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__ = [

Géry Ogam

1cee216

2018-05-29 22:10:30 +0200

[diff] [blame]

86

"HTTPServer", "ThreadingHTTPServer", "BaseHTTPRequestHandler",

Berker Peksag

366c570

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

[diff] [blame]

87

"SimpleHTTPRequestHandler", "CGIHTTPRequestHandler",

88

]

Georg Brandl

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

[diff] [blame]

89

Pierre Quentel

2017-04-02 12:26:12 +0200

[diff] [blame]

90

import copy

91

import datetime

Berker Peksag

04bc5b9

2016-03-14 06:06:03 +0200

[diff] [blame]

92

import email.utils

Georg Brandl

1f7fffb

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

[diff] [blame]

93

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

Jason R. Coombs

7cdc31a

2020-01-06 07:59:36 -0500

[diff] [blame]

106

import contextlib

Stéphane Wirtel

2017-05-24 09:29:06 +0200

[diff] [blame]

107

from functools import partial

Senthil Kumaran

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

[diff] [blame]

108

Serhiy Storchaka

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

[diff] [blame]

109

from http import HTTPStatus

110

Georg Brandl

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

[diff] [blame]

111

112

# Default error message template

113

DEFAULT_ERROR_MESSAGE = """\

Senthil Kumaran

1b407fe

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

[diff] [blame]

114

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

115

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

Ezio Melotti

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

[diff] [blame]

116

<html>

Senthil Kumaran

b253c9f

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

[diff] [blame]

117

<head>

Senthil Kumaran

1b407fe

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

[diff] [blame]

118

Senthil Kumaran

b253c9f

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

[diff] [blame]

119

<title>Error response</title>

120

</head>

121

<body>

122

<h1>Error response</h1>

123

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

124

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

125

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

126

</body>

127

</html>

Georg Brandl

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

[diff] [blame]

128

"""

129

130

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

131

Georg Brandl

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

[diff] [blame]

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)

Martin Panter

2016-04-03 01:28:53 +0000

[diff] [blame]

139

host, port = self.server_address[:2]

Georg Brandl

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

[diff] [blame]

140

self.server_name = socket.getfqdn(host)

141

self.server_port = port

142

143

Géry Ogam

1cee216

2018-05-29 22:10:30 +0200

[diff] [blame]

144

class ThreadingHTTPServer(socketserver.ThreadingMixIn, HTTPServer):

Julien Palard

8bcfa02

2018-03-23 17:40:33 +0100

[diff] [blame]

145

daemon_threads = True

146

147

Georg Brandl

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

[diff] [blame]

148

class BaseHTTPRequestHandler(socketserver.StreamRequestHandler):

149

150

"""HTTP request handler base class.

151

152

The following explanation of HTTP serves to guide you through the

153

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

154

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

155

:-).

156

157

HTTP (HyperText Transfer Protocol) is an extensible protocol on

158

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

159

recognizes three parts to a request:

160

161

1. One line identifying the request type and path

162

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

163

3. An optional data part

164

165

The headers and data are separated by a blank line.

166

167

The first line of the request has the form

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

172

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

173

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

174

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

175

the ASCII character with hex code xx).

176

177

The specification specifies that lines are separated by CRLF but

178

for compatibility with the widest range of clients recommends

179

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

180

is treated sensibly (allowing multiple spaces between components

181

and allowing trailing whitespace).

182

183

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

184

but most clients grok LF characters just fine.

185

186

If the first line of the request has the form

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

191

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

192

the reply consists of just the data.

193

194

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

195

196

1. One line giving the response code

197

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

198

3. The data

199

200

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

201

202

The response code line has the form

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

207

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

208

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

209

human-readable string explaining what the response code means.

210

211

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

212

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

213

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

214

such method exists the server sends an error response to the

215

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

220

are different requests).

221

222

The various request details are stored in instance variables:

223

224

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

225

port);

226

227

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

228

Barry Warsaw

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

[diff] [blame]

229

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

Georg Brandl

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

[diff] [blame]

230

class) containing the header information;

231

232

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

233

start of the optional input data part;

234

235

- wfile is a file object open for writing.

236

237

IT IS IMPORTANT TO ADHERE TO THE PROTOCOL FOR WRITING!

238

239

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

240

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

241

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

242

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

243

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

244

245

Content-type: <type>/<subtype>

246

247

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

248

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

"""

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

253

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

254

255

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

256

# The format is multiple whitespace-separated strings,

257

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

258

server_version = "BaseHTTP/" + __version__

259

260

error_message_format = DEFAULT_ERROR_MESSAGE

261

error_content_type = DEFAULT_ERROR_CONTENT_TYPE

262

263

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

264

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

265

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

266

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

267

default_request_version = "HTTP/0.9"

268

269

def parse_request(self):

270

"""Parse a request (internal).

271

272

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

273

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

274

self.headers.

275

Martin Panter

2016-11-19 01:06:37 +0000

[diff] [blame]

276

Return True for success, False for failure; on failure, any relevant

277

error response has already been sent back.

Georg Brandl

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

[diff] [blame]

278

279

"""

280

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

281

self.request_version = version = self.default_request_version

Benjamin Peterson

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

[diff] [blame]

282

self.close_connection = True

Georg Brandl

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

[diff] [blame]

283

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

Senthil Kumaran

3075549

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

[diff] [blame]

284

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

Georg Brandl

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

[diff] [blame]

285

self.requestline = requestline

286

words = requestline.split()

Martin Panter

2016-11-19 01:06:37 +0000

[diff] [blame]

if len(words) == 0:

return False

if len(words) >= 3: # Enough to determine protocol version

291

version = words[-1]

Georg Brandl

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

[diff] [blame]

292

try:

Martin Panter

2016-11-19 01:06:37 +0000

[diff] [blame]

293

if not version.startswith('HTTP/'):

Martin Panter

2016-04-03 01:28:53 +0000

[diff] [blame]

294

raise ValueError

Georg Brandl

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

[diff] [blame]

295

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

296

version_number = base_version_number.split(".")

297

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

298

# - major and minor numbers MUST be treated as

299

# separate integers;

300

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

301

# turn is lower than HTTP/12.3;

302

# - Leading zeros MUST be ignored by recipients.

303

if len(version_number) != 2:

304

raise ValueError

305

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

306

except (ValueError, IndexError):

Serhiy Storchaka

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

[diff] [blame]

307

self.send_error(

308

HTTPStatus.BAD_REQUEST,

309

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

Georg Brandl

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

[diff] [blame]

310

return False

311

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

Benjamin Peterson

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

[diff] [blame]

312

self.close_connection = False

Georg Brandl

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

[diff] [blame]

313

if version_number >= (2, 0):

Serhiy Storchaka

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

[diff] [blame]

314

self.send_error(

315

HTTPStatus.HTTP_VERSION_NOT_SUPPORTED,

Martin Panter

2016-04-03 01:28:53 +0000

[diff] [blame]

316

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

Georg Brandl

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

[diff] [blame]

317

return False

Martin Panter

2016-11-19 01:06:37 +0000

[diff] [blame]

318

self.request_version = version

319

320

if not 2 <= len(words) <= 3:

321

self.send_error(

322

HTTPStatus.BAD_REQUEST,

323

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

324

return False

325

command, path = words[:2]

326

if len(words) == 2:

Benjamin Peterson

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

[diff] [blame]

327

self.close_connection = True

Georg Brandl

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

[diff] [blame]

328

if command != 'GET':

Serhiy Storchaka

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

[diff] [blame]

329

self.send_error(

330

HTTPStatus.BAD_REQUEST,

331

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

Georg Brandl

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

[diff] [blame]

332

return False

Martin Panter

2016-11-19 01:06:37 +0000

[diff] [blame]

333

self.command, self.path = command, path

Georg Brandl

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

[diff] [blame]

334

335

# Examine the headers and look for a Connection directive.

Senthil Kumaran

5466bf1

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

[diff] [blame]

336

try:

337

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

338

_class=self.MessageClass)

Martin Panter

2016-04-03 01:28:53 +0000

[diff] [blame]

339

except http.client.LineTooLong as err:

Serhiy Storchaka

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

[diff] [blame]

340

self.send_error(

Martin Panter

2016-04-03 01:28:53 +0000

[diff] [blame]

341

HTTPStatus.REQUEST_HEADER_FIELDS_TOO_LARGE,

342

"Line too long",

343

str(err))

Senthil Kumaran

5466bf1

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

[diff] [blame]

344

return False

Martin Panter

acc0319

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

[diff] [blame]

345

except http.client.HTTPException as err:

346

self.send_error(

347

HTTPStatus.REQUEST_HEADER_FIELDS_TOO_LARGE,

"Too many headers",

str(err)

)

return False

Georg Brandl

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

[diff] [blame]

352

353

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

354

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

Benjamin Peterson

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

[diff] [blame]

355

self.close_connection = True

Georg Brandl

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

[diff] [blame]

356

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

357

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

Benjamin Peterson

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

[diff] [blame]

358

self.close_connection = False

Senthil Kumaran

0f476d4

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

[diff] [blame]

359

# Examine the headers and look for an Expect directive

360

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

361

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

362

self.protocol_version >= "HTTP/1.1" and

363

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

364

if not self.handle_expect_100():

return False

return True

def handle_expect_100(self):

369

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

370

371

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

372

respond with either a 100 Continue or a final response before

373

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

374

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

375

reject unauthorized requests) by overriding this method.

376

377

This method should either return True (possibly after sending

378

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

379

False.

380

381

"""

Serhiy Storchaka

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

[diff] [blame]

382

self.send_response_only(HTTPStatus.CONTINUE)

Benjamin Peterson

0442423

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

[diff] [blame]

383

self.end_headers()

Georg Brandl

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

[diff] [blame]

384

return True

385

386

def handle_one_request(self):

387

"""Handle a single HTTP request.

388

389

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

390

__doc__ string for information on how to handle specific HTTP

391

commands such as GET and POST.

392

393

"""

Kristján Valur Jónsson

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

[diff] [blame]

394

try:

Antoine Pitrou

c492437

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

[diff] [blame]

395

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

396

if len(self.raw_requestline) > 65536:

397

self.requestline = ''

398

self.request_version = ''

399

self.command = ''

Serhiy Storchaka

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

[diff] [blame]

400

self.send_error(HTTPStatus.REQUEST_URI_TOO_LONG)

Antoine Pitrou

c492437

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

[diff] [blame]

401

return

Kristján Valur Jónsson

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

[diff] [blame]

402

if not self.raw_requestline:

Benjamin Peterson

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

[diff] [blame]

403

self.close_connection = True

Kristján Valur Jónsson

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

[diff] [blame]

404

return

405

if not self.parse_request():

406

# An error code has been sent, just exit

407

return

408

mname = 'do_' + self.command

409

if not hasattr(self, mname):

Serhiy Storchaka

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

[diff] [blame]

410

self.send_error(

411

HTTPStatus.NOT_IMPLEMENTED,

412

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

Kristján Valur Jónsson

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

[diff] [blame]

413

return

414

method = getattr(self, mname)

415

method()

416

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

Christian Heimes

03c8ddd

2020-11-20 09:26:07 +0100

[diff] [blame]

417

except TimeoutError as e:

Kristján Valur Jónsson

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

[diff] [blame]

418

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

419

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

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

return

Georg Brandl

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

[diff] [blame]

422

423

def handle(self):

424

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

Benjamin Peterson

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

[diff] [blame]

425

self.close_connection = True

Georg Brandl

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

[diff] [blame]

426

427

self.handle_one_request()

428

while not self.close_connection:

429

self.handle_one_request()

430

Senthil Kumaran

2688644

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

[diff] [blame]

431

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

Georg Brandl

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

[diff] [blame]

432

"""Send and log an error reply.

433

Senthil Kumaran

2688644

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

[diff] [blame]

434

Arguments are

435

* code: an HTTP error code

436

3 digits

437

* message: a simple optional 1 line reason phrase.

438

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

439

defaults to short entry matching the response code

440

* explain: a detailed message defaults to the long entry

441

matching the response code.

Georg Brandl

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

[diff] [blame]

442

443

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

444

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

445

a piece of HTML explaining the error to the user.

"""

try:

shortmsg, longmsg = self.responses[code]

451

except KeyError:

452

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

453

if message is None:

454

message = shortmsg

Senthil Kumaran

2688644

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

[diff] [blame]

455

if explain is None:

456

explain = longmsg

Georg Brandl

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

[diff] [blame]

457

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

Senthil Kumaran

1e7551d

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

[diff] [blame]

458

self.send_response(code, message)

Georg Brandl

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

[diff] [blame]

459

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

Martin Panter

e42e129

2016-06-08 08:29:13 +0000

[diff] [blame]

460

461

# Message body is omitted for cases described in:

462

# - RFC7230: 3.3. 1xx, 204(No Content), 304(Not Modified)

463

# - RFC7231: 6.3.6. 205(Reset Content)

464

body = None

465

if (code >= 200 and

466

code not in (HTTPStatus.NO_CONTENT,

467

HTTPStatus.RESET_CONTENT,

468

HTTPStatus.NOT_MODIFIED)):

469

# HTML encode to prevent Cross Site Scripting attacks

470

# (see bug #1100201)

471

content = (self.error_message_format % {

472

'code': code,

Martin Panter

40de69a

2016-06-08 09:45:58 +0000

[diff] [blame]

473

'message': html.escape(message, quote=False),

474

'explain': html.escape(explain, quote=False)

Martin Panter

e42e129

2016-06-08 08:29:13 +0000

[diff] [blame]

475

})

476

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

477

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

ValeriyaSinevich

b36b0a3

2018-06-18 14:17:53 -0700

[diff] [blame]

478

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

Georg Brandl

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

[diff] [blame]

479

self.end_headers()

Serhiy Storchaka

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

[diff] [blame]

480

Martin Panter

e42e129

2016-06-08 08:29:13 +0000

[diff] [blame]

481

if self.command != 'HEAD' and body:

Senthil Kumaran

52d2720

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

[diff] [blame]

482

self.wfile.write(body)

Georg Brandl

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

[diff] [blame]

483

484

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

Senthil Kumaran

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

[diff] [blame]

485

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

486

response code.

Georg Brandl

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

[diff] [blame]

487

488

Also send two standard headers with the server software

489

version and the current date.

490

491

"""

492

self.log_request(code)

Senthil Kumaran

0f476d4

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

[diff] [blame]

493

self.send_response_only(code, message)

494

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

495

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

496

497

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

498

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

Georg Brandl

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

[diff] [blame]

499

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

Martin Panter

2016-04-03 01:28:53 +0000

[diff] [blame]

500

if message is None:

501

if code in self.responses:

502

message = self.responses[code][0]

503

else:

504

message = ''

Senthil Kumaran

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

[diff] [blame]

505

if not hasattr(self, '_headers_buffer'):

506

self._headers_buffer = []

507

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

508

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

509

'latin-1', 'strict'))

Georg Brandl

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

[diff] [blame]

510

511

def send_header(self, keyword, value):

Senthil Kumaran

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

[diff] [blame]

512

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

Georg Brandl

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

[diff] [blame]

513

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

Senthil Kumaran

e4dad4f

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

[diff] [blame]

514

if not hasattr(self, '_headers_buffer'):

515

self._headers_buffer = []

516

self._headers_buffer.append(

Marc-André Lemburg

8f36af7

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

[diff] [blame]

517

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

Georg Brandl

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

[diff] [blame]

518

519

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

520

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

Benjamin Peterson

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

[diff] [blame]

521

self.close_connection = True

Georg Brandl

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

[diff] [blame]

522

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

Benjamin Peterson

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

[diff] [blame]

523

self.close_connection = False

Georg Brandl

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

[diff] [blame]

524

525

def end_headers(self):

526

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

527

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

Senthil Kumaran

e4dad4f

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

[diff] [blame]

528

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

Senthil Kumaran

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

[diff] [blame]

529

self.flush_headers()

530

531

def flush_headers(self):

532

if hasattr(self, '_headers_buffer'):

Senthil Kumaran

e4dad4f

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

[diff] [blame]

533

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

534

self._headers_buffer = []

Georg Brandl

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

[diff] [blame]

535

536

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

537

"""Log an accepted request.

538

539

This is called by send_response().

540

541

"""

Serhiy Storchaka

c0a23e6

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

[diff] [blame]

542

if isinstance(code, HTTPStatus):

543

code = code.value

Georg Brandl

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

[diff] [blame]

544

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

545

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

546

547

def log_error(self, format, *args):

548

"""Log an error.

549

550

This is called when a request cannot be fulfilled. By

551

default it passes the message on to log_message().

552

553

Arguments are the same as for log_message().

554

555

XXX This should go to the separate error log.

"""

self.log_message(format, *args)

560

561

def log_message(self, format, *args):

562

"""Log an arbitrary message.

563

564

This is used by all other logging functions. Override

565

it if you have specific logging wishes.

566

567

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

568

message to be logged. If the format string contains

569

any % escapes requiring parameters, they should be

570

specified as subsequent arguments (it's just like

571

printf!).

572

Senthil Kumaran

db727b4

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

[diff] [blame]

573

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" %

579

(self.address_string(),

580

self.log_date_time_string(),

581

format%args))

582

583

def version_string(self):

584

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

585

return self.server_version + ' ' + self.sys_version

586

587

def date_time_string(self, timestamp=None):

588

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

589

if timestamp is None:

590

timestamp = time.time()

Berker Peksag

04bc5b9

2016-03-14 06:06:03 +0200

[diff] [blame]

591

return email.utils.formatdate(timestamp, usegmt=True)

Georg Brandl

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

[diff] [blame]

592

593

def log_date_time_string(self):

594

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

595

now = time.time()

596

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

597

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

598

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

599

return s

600

601

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

602

603

monthname = [None,

604

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

605

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

606

607

def address_string(self):

Senthil Kumaran

1aacba4

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

[diff] [blame]

608

"""Return the client address."""

Georg Brandl

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

[diff] [blame]

609

Senthil Kumaran

1aacba4

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

[diff] [blame]

610

return self.client_address[0]

Georg Brandl

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

[diff] [blame]

611

612

# Essentially static class variables

613

614

# The version of the HTTP protocol we support.

615

# Set this to HTTP/1.1 to enable automatic keepalive

616

protocol_version = "HTTP/1.0"

617

Barry Warsaw

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

[diff] [blame]

618

# MessageClass used to parse headers

Barry Warsaw

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

[diff] [blame]

619

MessageClass = http.client.HTTPMessage

Georg Brandl

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

[diff] [blame]

620

Serhiy Storchaka

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

[diff] [blame]

621

# hack to maintain backwards compatibility

Georg Brandl

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

[diff] [blame]

622

responses = {

Serhiy Storchaka

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

[diff] [blame]

623

v: (v.phrase, v.description)

624

for v in HTTPStatus.__members__.values()

625

}

Georg Brandl

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

[diff] [blame]

626

627

628

class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):

629

630

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

631

632

This serves files from the current directory and any of its

633

subdirectories. The MIME type for files is determined by

634

calling the .guess_type() method.

635

636

The GET and HEAD requests are identical except that the HEAD

637

request omits the actual contents of the file.

"""

server_version = "SimpleHTTP/" + __version__

An Long

5907e61

2020-01-09 02:28:14 +0800

[diff] [blame]

642

extensions_map = _encodings_map_default = {

643

'.gz': 'application/gzip',

644

'.Z': 'application/octet-stream',

645

'.bz2': 'application/x-bzip2',

646

'.xz': 'application/x-xz',

647

}

Georg Brandl

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

[diff] [blame]

648

Stéphane Wirtel

2017-05-24 09:29:06 +0200

[diff] [blame]

649

def __init__(self, *args, directory=None, **kwargs):

650

if directory is None:

651

directory = os.getcwd()

Géry Ogam

781266e

2019-09-11 15:03:46 +0200

[diff] [blame]

652

self.directory = os.fspath(directory)

Stéphane Wirtel

2017-05-24 09:29:06 +0200

[diff] [blame]

653

super().__init__(*args, **kwargs)

654

Georg Brandl

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

[diff] [blame]

655

def do_GET(self):

656

"""Serve a GET request."""

657

f = self.send_head()

658

if f:

Serhiy Storchaka

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

[diff] [blame]

659

try:

660

self.copyfile(f, self.wfile)

661

finally:

662

f.close()

Georg Brandl

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

[diff] [blame]

663

664

def do_HEAD(self):

665

"""Serve a HEAD request."""

f = self.send_head()

if f:

f.close()

def send_head(self):

"""Common code for GET and HEAD commands.

672

673

This sends the response code and MIME headers.

674

675

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

676

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

677

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

678

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

679

680

"""

681

path = self.translate_path(self.path)

682

f = None

683

if os.path.isdir(path):

Benjamin Peterson

94cb7a2

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

[diff] [blame]

684

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

685

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

Georg Brandl

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

[diff] [blame]

686

# redirect browser - doing basically what apache does

Serhiy Storchaka

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

[diff] [blame]

687

self.send_response(HTTPStatus.MOVED_PERMANENTLY)

Benjamin Peterson

94cb7a2

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

[diff] [blame]

688

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

689

parts[3], parts[4])

690

new_url = urllib.parse.urlunsplit(new_parts)

691

self.send_header("Location", new_url)

Georg Brandl

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

[diff] [blame]

692

self.end_headers()

693

return None

694

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

695

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

696

if os.path.exists(index):

path = index

break

else:

return self.list_directory(path)

701

ctype = self.guess_type(path)

Michael Felt

2062a20

2018-12-26 06:43:42 +0100

[diff] [blame]

702

# check for trailing "/" which should return 404. See Issue17324

703

# The test for this was added in test_httpserver.py

704

# However, some OS platforms accept a trailingSlash as a filename

705

# See discussion on python-dev and Issue34711 regarding

706

# parseing and rejection of filenames with a trailing slash

707

if path.endswith("/"):

708

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

709

return None

Georg Brandl

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

[diff] [blame]

710

try:

711

f = open(path, 'rb')

Andrew Svetlov

f7a17b4

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

[diff] [blame]

712

except OSError:

Serhiy Storchaka

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

[diff] [blame]

713

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

Georg Brandl

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

[diff] [blame]

714

return None

Pierre Quentel

2017-04-02 12:26:12 +0200

[diff] [blame]

715

Serhiy Storchaka

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

[diff] [blame]

716

try:

Pierre Quentel

2017-04-02 12:26:12 +0200

[diff] [blame]

717

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

718

# Use browser cache if possible

719

if ("If-Modified-Since" in self.headers

720

and "If-None-Match" not in self.headers):

721

# compare If-Modified-Since and time of last file modification

722

try:

723

ims = email.utils.parsedate_to_datetime(

724

self.headers["If-Modified-Since"])

725

except (TypeError, IndexError, OverflowError, ValueError):

726

# ignore ill-formed values

727

pass

728

else:

729

if ims.tzinfo is None:

730

# obsolete format with no timezone, cf.

731

# https://tools.ietf.org/html/rfc7231#section-7.1.1.1

732

ims = ims.replace(tzinfo=datetime.timezone.utc)

733

if ims.tzinfo is datetime.timezone.utc:

734

# compare to UTC datetime of last modification

735

last_modif = datetime.datetime.fromtimestamp(

736

fs.st_mtime, datetime.timezone.utc)

737

# remove microseconds, like in If-Modified-Since

738

last_modif = last_modif.replace(microsecond=0)

Serhiy Storchaka

13ad3b7

2017-09-14 09:38:36 +0300

[diff] [blame]

739

Pierre Quentel

2017-04-02 12:26:12 +0200

[diff] [blame]

740

if last_modif <= ims:

741

self.send_response(HTTPStatus.NOT_MODIFIED)

self.end_headers()

f.close()

return None

Serhiy Storchaka

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

[diff] [blame]

746

self.send_response(HTTPStatus.OK)

Serhiy Storchaka

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

[diff] [blame]

747

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

Serhiy Storchaka

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

[diff] [blame]

748

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

Serhiy Storchaka

13ad3b7

2017-09-14 09:38:36 +0300

[diff] [blame]

749

self.send_header("Last-Modified",

Pierre Quentel

2017-04-02 12:26:12 +0200

[diff] [blame]

750

self.date_time_string(fs.st_mtime))

Serhiy Storchaka

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

[diff] [blame]

self.end_headers()

return f

except:

f.close()

raise

Georg Brandl

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

[diff] [blame]

756

757

def list_directory(self, path):

758

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

759

760

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

761

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

762

interface the same as for send_head().

"""

try:

list = os.listdir(path)

Andrew Svetlov

ad28c7f

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

[diff] [blame]

767

except OSError:

Serhiy Storchaka

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

[diff] [blame]

768

self.send_error(

769

HTTPStatus.NOT_FOUND,

770

"No permission to list directory")

Georg Brandl

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

[diff] [blame]

771

return None

772

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

773

r = []

Serhiy Storchaka

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

[diff] [blame]

774

try:

775

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

776

errors='surrogatepass')

777

except UnicodeDecodeError:

778

displaypath = urllib.parse.unquote(path)

Martin Panter

da3bb38

2016-04-11 00:40:08 +0000

[diff] [blame]

779

displaypath = html.escape(displaypath, quote=False)

Ezio Melotti

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

[diff] [blame]

780

enc = sys.getfilesystemencoding()

781

title = 'Directory listing for %s' % displaypath

782

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

783

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

784

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

785

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

786

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

787

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

788

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

789

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

Georg Brandl

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

[diff] [blame]

790

for name in list:

791

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

792

displayname = linkname = name

793

# Append / for directories or @ for symbolic links

794

if os.path.isdir(fullname):

795

displayname = name + "/"

796

linkname = name + "/"

797

if os.path.islink(fullname):

798

displayname = name + "@"

799

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

Ezio Melotti

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

[diff] [blame]

800

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

Serhiy Storchaka

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

[diff] [blame]

801

% (urllib.parse.quote(linkname,

802

errors='surrogatepass'),

Martin Panter

da3bb38

2016-04-11 00:40:08 +0000

[diff] [blame]

803

html.escape(displayname, quote=False)))

Ezio Melotti

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

[diff] [blame]

804

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

Serhiy Storchaka

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

[diff] [blame]

805

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

Georg Brandl

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

[diff] [blame]

806

f = io.BytesIO()

807

f.write(encoded)

808

f.seek(0)

Serhiy Storchaka

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

[diff] [blame]

809

self.send_response(HTTPStatus.OK)

Georg Brandl

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

[diff] [blame]

810

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

811

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

self.end_headers()

return f

def translate_path(self, path):

816

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

817

818

Components that mean special things to the local file system

819

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

820

probably be diagnosed.)

821

822

"""

823

# abandon query parameters

824

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

825

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

Senthil Kumaran

72c238e

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

[diff] [blame]

826

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

Senthil Kumaran

600b735

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

[diff] [blame]

827

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

Serhiy Storchaka

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

[diff] [blame]

828

try:

829

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

830

except UnicodeDecodeError:

831

path = urllib.parse.unquote(path)

832

path = posixpath.normpath(path)

Georg Brandl

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

[diff] [blame]

833

words = path.split('/')

834

words = filter(None, words)

Stéphane Wirtel

2017-05-24 09:29:06 +0200

[diff] [blame]

835

path = self.directory

Georg Brandl

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

[diff] [blame]

836

for word in words:

Martin Panter

d274b3f

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

[diff] [blame]

837

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

838

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

839

continue

Georg Brandl

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

[diff] [blame]

840

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

Senthil Kumaran

72c238e

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

[diff] [blame]

841

if trailing_slash:

842

path += '/'

Georg Brandl

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

[diff] [blame]

843

return path

844

845

def copyfile(self, source, outputfile):

846

"""Copy all data between two file objects.

847

848

The SOURCE argument is a file object open for reading

849

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

850

argument is a file object open for writing (or

851

anything with a write() method).

852

853

The only reason for overriding this would be to change

854

the block size or perhaps to replace newlines by CRLF

855

-- note however that this the default server uses this

856

to copy binary data as well.

857

858

"""

859

shutil.copyfileobj(source, outputfile)

860

861

def guess_type(self, path):

862

"""Guess the type of a file.

863

864

Argument is a PATH (a filename).

865

866

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

867

usable for a MIME Content-type header.

868

869

The default implementation looks the file's extension

870

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

871

as a default; however it would be permissible (if

872

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

873

874

"""

Georg Brandl

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

[diff] [blame]

875

base, ext = posixpath.splitext(path)

876

if ext in self.extensions_map:

877

return self.extensions_map[ext]

878

ext = ext.lower()

879

if ext in self.extensions_map:

880

return self.extensions_map[ext]

An Long

5907e61

2020-01-09 02:28:14 +0800

[diff] [blame]

881

guess, _ = mimetypes.guess_type(path)

882

if guess:

883

return guess

884

return 'application/octet-stream'

Georg Brandl

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

[diff] [blame]

885

886

887

# Utilities for CGIHTTPRequestHandler

888

Senthil Kumaran

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

[diff] [blame]

889

def _url_collapse_path(path):

Benjamin Peterson

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

[diff] [blame]

890

"""

891

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

Martin Panter

9955a37

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

[diff] [blame]

892

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

Benjamin Peterson

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

[diff] [blame]

893

894

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]

895

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

896

preventing some security attacks.

Benjamin Peterson

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

[diff] [blame]

897

Martin Panter

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

[diff] [blame]

898

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

Benjamin Peterson

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

[diff] [blame]

899

900

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

Senthil Kumaran

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

[diff] [blame]

901

Benjamin Peterson

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

[diff] [blame]

902

"""

Martin Panter

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

[diff] [blame]

903

# Query component should not be involved.

904

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

905

path = urllib.parse.unquote(path)

906

Benjamin Peterson

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

[diff] [blame]

907

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

908

# path semantics rather than local operating system semantics.

Senthil Kumaran

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

[diff] [blame]

909

path_parts = path.split('/')

910

head_parts = []

911

for part in path_parts[:-1]:

912

if part == '..':

913

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

914

elif part and part != '.':

915

head_parts.append( part )

Benjamin Peterson

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

[diff] [blame]

916

if path_parts:

Senthil Kumaran

dbb369d

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

[diff] [blame]

917

tail_part = path_parts.pop()

Senthil Kumaran

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

[diff] [blame]

918

if tail_part:

919

if tail_part == '..':

920

head_parts.pop()

921

tail_part = ''

922

elif tail_part == '.':

923

tail_part = ''

Benjamin Peterson

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

[diff] [blame]

924

else:

925

tail_part = ''

Senthil Kumaran

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

[diff] [blame]

926

Martin Panter

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

[diff] [blame]

927

if query:

928

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

929

Senthil Kumaran

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

[diff] [blame]

930

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

931

collapsed_path = "/".join(splitpath)

932

933

return collapsed_path

934

Benjamin Peterson

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

[diff] [blame]

935

936

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]

946

except ImportError:

Georg Brandl

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

[diff] [blame]

947

return -1

948

try:

949

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

950

except KeyError:

Georg Brandl

cbd2ab1

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

[diff] [blame]

951

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):

956

"""Test for executable file."""

Victor Stinner

fb25ba9

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

[diff] [blame]

957

return os.access(path, os.X_OK)

Georg Brandl

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

[diff] [blame]

958

959

960

class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):

961

962

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

963

964

GET and HEAD also support running CGI scripts.

965

966

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

"""

# Determine platform specifics

971

have_fork = hasattr(os, 'fork')

Georg Brandl

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

[diff] [blame]

972

973

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

974

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

rbufsize = 0

def do_POST(self):

"""Serve a POST request.

979

980

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]

987

self.send_error(

988

HTTPStatus.NOT_IMPLEMENTED,

989

"Can only POST to CGI scripts")

Georg Brandl

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

[diff] [blame]

990

991

def send_head(self):

992

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

993

if self.is_cgi():

994

return self.run_cgi()

995

else:

996

return SimpleHTTPRequestHandler.send_head(self)

997

998

def is_cgi(self):

999

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

1000

Benjamin Peterson

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

[diff] [blame]

1001

Returns True and updates the cgi_info attribute to the tuple

1002

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

1003

Returns False otherwise.

Georg Brandl

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

[diff] [blame]

1004

Benjamin Peterson

a7deeee

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

[diff] [blame]

1005

If any exception is raised, the caller should assume that

1006

self.path was rejected as invalid and act accordingly.

1007

Benjamin Peterson

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

[diff] [blame]

1008

The default implementation tests whether the normalized url

1009

path begins with one of the strings in self.cgi_directories

1010

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

Georg Brandl

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

[diff] [blame]

1011

1012

"""

Martin Panter

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

[diff] [blame]

1013

collapsed_path = _url_collapse_path(self.path)

Senthil Kumaran

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

[diff] [blame]

1014

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

Siwon Kang

91daa9d

2019-11-22 18:13:05 +0900

[diff] [blame]

1015

while dir_sep > 0 and not collapsed_path[:dir_sep] in self.cgi_directories:

1016

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

1017

if dir_sep > 0:

1018

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

Senthil Kumaran

dbb369d

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

[diff] [blame]

1019

self.cgi_info = head, tail

Benjamin Peterson

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

[diff] [blame]

1020

return True

Georg Brandl

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

[diff] [blame]

1021

return False

1022

Senthil Kumaran

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

[diff] [blame]

1023

Georg Brandl

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

[diff] [blame]

1024

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

1025

1026

def is_executable(self, path):

1027

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

1028

return executable(path)

1029

1030

def is_python(self, path):

1031

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

1032

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

1033

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

1034

1035

def run_cgi(self):

1036

"""Execute a CGI script."""

Georg Brandl

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

[diff] [blame]

1037

dir, rest = self.cgi_info

Ned Deily

915a30f

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

[diff] [blame]

1038

path = dir + '/' + rest

1039

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

Georg Brandl

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

[diff] [blame]

1040

while i >= 0:

Ned Deily

915a30f

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

[diff] [blame]

1041

nextdir = path[:i]

1042

nextrest = path[i+1:]

Georg Brandl

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

[diff] [blame]

1043

1044

scriptdir = self.translate_path(nextdir)

1045

if os.path.isdir(scriptdir):

1046

dir, rest = nextdir, nextrest

Ned Deily

915a30f

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

[diff] [blame]

1047

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]

1052

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

Georg Brandl

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

[diff] [blame]

1053

1054

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

1055

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

1056

i = rest.find('/')

1057

if i >= 0:

1058

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

1059

else:

1060

script, rest = rest, ''

1061

1062

scriptname = dir + '/' + script

1063

scriptfile = self.translate_path(scriptname)

1064

if not os.path.exists(scriptfile):

Serhiy Storchaka

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

[diff] [blame]

1065

self.send_error(

1066

HTTPStatus.NOT_FOUND,

1067

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

Georg Brandl

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

[diff] [blame]

1068

return

1069

if not os.path.isfile(scriptfile):

Serhiy Storchaka

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

[diff] [blame]

1070

self.send_error(

1071

HTTPStatus.FORBIDDEN,

1072

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

Georg Brandl

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

[diff] [blame]

1073

return

1074

ispy = self.is_python(scriptname)

Victor Stinner

fb25ba9

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

[diff] [blame]

1075

if self.have_fork or not ispy:

Georg Brandl

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

[diff] [blame]

1076

if not self.is_executable(scriptfile):

Serhiy Storchaka

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

[diff] [blame]

1077

self.send_error(

1078

HTTPStatus.FORBIDDEN,

1079

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

Georg Brandl

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

[diff] [blame]

1080

return

1081

1082

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

1083

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

Senthil Kumaran

4271372

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

[diff] [blame]

1084

env = copy.deepcopy(os.environ)

Georg Brandl

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

[diff] [blame]

1085

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

1086

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

1087

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

1088

env['SERVER_PROTOCOL'] = self.protocol_version

1089

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

1090

env['REQUEST_METHOD'] = self.command

Jeremy Hylton

1afc169

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

[diff] [blame]

1091

uqrest = urllib.parse.unquote(rest)

Georg Brandl

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

[diff] [blame]

1092

env['PATH_INFO'] = uqrest

1093

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

1094

env['SCRIPT_NAME'] = scriptname

Senthil Kumaran

3ec9d01

2020-12-02 19:48:14 -0800

[diff] [blame]

1095

env['QUERY_STRING'] = query

Georg Brandl

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

[diff] [blame]

1096

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

Barry Warsaw

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

[diff] [blame]

1097

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

Georg Brandl

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

[diff] [blame]

1098

if authorization:

1099

authorization = authorization.split()

1100

if len(authorization) == 2:

1101

import base64, binascii

1102

env['AUTH_TYPE'] = authorization[0]

1103

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

1104

try:

1105

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

Georg Brandl

706824f

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

[diff] [blame]

1106

authorization = base64.decodebytes(authorization).\

Georg Brandl

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

[diff] [blame]

1107

decode('ascii')

1108

except (binascii.Error, UnicodeError):

1109

pass

1110

else:

1111

authorization = authorization.split(':')

1112

if len(authorization) == 2:

1113

env['REMOTE_USER'] = authorization[0]

1114

# XXX REMOTE_IDENT

Barry Warsaw

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

[diff] [blame]

1115

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

1116

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

Georg Brandl

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

[diff] [blame]

1117

else:

Barry Warsaw

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

[diff] [blame]

1118

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

1119

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

Georg Brandl

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

[diff] [blame]

1120

if length:

1121

env['CONTENT_LENGTH'] = length

Barry Warsaw

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

[diff] [blame]

1122

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

Georg Brandl

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

[diff] [blame]

1123

if referer:

1124

env['HTTP_REFERER'] = referer

Senthil Kumaran

da3d2ab

2020-12-05 05:26:24 -0800

[diff] [blame]

1125

accept = self.headers.get_all('accept', ())

Georg Brandl

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

[diff] [blame]

1126

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

Barry Warsaw

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

[diff] [blame]

1127

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

Georg Brandl

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

[diff] [blame]

1128

if ua:

1129

env['HTTP_USER_AGENT'] = ua

Barry Warsaw

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

[diff] [blame]

1130

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

Georg Brandl

62e2ca2

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

[diff] [blame]

1131

cookie_str = ', '.join(co)

1132

if cookie_str:

1133

env['HTTP_COOKIE'] = cookie_str

Georg Brandl

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

[diff] [blame]

1134

# XXX Other HTTP_* headers

1135

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

1136

# values to override previously set values

1137

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

1138

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

1139

env.setdefault(k, "")

Georg Brandl

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

[diff] [blame]

1140

Serhiy Storchaka

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

[diff] [blame]

1141

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

Senthil Kumaran

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

[diff] [blame]

1142

self.flush_headers()

Georg Brandl

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

[diff] [blame]

1143

1144

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

1145

1146

if self.have_fork:

1147

# Unix -- fork as we should

1148

args = [script]

1149

if '=' not in decoded_query:

1150

args.append(decoded_query)

1151

nobody = nobody_uid()

1152

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

pid = os.fork()

if pid != 0:

# Parent

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

1157

# throw away additional data [see bug #427345]

1158

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

1159

if not self.rfile.read(1):

1160

break

Victor Stinner

9a679a0

2020-04-02 03:42:05 +0200

[diff] [blame]

1161

exitcode = os.waitstatus_to_exitcode(sts)

1162

if exitcode:

1163

self.log_error(f"CGI script exit code {exitcode}")

Georg Brandl

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

[diff] [blame]

return

# Child

try:

try:

os.setuid(nobody)

Andrew Svetlov

ad28c7f

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

[diff] [blame]

1169

except OSError:

Georg Brandl

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

[diff] [blame]

1170

pass

1171

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

1172

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

Senthil Kumaran

4271372

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

[diff] [blame]

1173

os.execve(scriptfile, args, env)

Georg Brandl

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

[diff] [blame]

1174

except:

1175

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

1176

os._exit(127)

1177

Amaury Forgeot d'Arc

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

[diff] [blame]

1178

else:

1179

# Non-Unix -- use subprocess

1180

import subprocess

Senthil Kumaran

e29cd16

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

[diff] [blame]

1181

cmdline = [scriptfile]

Georg Brandl

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

[diff] [blame]

1182

if self.is_python(scriptfile):

1183

interp = sys.executable

1184

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

1185

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

1186

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

Senthil Kumaran

e29cd16

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

[diff] [blame]

1187

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

1188

if '=' not in query:

1189

cmdline.append(query)

1190

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

Georg Brandl

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

[diff] [blame]

1191

try:

1192

nbytes = int(length)

1193

except (TypeError, ValueError):

1194

nbytes = 0

Amaury Forgeot d'Arc

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

[diff] [blame]

1195

p = subprocess.Popen(cmdline,

1196

stdin=subprocess.PIPE,

1197

stdout=subprocess.PIPE,

Senthil Kumaran

4271372

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

[diff] [blame]

1198

stderr=subprocess.PIPE,

1199

env = env

Amaury Forgeot d'Arc

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

[diff] [blame]

1200

)

Georg Brandl

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

[diff] [blame]

1201

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

1202

data = self.rfile.read(nbytes)

Amaury Forgeot d'Arc

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

[diff] [blame]

1203

else:

1204

data = None

Georg Brandl

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

[diff] [blame]

1205

# throw away additional data [see bug #427345]

1206

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

1207

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

1208

break

Amaury Forgeot d'Arc

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

[diff] [blame]

1209

stdout, stderr = p.communicate(data)

1210

self.wfile.write(stdout)

1211

if stderr:

1212

self.log_error('%s', stderr)

Brian Curtin

cbad4df

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

[diff] [blame]

1213

p.stderr.close()

1214

p.stdout.close()

Amaury Forgeot d'Arc

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

[diff] [blame]

1215

status = p.returncode

1216

if status:

1217

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

Georg Brandl

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

[diff] [blame]

1218

else:

1219

self.log_message("CGI script exited OK")

1220

1221

Jason R. Coombs

2019-02-07 08:22:45 -0500

[diff] [blame]

1222

def _get_best_family(*address):

1223

infos = socket.getaddrinfo(

1224

*address,

1225

type=socket.SOCK_STREAM,

1226

flags=socket.AI_PASSIVE,

1227

)

1228

family, type, proto, canonname, sockaddr = next(iter(infos))

1229

return family, sockaddr

1230

1231

Senthil Kumaran

defe7f4

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

[diff] [blame]

1232

def test(HandlerClass=BaseHTTPRequestHandler,

Géry Ogam

1cee216

2018-05-29 22:10:30 +0200

[diff] [blame]

1233

ServerClass=ThreadingHTTPServer,

Jason R. Coombs

2019-02-07 08:22:45 -0500

[diff] [blame]

1234

protocol="HTTP/1.0", port=8000, bind=None):

Georg Brandl

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

[diff] [blame]

1235

"""Test the HTTP request handler class.

1236

Robert Collins

9644f24

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

[diff] [blame]

1237

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

Georg Brandl

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

[diff] [blame]

1238

1239

"""

Jason R. Coombs

2019-02-07 08:22:45 -0500

[diff] [blame]

1240

ServerClass.address_family, addr = _get_best_family(bind, port)

Lisa Roach

433433f

2018-11-26 10:43:38 -0800

[diff] [blame]

1241

Georg Brandl

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

[diff] [blame]

1242

HandlerClass.protocol_version = protocol

Jason R. Coombs

2019-02-07 08:22:45 -0500

[diff] [blame]

1243

with ServerClass(addr, HandlerClass) as httpd:

1244

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

1245

url_host = f'[{host}]' if ':' in host else host

1246

print(

1247

f"Serving HTTP on {host} port {port} "

1248

f"(http://{url_host}:{port}/) ..."

1249

)

Martin Panter

0cab9c1

2016-04-13 00:36:52 +0000

[diff] [blame]

1250

try:

1251

httpd.serve_forever()

1252

except KeyboardInterrupt:

1253

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

1254

sys.exit(0)

Georg Brandl

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

[diff] [blame]

1255

1256

if __name__ == '__main__':

Serhiy Storchaka

7e4db2f

2017-05-04 08:17:47 +0300

[diff] [blame]

1257

import argparse

1258

Senthil Kumaran

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

[diff] [blame]

1259

parser = argparse.ArgumentParser()

1260

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

1261

help='Run as CGI Server')

Jason R. Coombs

2019-02-07 08:22:45 -0500

[diff] [blame]

1262

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

Senthil Kumaran

defe7f4

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

[diff] [blame]

1263

help='Specify alternate bind address '

1264

'[default: all interfaces]')

Stéphane Wirtel

2017-05-24 09:29:06 +0200

[diff] [blame]

1265

parser.add_argument('--directory', '-d', default=os.getcwd(),

1266

help='Specify alternative directory '

1267

'[default:current directory]')

Senthil Kumaran

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

[diff] [blame]

1268

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

1269

default=8000, type=int,

1270

nargs='?',

1271

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

1272

args = parser.parse_args()

1273

if args.cgi:

Senthil Kumaran

defe7f4

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

[diff] [blame]

1274

handler_class = CGIHTTPRequestHandler

Senthil Kumaran

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

[diff] [blame]

1275

else:

Stéphane Wirtel