):

974

"""Test whether argument path is an executable file."""

975

return executable(path)

976

977

def is_python(self, path):

978

"""Test whether argument path is a Python script."""

979

head, tail = os.path.splitext(path)

980

return tail.lower() in (".py", ".pyw")

981

982

def run_cgi(self):

983

"""Execute a CGI script."""

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

984

dir, rest = self.cgi_info

Ned Deily

915a30f

2014-07-12 22:06:26 -0700

[diff] [blame]

985

path = dir + '/' + rest

986

i = path.find('/', len(dir)+1)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

987

while i >= 0:

Ned Deily

915a30f

2014-07-12 22:06:26 -0700

[diff] [blame]

988

nextdir = path[:i]

989

nextrest = path[i+1:]

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

990

991

scriptdir = self.translate_path(nextdir)

992

if os.path.isdir(scriptdir):

993

dir, rest = nextdir, nextrest

Ned Deily

915a30f

2014-07-12 22:06:26 -0700

[diff] [blame]

994

i = path.find('/', len(dir)+1)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

else:

break

# find an explicit query string, if present.

Martin Panter

a02e18a

2015-10-03 05:38:07 +0000

[diff] [blame]

999

rest, _, query = rest.partition('?')

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1000

1001

# dissect the part after the directory name into a script name &

1002

# a possible additional path, to be stored in PATH_INFO.

1003

i = rest.find('/')

1004

if i >= 0:

1005

script, rest = rest[:i], rest[i:]

1006

else:

1007

script, rest = rest, ''

1008

1009

scriptname = dir + '/' + script

1010

scriptfile = self.translate_path(scriptname)

1011

if not os.path.exists(scriptfile):

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

1012

self.send_error(

1013

HTTPStatus.NOT_FOUND,

1014

"No such CGI script (%r)" % scriptname)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1015

return

1016

if not os.path.isfile(scriptfile):

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

1017

self.send_error(

1018

HTTPStatus.FORBIDDEN,

1019

"CGI script is not a plain file (%r)" % scriptname)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1020

return

1021

ispy = self.is_python(scriptname)

Victor Stinner

fb25ba9

2011-06-20 17:45:54 +0200

[diff] [blame]

1022

if self.have_fork or not ispy:

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1023

if not self.is_executable(scriptfile):

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

1024

self.send_error(

1025

HTTPStatus.FORBIDDEN,

1026

"CGI script is not executable (%r)" % scriptname)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1027

return

1028

1029

# Reference: http://hoohoo.ncsa.uiuc.edu/cgi/env.html

1030

# XXX Much of the following could be prepared ahead of time!

Senthil Kumaran

2010-10-03 17:55:45 +0000

[diff] [blame]

1031

env = copy.deepcopy(os.environ)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1032

env['SERVER_SOFTWARE'] = self.version_string()

1033

env['SERVER_NAME'] = self.server.server_name

1034

env['GATEWAY_INTERFACE'] = 'CGI/1.1'

1035

env['SERVER_PROTOCOL'] = self.protocol_version

1036

env['SERVER_PORT'] = str(self.server.server_port)

1037

env['REQUEST_METHOD'] = self.command

Jeremy Hylton

1afc169

2008-06-18 20:49:58 +0000

[diff] [blame]

1038

uqrest = urllib.parse.unquote(rest)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1039

env['PATH_INFO'] = uqrest

1040

env['PATH_TRANSLATED'] = self.translate_path(uqrest)

1041

env['SCRIPT_NAME'] = scriptname

1042

if query:

1043

env['QUERY_STRING'] = query

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1044

env['REMOTE_ADDR'] = self.client_address[0]

Barry Warsaw

2008-06-12 04:06:45 +0000

[diff] [blame]

1045

authorization = self.headers.get("authorization")

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1046

if authorization:

1047

authorization = authorization.split()

1048

if len(authorization) == 2:

1049

import base64, binascii

1050

env['AUTH_TYPE'] = authorization[0]

1051

if authorization[0].lower() == "basic":

1052

try:

1053

authorization = authorization[1].encode('ascii')

Georg Brandl

706824f

2009-06-04 09:42:55 +0000

[diff] [blame]

1054

authorization = base64.decodebytes(authorization).\

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1055

decode('ascii')

1056

except (binascii.Error, UnicodeError):

1057

pass

1058

else:

1059

authorization = authorization.split(':')

1060

if len(authorization) == 2:

1061

env['REMOTE_USER'] = authorization[0]

1062

# XXX REMOTE_IDENT

Barry Warsaw

2008-06-12 04:06:45 +0000

[diff] [blame]

1063

if self.headers.get('content-type') is None:

1064

env['CONTENT_TYPE'] = self.headers.get_content_type()

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1065

else:

Barry Warsaw

2008-06-12 04:06:45 +0000

[diff] [blame]

1066

env['CONTENT_TYPE'] = self.headers['content-type']

1067

length = self.headers.get('content-length')

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1068

if length:

1069

env['CONTENT_LENGTH'] = length

Barry Warsaw

2008-06-12 04:06:45 +0000

[diff] [blame]

1070

referer = self.headers.get('referer')

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1071

if referer:

1072

env['HTTP_REFERER'] = referer

1073

accept = []

1074

for line in self.headers.getallmatchingheaders('accept'):

1075

if line[:1] in "\t\n\r ":

1076

accept.append(line.strip())

1077

else:

1078

accept = accept + line[7:].split(',')

1079

env['HTTP_ACCEPT'] = ','.join(accept)

Barry Warsaw

2008-06-12 04:06:45 +0000

[diff] [blame]

1080

ua = self.headers.get('user-agent')

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1081

if ua:

1082

env['HTTP_USER_AGENT'] = ua

Barry Warsaw

2008-06-12 04:06:45 +0000

[diff] [blame]

1083

co = filter(None, self.headers.get_all('cookie', []))

Georg Brandl

62e2ca2

2010-07-31 21:54:24 +0000

[diff] [blame]

1084

cookie_str = ', '.join(co)

1085

if cookie_str:

1086

env['HTTP_COOKIE'] = cookie_str

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1087

# XXX Other HTTP_* headers

1088

# Since we're setting the env in the parent, provide empty

1089

# values to override previously set values

1090

for k in ('QUERY_STRING', 'REMOTE_HOST', 'CONTENT_LENGTH',

1091

'HTTP_USER_AGENT', 'HTTP_COOKIE', 'HTTP_REFERER'):

1092

env.setdefault(k, "")

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1093

Serhiy Storchaka

2014-12-23 16:28:28 +0200

[diff] [blame]

1094

self.send_response(HTTPStatus.OK, "Script output follows")

Senthil Kumaran

2011-05-09 23:25:02 +0800

[diff] [blame]

1095

self.flush_headers()

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1096

1097

decoded_query = query.replace('+', ' ')

1098

1099

if self.have_fork:

1100

# Unix -- fork as we should

1101

args = [script]

1102

if '=' not in decoded_query:

1103

args.append(decoded_query)

1104

nobody = nobody_uid()

1105

self.wfile.flush() # Always flush before forking

pid = os.fork()

if pid != 0:

# Parent

pid, sts = os.waitpid(pid, 0)

1110

# throw away additional data [see bug #427345]

1111

while select.select([self.rfile], [], [], 0)[0]:

1112

if not self.rfile.read(1):

1113

break

1114

if sts:

1115

self.log_error("CGI script exit status %#x", sts)

return

# Child

try:

try:

os.setuid(nobody)

Andrew Svetlov

ad28c7f

2012-12-18 22:02:39 +0200

[diff] [blame]

1121

except OSError:

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1122

pass

1123

os.dup2(self.rfile.fileno(), 0)

1124

os.dup2(self.wfile.fileno(), 1)

Senthil Kumaran

2010-10-03 17:55:45 +0000

[diff] [blame]

1125

os.execve(scriptfile, args, env)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1126

except:

1127

self.server.handle_error(self.request, self.client_address)

1128

os._exit(127)

1129

Amaury Forgeot d'Arc

2008-06-18 22:19:22 +0000

[diff] [blame]

1130

else:

1131

# Non-Unix -- use subprocess

1132

import subprocess

Senthil Kumaran

e29cd16

2009-11-11 04:17:53 +0000

[diff] [blame]

1133

cmdline = [scriptfile]

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1134

if self.is_python(scriptfile):

1135

interp = sys.executable

1136

if interp.lower().endswith("w.exe"):

1137

# On Windows, use python.exe, not pythonw.exe

1138

interp = interp[:-5] + interp[-4:]

Senthil Kumaran

e29cd16

2009-11-11 04:17:53 +0000

[diff] [blame]

1139

cmdline = [interp, '-u'] + cmdline

1140

if '=' not in query:

1141

cmdline.append(query)

1142

self.log_message("command: %s", subprocess.list2cmdline(cmdline))

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1143

try:

1144

nbytes = int(length)

1145

except (TypeError, ValueError):

1146

nbytes = 0

Amaury Forgeot d'Arc

2008-06-18 22:19:22 +0000

[diff] [blame]

1147

p = subprocess.Popen(cmdline,

1148

stdin=subprocess.PIPE,

1149

stdout=subprocess.PIPE,

Senthil Kumaran

2010-10-03 17:55:45 +0000

[diff] [blame]

1150

stderr=subprocess.PIPE,

1151

env = env

Amaury Forgeot d'Arc

2008-06-18 22:19:22 +0000

[diff] [blame]

1152

)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1153

if self.command.lower() == "post" and nbytes > 0:

1154

data = self.rfile.read(nbytes)

Amaury Forgeot d'Arc

2008-06-18 22:19:22 +0000

[diff] [blame]

1155

else:

1156

data = None

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1157

# throw away additional data [see bug #427345]

1158

while select.select([self.rfile._sock], [], [], 0)[0]:

1159

if not self.rfile._sock.recv(1):

1160

break

Amaury Forgeot d'Arc

2008-06-18 22:19:22 +0000

[diff] [blame]

1161

stdout, stderr = p.communicate(data)

1162

self.wfile.write(stdout)

1163

if stderr:

1164

self.log_error('%s', stderr)

Brian Curtin

cbad4df

2010-11-05 15:04:48 +0000

[diff] [blame]

1165

p.stderr.close()

1166

p.stdout.close()

Amaury Forgeot d'Arc

2008-06-18 22:19:22 +0000

[diff] [blame]

1167

status = p.returncode

1168

if status:

1169

self.log_error("CGI script exit status %#x", status)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1170

else:

1171

self.log_message("CGI script exited OK")

1172

1173

Senthil Kumaran

2013-09-15 09:37:27 -0700

[diff] [blame]

1174

def test(HandlerClass=BaseHTTPRequestHandler,

1175

ServerClass=HTTPServer, protocol="HTTP/1.0", port=8000, bind=""):

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1176

"""Test the HTTP request handler class.

1177

Robert Collins

9644f24

2015-08-17 12:18:35 +1200

[diff] [blame]

1178

This runs an HTTP server on port 8000 (or the port argument).

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1179

1180

"""

Senthil Kumaran

2013-09-15 09:37:27 -0700

[diff] [blame]

1181

server_address = (bind, port)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1182

1183

HandlerClass.protocol_version = protocol

1184

httpd = ServerClass(server_address, HandlerClass)

1185

1186

sa = httpd.socket.getsockname()

1187

print("Serving HTTP on", sa[0], "port", sa[1], "...")

Alexandre Vassalotti

b5292a2

2009-04-03 07:16:55 +0000

[diff] [blame]

1188

try:

1189

httpd.serve_forever()

1190

except KeyboardInterrupt:

1191

print("\nKeyboard interrupt received, exiting.")

1192

httpd.server_close()

1193

sys.exit(0)

Georg Brandl

2008-05-26 16:32:26 +0000

[diff] [blame]

1194

1195

if __name__ == '__main__':

Senthil Kumaran

2012-06-03 16:15:54 +0800

[diff] [blame]

1196

parser = argparse.ArgumentParser()

1197

parser.add_argument('--cgi', action='store_true',

1198

help='Run as CGI Server')

Senthil Kumaran

2013-09-15 09:37:27 -0700

[diff] [blame]

1199

parser.add_argument('--bind', '-b', default='', metavar='ADDRESS',

1200

help='Specify alternate bind address '

1201

'[default: all interfaces]')

Senthil Kumaran

2012-06-03 16:15:54 +0800

[diff] [blame]

1202

parser.add_argument('port', action='store',

1203

default=8000, type=int,

1204

nargs='?',

1205

help='Specify alternate port [default: 8000]')

1206

args = parser.parse_args()

1207

if args.cgi:

Senthil Kumaran

2013-09-15 09:37:27 -0700

[diff] [blame]

1208

handler_class = CGIHTTPRequestHandler

Senthil Kumaran

2012-06-03 16:15:54 +0800

[diff] [blame]

1209

else:

Senthil Kumaran