aboutsummaryrefslogtreecommitdiff
path: root/docs/INTERNALS
blob: e667c330767220a79ddde9ea0254218a3a44ce15 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
                                  _   _ ____  _
                              ___| | | |  _ \| |
                             / __| | | | |_) | |
                            | (__| |_| |  _ <| |___
                             \___|\___/|_| \_\_____|

INTERNALS

 The project is split in two. The library and the client. The client part uses
 the library, but the library is designed to allow other applications to use
 it.

 The largest amount of code and complexity is in the library part.

GIT
===
 All changes to the sources are committed to the git repository as soon as
 they're somewhat verified to work. Changes shall be commited as independently
 as possible so that individual changes can be easier spotted and tracked
 afterwards.

 Tagging shall be used extensively, and by the time we release new archives we
 should tag the sources with a name similar to the released version number.

Portability
===========

 We write curl and libcurl to compile with C89 compilers.  On 32bit and up
 machines. Most of libcurl assumes more or less POSIX compliance but that's
 not a requirement.

 We write libcurl to build and work with lots of third party tools, and we
 want it to remain functional and buildable with these and later versions
 (older versions may still work but is not what we work hard to maintain):

 OpenSSL      0.9.6
 GnuTLS       1.2
 zlib         1.1.4
 libssh2      0.16
 c-ares       1.6.0
 libidn       0.4.1
 *yassl       1.4.0 (http://curl.haxx.se/mail/lib-2008-02/0093.html)
 openldap     2.0
 MIT krb5 lib 1.2.4
 qsossl       V5R2M0
 NSS          3.11.x
 axTLS        1.2.7
 Heimdal      ?

 * = only partly functional, but that's due to bugs in the third party lib, not
     because of libcurl code

 On systems where configure runs, we aim at working on them all - if they have
 a suitable C compiler. On systems that don't run configure, we strive to keep
 curl running fine on:

 Windows      98
 AS/400       V5R2M0
 Symbian      9.1
 Windows CE   ?
 TPF          ?

 When writing code (mostly for generating stuff included in release tarballs)
 we use a few "build tools" and we make sure that we remain functional with
 these versions:

 GNU Libtool  1.4.2
 GNU Autoconf 2.57
 GNU Automake 1.7 (we currently avoid 1.10 due to Solaris-related bugs)
 GNU M4       1.4
 perl         4
 roffit       0.5
 groff        ? (any version that supports "groff -Tps -man [in] [out]")
 ps2pdf (gs)  ?

Windows vs Unix
===============

 There are a few differences in how to program curl the unix way compared to
 the Windows way. The four perhaps most notable details are:

 1. Different function names for socket operations.

   In curl, this is solved with defines and macros, so that the source looks
   the same at all places except for the header file that defines them. The
   macros in use are sclose(), sread() and swrite().

 2. Windows requires a couple of init calls for the socket stuff.

   That's taken care of by the curl_global_init() call, but if other libs also
   do it etc there might be reasons for applications to alter that behaviour.

 3. The file descriptors for network communication and file operations are
    not easily interchangable as in unix.

   We avoid this by not trying any funny tricks on file descriptors.

 4. When writing data to stdout, Windows makes end-of-lines the DOS way, thus
    destroying binary data, although you do want that conversion if it is
    text coming through... (sigh)

   We set stdout to binary under windows

 Inside the source code, We make an effort to avoid '#ifdef [Your OS]'. All
 conditionals that deal with features *should* instead be in the format
 '#ifdef HAVE_THAT_WEIRD_FUNCTION'. Since Windows can't run configure scripts,
 we maintain two curl_config-win32.h files (one in lib/ and one in src/) that
 are supposed to look exactly as a curl_config.h file would have looked like on
 a Windows machine!

 Generally speaking: always remember that this will be compiled on dozens of
 operating systems. Don't walk on the edge.

Library
=======

 There are plenty of entry points to the library, namely each publicly defined
 function that libcurl offers to applications. All of those functions are
 rather small and easy-to-follow. All the ones prefixed with 'curl_easy' are
 put in the lib/easy.c file.

 curl_global_init_() and curl_global_cleanup() should be called by the
 application to initialize and clean up global stuff in the library. As of
 today, it can handle the global SSL initing if SSL is enabled and it can init
 the socket layer on windows machines. libcurl itself has no "global" scope.

 All printf()-style functions use the supplied clones in lib/mprintf.c. This
 makes sure we stay absolutely platform independent.

 curl_easy_init() allocates an internal struct and makes some initializations.
 The returned handle does not reveal internals. This is the 'SessionHandle'
 struct which works as an "anchor" struct for all curl_easy functions. All
 connections performed will get connect-specific data allocated that should be
 used for things related to particular connections/requests.

 curl_easy_setopt() takes three arguments, where the option stuff must be
 passed in pairs: the parameter-ID and the parameter-value. The list of
 options is documented in the man page. This function mainly sets things in
 the 'SessionHandle' struct.

 curl_easy_perform() does a whole lot of things:

 It starts off in the lib/easy.c file by calling Curl_perform() and the main
 work then continues in lib/url.c. The flow continues with a call to
 Curl_connect() to connect to the remote site.

 o Curl_connect()

   ... analyzes the URL, it separates the different components and connects to
   the remote host. This may involve using a proxy and/or using SSL. The
   Curl_resolv() function in lib/hostip.c is used for looking up host names
   (it does then use the proper underlying method, which may vary between
   platforms and builds).

   When Curl_connect is done, we are connected to the remote site. Then it is
   time to tell the server to get a document/file. Curl_do() arranges this.

   This function makes sure there's an allocated and initiated 'connectdata'
   struct that is used for this particular connection only (although there may
   be several requests performed on the same connect). A bunch of things are
   inited/inherited from the SessionHandle struct.

 o Curl_do()

   Curl_do() makes sure the proper protocol-specific function is called. The
   functions are named after the protocols they handle. Curl_ftp(),
   Curl_http(), Curl_dict(), etc. They all reside in their respective files
   (ftp.c, http.c and dict.c). HTTPS is handled by Curl_http() and FTPS by
   Curl_ftp().

   The protocol-specific functions of course deal with protocol-specific
   negotiations and setup. They have access to the Curl_sendf() (from
   lib/sendf.c) function to send printf-style formatted data to the remote
   host and when they're ready to make the actual file transfer they call the
   Curl_Transfer() function (in lib/transfer.c) to setup the transfer and
   returns.

   If this DO function fails and the connection is being re-used, libcurl will
   then close this connection, setup a new connection and re-issue the DO
   request on that. This is because there is no way to be perfectly sure that
   we have discovered a dead connection before the DO function and thus we
   might wrongly be re-using a connection that was closed by the remote peer.

   Some time during the DO function, the Curl_setup_transfer() function must
   be called with some basic info about the upcoming transfer: what socket(s)
   to read/write and the expected file tranfer sizes (if known).

 o Transfer()

   Curl_perform() then calls Transfer() in lib/transfer.c that performs the
   entire file transfer.

   During transfer, the progress functions in lib/progress.c are called at a
   frequent interval (or at the user's choice, a specified callback might get
   called). The speedcheck functions in lib/speedcheck.c are also used to
   verify that the transfer is as fast as required.

 o Curl_done()

   Called after a transfer is done. This function takes care of everything
   that has to be done after a transfer. This function attempts to leave
   matters in a state so that Curl_do() should be possible to call again on
   the same connection (in a persistent connection case). It might also soon
   be closed with Curl_disconnect().

 o Curl_disconnect()

   When doing normal connections and transfers, no one ever tries to close any
   connections so this is not normally called when curl_easy_perform() is
   used. This function is only used when we are certain that no more transfers
   is going to be made on the connection. It can be also closed by force, or
   it can be called to make sure that libcurl doesn't keep too many
   connections alive at the same time (there's a default amount of 5 but that
   can be changed with the CURLOPT_MAXCONNECTS option).

   This function cleans up all resources that are associated with a single
   connection.

 Curl_perform() is the function that does the main "connect - do - transfer -
 done" loop. It loops if there's a Location: to follow.

 When completed, the curl_easy_cleanup() should be called to free up used
 resources. It runs Curl_disconnect() on all open connectons.

 A quick roundup on internal function sequences (many of these call
 protocol-specific function-pointers):

  curl_connect - connects to a remote site and does initial connect fluff
   This also checks for an existing connection to the requested site and uses
   that one if it is possible.

   curl_do - starts a transfer
    curl_transfer() - transfers data
   curl_done - ends a transfer

  curl_disconnect - disconnects from a remote site. This is called when the
   disconnect is really requested, which doesn't necessarily have to be
   exactly after curl_done in case we want to keep the connection open for
   a while.

 HTTP(S)

 HTTP offers a lot and is the protocol in curl that uses the most lines of
 code. There is a special file (lib/formdata.c) that offers all the multipart
 post functions.

 base64-functions for user+password stuff (and more) is in (lib/base64.c) and
 all functions for parsing and sending cookies are found in (lib/cookie.c).

 HTTPS uses in almost every means the same procedure as HTTP, with only two
 exceptions: the connect procedure is different and the function used to read
 or write from the socket is different, although the latter fact is hidden in
 the source by the use of curl_read() for reading and curl_write() for writing
 data to the remote server.

 http_chunks.c contains functions that understands HTTP 1.1 chunked transfer
 encoding.

 An interesting detail with the HTTP(S) request, is the add_buffer() series of
 functions we use. They append data to one single buffer, and when the
 building is done the entire request is sent off in one single write. This is
 done this way to overcome problems with flawed firewalls and lame servers.

 FTP

 The Curl_if2ip() function can be used for getting the IP number of a
 specified network interface, and it resides in lib/if2ip.c.

 Curl_ftpsendf() is used for sending FTP commands to the remote server. It was
 made a separate function to prevent us programmers from forgetting that they
 must be CRLF terminated. They must also be sent in one single write() to make
 firewalls and similar happy.

 Kerberos

 The kerberos support is mainly in lib/krb4.c and lib/security.c.

 TELNET

 Telnet is implemented in lib/telnet.c.

 FILE

 The file:// protocol is dealt with in lib/file.c.

 LDAP

 Everything LDAP is in lib/ldap.c.

 GENERAL

 URL encoding and decoding, called escaping and unescaping in the source code,
 is found in lib/escape.c.

 While transfering data in Transfer() a few functions might get used.
 curl_getdate() in lib/parsedate.c is for HTTP date comparisons (and more).

 lib/getenv.c offers curl_getenv() which is for reading environment variables
 in a neat platform independent way. That's used in the client, but also in
 lib/url.c when checking the proxy environment variables. Note that contrary
 to the normal unix getenv(), this returns an allocated buffer that must be
 free()ed after use.

 lib/netrc.c holds the .netrc parser

 lib/timeval.c features replacement functions for systems that don't have
 gettimeofday() and a few support functions for timeval convertions.

 A function named curl_version() that returns the full curl version string is
 found in lib/version.c.

Persistent Connections
======================

 The persistent connection support in libcurl requires some considerations on
 how to do things inside of the library.

 o The 'SessionHandle' struct returned in the curl_easy_init() call must never
   hold connection-oriented data. It is meant to hold the root data as well as
   all the options etc that the library-user may choose.
 o The 'SessionHandle' struct holds the "connection cache" (an array of
   pointers to 'connectdata' structs). There's one connectdata struct
   allocated for each connection that libcurl knows about. Note that when you
   use the multi interface, the multi handle will hold the connection cache
   and not the particular easy handle. This of course to allow all easy handles
   in a multi stack to be able to share and re-use connections.
 o This enables the 'curl handle' to be reused on subsequent transfers.
 o When we are about to perform a transfer with curl_easy_perform(), we first
   check for an already existing connection in the cache that we can use,
   otherwise we create a new one and add to the cache. If the cache is full
   already when we add a new connection, we close one of the present ones. We
   select which one to close dependent on the close policy that may have been
   previously set.
 o When the transfer operation is complete, we try to leave the connection
   open. Particular options may tell us not to, and protocols may signal
   closure on connections and then we don't keep it open of course.
 o When curl_easy_cleanup() is called, we close all still opened connections,
   unless of course the multi interface "owns" the connections.

 You do realize that the curl handle must be re-used in order for the
 persistent connections to work.

multi interface/non-blocking
============================

 We make an effort to provide a non-blocking interface to the library, the
 multi interface. To make that interface work as good as possible, no
 low-level functions within libcurl must be written to work in a blocking
 manner.

 One of the primary reasons we introduced c-ares support was to allow the name
 resolve phase to be perfectly non-blocking as well.

 The ultimate goal is to provide the easy interface simply by wrapping the
 multi interface functions and thus treat everything internally as the multi
 interface is the single interface we have.

 The FTP and the SFTP/SCP protocols are thus perfect examples of how we adapt
 and adjust the code to allow non-blocking operations even on multi-stage
 protocols. The DICT, LDAP and TELNET are crappy examples and they are subject
 for rewrite in the future to better fit the libcurl protocol family.

SSL libraries
=============

 Originally libcurl supported SSLeay for SSL/TLS transports, but that was then
 extended to its successor OpenSSL but has since also been extended to several
 other SSL/TLS libraries and we expect and hope to further extend the support
 in future libcurl versions.

 To deal with this internally in the best way possible, we have a generic SSL
 function API as provided by the sslgen.[ch] system, and they are the only SSL
 functions we must use from within libcurl. sslgen is then crafted to use the
 appropriate lower-level function calls to whatever SSL library that is in
 use.

Library Symbols
===============

 All symbols used internally in libcurl must use a 'Curl_' prefix if they're
 used in more than a single file. Single-file symbols must be made static.
 Public ("exported") symbols must use a 'curl_' prefix. (There are exceptions,
 but they are to be changed to follow this pattern in future versions.)

Return Codes and Informationals
===============================

 I've made things simple. Almost every function in libcurl returns a CURLcode,
 that must be CURLE_OK if everything is OK or otherwise a suitable error code
 as the curl/curl.h include file defines. The very spot that detects an error
 must use the Curl_failf() function to set the human-readable error
 description.

 In aiding the user to understand what's happening and to debug curl usage, we
 must supply a fair amount of informational messages by using the Curl_infof()
 function. Those messages are only displayed when the user explicitly asks for
 them. They are best used when revealing information that isn't otherwise
 obvious.

API/ABI
=======

 We make an effort to not export or show internals or how internals work, as
 that makes it easier to keep a solid API/ABI over time. See docs/libcurl/ABI
 for our promise to users.

Client
======

 main() resides in src/main.c together with most of the client code.

 src/hugehelp.c is automatically generated by the mkhelp.pl perl script to
 display the complete "manual" and the src/urlglob.c file holds the functions
 used for the URL-"globbing" support. Globbing in the sense that the {} and []
 expansion stuff is there.

 The client mostly messes around to setup its 'config' struct properly, then
 it calls the curl_easy_*() functions of the library and when it gets back
 control after the curl_easy_perform() it cleans up the library, checks status
 and exits.

 When the operation is done, the ourWriteOut() function in src/writeout.c may
 be called to report about the operation. That function is using the
 curl_easy_getinfo() function to extract useful information from the curl
 session.

 Recent versions may loop and do all this several times if many URLs were
 specified on the command line or config file.

Memory Debugging
================

 The file lib/memdebug.c contains debug-versions of a few functions. Functions
 such as malloc, free, fopen, fclose, etc that somehow deal with resources
 that might give us problems if we "leak" them. The functions in the memdebug
 system do nothing fancy, they do their normal function and then log
 information about what they just did. The logged data can then be analyzed
 after a complete session,

 memanalyze.pl is the perl script present in tests/ that analyzes a log file
 generated by the memory tracking system. It detects if resources are
 allocated but never freed and other kinds of errors related to resource
 management.

 Internally, definition of preprocessor symbol DEBUGBUILD restricts code which
 is only compiled for debug enabled builds. And symbol CURLDEBUG is used to
 differentiate code which is _only_ used for memory tracking/debugging.

 Use -DCURLDEBUG when compiling to enable memory debugging, this is also
 switched on by running configure with --enable-curldebug. Use -DDEBUGBUILD
 when compiling to enable a debug build or run configure with --enable-debug.

 curl --version will list 'Debug' feature for debug enabled builds, and
 will list 'TrackMemory' feature for curl debug memory tracking capable
 builds. These features are independent and can be controlled when running
 the configure script. When --enable-debug is given both features will be
 enabled, unless some restriction prevents memory tracking from being used.

Test Suite
==========

 Since November 2000, a test suite has evolved. It is placed in its own
 subdirectory directly off the root in the curl archive tree, and it contains
 a bunch of scripts and a lot of test case data.

 The main test script is runtests.pl that will invoke the two servers
 httpserver.pl and ftpserver.pl before all the test cases are performed. The
 test suite currently only runs on unix-like platforms.

 You'll find a complete description of the test case data files in the
 tests/README file.

 The test suite automatically detects if curl was built with the memory
 debugging enabled, and if it was it will detect memory leaks too.

Building Releases
=================

 There's no magic to this. When you consider everything stable enough to be
 released, run the 'maketgz' script (using 'make distcheck' will give you a
 pretty good view on the status of the current sources). maketgz prompts for
 version number of the client and the library before it creates a release
 archive. maketgz uses 'make dist' for the actual archive building, why you
 need to fill in the Makefile.am files properly for which files that should
 be included in the release archives.

 NOTE: you need to have curl checked out from git to be able to do a proper
 release build. The release tarballs do not have everything setup in order to
 do releases properly.