1 $Id: libcurl-the-guide,v 1.23 2004/01/29 16:17:25 bagder Exp $
6 \___|\___/|_| \_\_____|
8 PROGRAMMING WITH LIBCURL
12 This document attempts to describe the general principles and some basic
13 approaches to consider when programming with libcurl. The text will focus
14 mainly on the C interface but might apply fairly well on other interfaces as
15 well as they usually follow the C one pretty closely.
17 This document will refer to 'the user' as the person writing the source code
18 that uses libcurl. That would probably be you or someone in your position.
19 What will be generally refered to as 'the program' will be the collected
20 source code that you write that is using libcurl for transfers. The program
21 is outside libcurl and libcurl is outside of the program.
23 To get the more details on all options and functions described herein, please
24 refer to their respective man pages.
28 There are many different ways to build C programs. This chapter will assume a
29 unix-style build process. If you use a different build system, you can still
30 read this to get general information that may apply to your environment as
35 Your compiler needs to know where the libcurl headers are
36 located. Therefore you must set your compiler's include path to point to
37 the directory where you installed them. The 'curl-config'[3] tool can be
38 used to get this information:
40 $ curl-config --cflags
42 Linking the Program with libcurl
44 When having compiled the program, you need to link your object files to
45 create a single executable. For that to succeed, you need to link with
46 libcurl and possibly also with other libraries that libcurl itself depends
47 on. Like OpenSSL librararies, but even some standard OS libraries may be
48 needed on the command line. To figure out which flags to use, once again
49 the 'curl-config' tool comes to the rescue:
55 libcurl can be built and customized in many ways. One of the things that
56 varies from different libraries and builds is the support for SSL-based
57 transfers, like HTTPS and FTPS. If OpenSSL was detected properly at
58 build-time, libcurl will be built with SSL support. To figure out if an
59 installed libcurl has been built with SSL support enabled, use
60 'curl-config' like this:
62 $ curl-config --feature
64 And if SSL is supported, the keyword 'SSL' will be written to stdout,
65 possibly together with a few other features that can be on and off on
69 Portable Code in a Portable World
71 The people behind libcurl have put a considerable effort to make libcurl work
72 on a large amount of different operating systems and environments.
74 You program libcurl the same way on all platforms that libcurl runs on. There
75 are only very few minor considerations that differs. If you just make sure to
76 write your code portable enough, you may very well create yourself a very
77 portable program. libcurl shouldn't stop you from that.
82 The program must initialize some of the libcurl functionality globally. That
83 means it should be done exactly once, no matter how many times you intend to
84 use the library. Once for your program's entire life time. This is done using
88 and it takes one parameter which is a bit pattern that tells libcurl what to
89 intialize. Using CURL_GLOBAL_ALL will make it initialize all known internal
90 sub modules, and might be a good default option. The current two bits that
93 CURL_GLOBAL_WIN32 which only does anything on Windows machines. When used on
94 a Windows machine, it'll make libcurl intialize the win32 socket
95 stuff. Without having that initialized properly, your program cannot use
96 sockets properly. You should only do this once for each application, so if
97 your program already does this or of another library in use does it, you
98 should not tell libcurl to do this as well.
100 CURL_GLOBAL_SSL which only does anything on libcurls compiled and built
101 SSL-enabled. On these systems, this will make libcurl init OpenSSL properly
102 for this application. This is only needed to do once for each application so
103 if your program or another library already does this, this bit should not be
106 libcurl has a default protection mechanism that detects if curl_global_init()
107 hasn't been called by the time curl_easy_perform() is called and if that is
108 the case, libcurl runs the function itself with a guessed bit pattern. Please
109 note that depending solely on this is not considered nice nor very good.
111 When the program no longer uses libcurl, it should call
112 curl_global_cleanup(), which is the opposite of the init call. It will then
113 do the reversed operations to cleanup the resources the curl_global_init()
116 Repeated calls to curl_global_init() and curl_global_cleanup() should be
117 avoided. They should only be called once each.
120 Handle the Easy libcurl
122 libcurl version 7 is oriented around the so called easy interface. All
123 operations in the easy interface are prefixed with 'curl_easy'.
125 Future libcurls will also offer the multi interface. More about that
126 interface, what it is targeted for and how to use it is still only debated on
127 the libcurl mailing list and developer web pages. Join up to discuss and
130 To use the easy interface, you must first create yourself an easy handle. You
131 need one handle for each easy session you want to perform. Basicly, you
132 should use one handle for every thread you plan to use for transferring. You
133 must never share the same handle in multiple threads.
135 Get an easy handle with
137 easyhandle = curl_easy_init();
139 It returns an easy handle. Using that you proceed to the next step: setting
140 up your preferred actions. A handle is just a logic entity for the upcoming
141 transfer or series of transfers.
143 You set properties and options for this handle using curl_easy_setopt(). They
144 control how the subsequent transfer or transfers will be made. Options remain
145 set in the handle until set again to something different. Alas, multiple
146 requests using the same handle will use the same options.
148 Many of the informationals you set in libcurl are "strings", pointers to data
149 terminated with a zero byte. Keep in mind that when you set strings with
150 curl_easy_setopt(), libcurl will not copy the data. It will merely point to
151 the data. You MUST make sure that the data remains available for libcurl to
152 use until finished or until you use the same option again to point to
155 One of the most basic properties to set in the handle is the URL. You set
156 your preferred URL to transfer with CURLOPT_URL in a manner similar to:
158 curl_easy_setopt(easyhandle, CURLOPT_URL, "http://curl.haxx.se/");
160 Let's assume for a while that you want to receive data as the URL indentifies
161 a remote resource you want to get here. Since you write a sort of application
162 that needs this transfer, I assume that you would like to get the data passed
163 to you directly instead of simply getting it passed to stdout. So, you write
164 your own function that matches this prototype:
166 size_t write_data(void *buffer, size_t size, size_t nmemb, void *userp);
168 You tell libcurl to pass all data to this function by issuing a function
171 curl_easy_setopt(easyhandle, CURLOPT_WRITEFUNCTION, write_data);
173 You can control what data your function get in the forth argument by setting
176 curl_easy_setopt(easyhandle, CURLOPT_FILE, &internal_struct);
178 Using that property, you can easily pass local data between your application
179 and the function that gets invoked by libcurl. libcurl itself won't touch the
180 data you pass with CURLOPT_FILE.
182 libcurl offers its own default internal callback that'll take care of the
183 data if you don't set the callback with CURLOPT_WRITEFUNCTION. It will then
184 simply output the received data to stdout. You can have the default callback
185 write the data to a different file handle by passing a 'FILE *' to a file
186 opened for writing with the CURLOPT_FILE option.
188 Now, we need to take a step back and have a deep breath. Here's one of those
189 rare platform-dependent nitpicks. Did you spot it? On some platforms[2],
190 libcurl won't be able to operate on files opened by the program. Thus, if you
191 use the default callback and pass in a an open file with CURLOPT_FILE, it
192 will crash. You should therefore avoid this to make your program run fine
193 virtually everywhere.
195 There are of course many more options you can set, and we'll get back to a
196 few of them later. Let's instead continue to the actual transfer:
198 success = curl_easy_perform(easyhandle);
200 The curl_easy_perform() will connect to the remote site, do the necessary
201 commands and receive the transfer. Whenever it receives data, it calls the
202 callback function we previously set. The function may get one byte at a time,
203 or it may get many kilobytes at once. libcurl delivers as much as possible as
204 often as possible. Your callback function should return the number of bytes
205 it "took care of". If that is not the exact same amount of bytes that was
206 passed to it, libcurl will abort the operation and return with an error code.
208 When the transfer is complete, the function returns a return code that
209 informs you if it succeeded in its mission or not. If a return code isn't
210 enough for you, you can use the CURLOPT_ERRORBUFFER to point libcurl to a
211 buffer of yours where it'll store a human readable error message as well.
213 If you then want to transfer another file, the handle is ready to be used
214 again. Mind you, it is even preferred that you re-use an existing handle if
215 you intend to make another transfer. libcurl will then attempt to re-use the
219 Multi-threading issues
221 libcurl is completely thread safe, except for two issues: signals and alarm
222 handlers. Signals are needed for a SIGPIPE handler, and the alarm() syscall
223 is used to catch timeouts (mostly during DNS lookup).
225 If you are accessing HTTPS or FTPS URLs in a multi-threaded manner, you are
226 then of course using OpenSSL multi-threaded and it has itself a few
227 requirements on this. Basicly, you need to provide one or two functions to
228 allow it to function properly. For all details, see this:
230 http://www.openssl.org/docs/crypto/threads.html#DESCRIPTION
232 When using multiple threads you should first ignore SIGPIPE in your main
233 thread and set the CURLOPT_NOSIGNAL option to TRUE for all handles.
235 Everything will work fine except that timeouts are not honored during the DNS
236 lookup - which you can work around by building libcurl with ares-support.
237 Ares is a library that provides asynchronous name resolves. Unfortunately,
238 ares does not yet support IPv6.
240 For SIGPIPE info see the UNIX Socket FAQ at
241 http://www.unixguide.net/network/socketfaq/2.22.shtml
243 Also, note that CURLOPT_DNS_USE_GLOBAL_CACHE is not thread-safe.
247 There will always be times when the transfer fails for some reason. You might
248 have set the wrong libcurl option or misunderstood what the libcurl option
249 actually does, or the remote server might return non-standard replies that
250 confuse the library which then confuses your program.
252 There's one golden rule when these things occur: set the CURLOPT_VERBOSE
253 option to TRUE. It'll cause the library to spew out the entire protocol
254 details it sends, some internal info and some received protcol data as well
255 (especially when using FTP). If you're using HTTP, adding the headers in the
256 received output to study is also a clever way to get a better understanding
257 wht the server behaves the way it does. Include headers in the normal body
258 output with CURLOPT_HEADER set TRUE.
260 Of course there are bugs left. We need to get to know about them to be able
261 to fix them, so we're quite dependent on your bug reports! When you do report
262 suspected bugs in libcurl, please include as much details you possibly can: a
263 protocol dump that CURLOPT_VERBOSE produces, library version, as much as
264 possible of your code that uses libcurl, operating system name and version,
265 compiler name and version etc.
267 If CURLOPT_VERBOSE is not enough, you increase the level of debug data your
268 application receive by using the CURLOPT_DEBUGFUNCTION.
270 Getting some in-depth knowledge about the protocols involved is never wrong,
271 and if you're trying to do funny things, you might very well understand
272 libcurl and how to use it better if you study the appropriate RFC documents
276 Upload Data to a Remote Site
278 libcurl tries to keep a protocol independent approach to most transfers, thus
279 uploading to a remote FTP site is very similar to uploading data to a HTTP
280 server with a PUT request.
282 Of course, first you either create an easy handle or you re-use one existing
283 one. Then you set the URL to operate on just like before. This is the remote
284 URL, that we now will upload.
286 Since we write an application, we most likely want libcurl to get the upload
287 data by asking us for it. To make it do that, we set the read callback and
288 the custom pointer libcurl will pass to our read callback. The read callback
289 should have a prototype similar to:
291 size_t function(char *bufptr, size_t size, size_t nitems, void *userp);
293 Where bufptr is the pointer to a buffer we fill in with data to upload and
294 size*nitems is the size of the buffer and therefore also the maximum amount
295 of data we can return to libcurl in this call. The 'userp' pointer is the
296 custom pointer we set to point to a struct of ours to pass private data
297 between the application and the callback.
299 curl_easy_setopt(easyhandle, CURLOPT_READFUNCTION, read_function);
301 curl_easy_setopt(easyhandle, CURLOPT_INFILE, &filedata);
303 Tell libcurl that we want to upload:
305 curl_easy_setopt(easyhandle, CURLOPT_UPLOAD, TRUE);
307 A few protocols won't behave properly when uploads are done without any prior
308 knowledge of the expected file size. So, set the upload file size using the
309 CURLOPT_INFILESIZE_LARGE for all known file sizes like this[1]:
311 /* in this example, file_size must be an off_t variable */
312 curl_easy_setopt(easyhandle, CURLOPT_INFILESIZE_LARGE, file_size);
314 When you call curl_easy_perform() this time, it'll perform all the necessary
315 operations and when it has invoked the upload it'll call your supplied
316 callback to get the data to upload. The program should return as much data as
317 possible in every invoke, as that is likely to make the upload perform as
318 fast as possible. The callback should return the number of bytes it wrote in
319 the buffer. Returning 0 will signal the end of the upload.
324 Many protocols use or even require that user name and password are provided
325 to be able to download or upload the data of your choice. libcurl offers
326 several ways to specify them.
328 Most protocols support that you specify the name and password in the URL
329 itself. libcurl will detect this and use them accordingly. This is written
332 protocol://user:password@example.com/path/
334 If you need any odd letters in your user name or password, you should enter
335 them URL encoded, as %XX where XX is a two-digit hexadecimal number.
337 libcurl also provides options to set various passwords. The user name and
338 password as shown embedded in the URL can instead get set with the
339 CURLOPT_USERPWD option. The argument passed to libcurl should be a char * to
340 a string in the format "user:password:". In a manner like this:
342 curl_easy_setopt(easyhandle, CURLOPT_USERPWD, "myname:thesecret");
344 Another case where name and password might be needed at times, is for those
345 users who need to athenticate themselves to a proxy they use. libcurl offers
346 another option for this, the CURLOPT_PROXYUSERPWD. It is used quite similar
347 to the CURLOPT_USERPWD option like this:
349 curl_easy_setopt(easyhandle, CURLOPT_PROXYUSERPWD, "myname:thesecret");
351 There's a long time unix "standard" way of storing ftp user names and
352 passwords, namely in the $HOME/.netrc file. The file should be made private
353 so that only the user may read it (see also the "Security Considerations"
354 chapter), as it might contain the password in plain text. libcurl has the
355 ability to use this file to figure out what set of user name and password to
356 use for a particular host. As an extension to the normal functionality,
357 libcurl also supports this file for non-FTP protocols such as HTTP. To make
358 curl use this file, use the CURLOPT_NETRC option:
360 curl_easy_setopt(easyhandle, CURLOPT_NETRC, TRUE);
362 And a very basic example of how such a .netrc file may look like:
364 machine myhost.mydomain.com
368 All these examples have been cases where the password has been optional, or
369 at least you could leave it out and have libcurl attempt to do its job
370 without it. There are times when the password isn't optional, like when
371 you're using an SSL private key for secure transfers.
373 To pass the known private key password to libcurl:
375 curl_easy_setopt(easyhandle, CURLOPT_SSLKEYPASSWD, "keypassword");
380 The previous chapter showed how to set user name and password for getting
381 URLs that require authentication. When using the HTTP protocol, there are
382 many different ways a client can provide those credentials to the server and
383 you can control what way libcurl will (attempt to) use. The default HTTP
384 authentication method is called 'Basic', which is sending the name and
385 password in clear-text in the HTTP request, base64-encoded. This is unsecure.
387 At the time of this writing libcurl can be built to use: Basic, Digest, NTLM,
388 Negotiate, GSS-Negotiate and SPNEGO. You can tell libcurl which one to use
389 with CURLOPT_HTTPAUTH as in:
391 curl_easy_setopt(easyhandle, CURLOPT_HTTPAUTH, CURLAUTH_DIGEST);
393 And when you send authentication to a proxy, you can also set authentication
394 type the same way but instead with CURLOPT_PROXYAUTH:
396 curl_easy_setopt(easyhandle, CURLOPT_PROXYAUTH, CURLAUTH_NTLM);
398 Both these options allow you to set multiple types (by ORing them together),
399 to make libcurl pick the most secure one out of the types the server/proxy
400 claims to support. This method does however add a round-trip since libcurl
401 must first ask the server what it supports:
403 curl_easy_setopt(easyhandle, CURLOPT_HTTPAUTH,
404 CURLAUTH_DIGEST|CURLAUTH_BASIC);
406 For convenience, you can use the 'CURLAUTH_ANY' define (instead of a list
407 with specific types) which allows libcurl to use whatever method it wants.
409 When asking for multiple types, libcurl will pick the available one it
410 considers "best" in its own internal order of preference.
415 We get many questions regarding how to issue HTTP POSTs with libcurl the
416 proper way. This chapter will thus include examples using both different
417 versions of HTTP POST that libcurl supports.
419 The first version is the simple POST, the most common version, that most HTML
420 pages using the <form> tag uses. We provide a pointer to the data and tell
421 libcurl to post it all to the remote site:
423 char *data="name=daniel&project=curl";
424 curl_easy_setopt(easyhandle, CURLOPT_POSTFIELDS, data);
425 curl_easy_setopt(easyhandle, CURLOPT_URL, "http://posthere.com/");
427 curl_easy_perform(easyhandle); /* post away! */
429 Simple enough, huh? Since you set the POST options with the
430 CURLOPT_POSTFIELDS, this automaticly switches the handle to use POST in the
433 Ok, so what if you want to post binary data that also requires you to set the
434 Content-Type: header of the post? Well, binary posts prevents libcurl from
435 being able to do strlen() on the data to figure out the size, so therefore we
436 must tell libcurl the size of the post data. Setting headers in libcurl
437 requests are done in a generic way, by building a list of our own headers and
438 then passing that list to libcurl.
440 struct curl_slist *headers=NULL;
441 headers = curl_slist_append(headers, "Content-Type: text/xml");
443 /* post binary data */
444 curl_easy_setopt(easyhandle, CURLOPT_POSTFIELDS, binaryptr);
446 /* set the size of the postfields data */
447 curl_easy_setopt(easyhandle, CURLOPT_POSTFIELDSIZE, 23);
449 /* pass our list of custom made headers */
450 curl_easy_setopt(easyhandle, CURLOPT_HTTPHEADER, headers);
452 curl_easy_perform(easyhandle); /* post away! */
454 curl_slist_free_all(headers); /* free the header list */
456 While the simple examples above cover the majority of all cases where HTTP
457 POST operations are required, they don't do multipart formposts. Multipart
458 formposts were introduced as a better way to post (possibly large) binary
459 data and was first documented in the RFC1867. They're called multipart
460 because they're built by a chain of parts, each being a single unit. Each
461 part has its own name and contents. You can in fact create and post a
462 multipart formpost with the regular libcurl POST support described above, but
463 that would require that you build a formpost yourself and provide to
464 libcurl. To make that easier, libcurl provides curl_formadd(). Using this
465 function, you add parts to the form. When you're done adding parts, you post
468 The following example sets two simple text parts with plain textual contents,
469 and then a file with binary contents and upload the whole thing.
471 struct curl_httppost *post=NULL;
472 struct curl_httppost *last=NULL;
473 curl_formadd(&post, &last,
474 CURLFORM_COPYNAME, "name",
475 CURLFORM_COPYCONTENTS, "daniel", CURLFORM_END);
476 curl_formadd(&post, &last,
477 CURLFORM_COPYNAME, "project",
478 CURLFORM_COPYCONTENTS, "curl", CURLFORM_END);
479 curl_formadd(&post, &last,
480 CURLFORM_COPYNAME, "logotype-image",
481 CURLFORM_FILECONTENT, "curl.png", CURLFORM_END);
483 /* Set the form info */
484 curl_easy_setopt(easyhandle, CURLOPT_HTTPPOST, post);
486 curl_easy_perform(easyhandle); /* post away! */
488 /* free the post data again */
491 Multipart formposts are chains of parts using MIME-style separators and
492 headers. It means that each one of these separate parts get a few headers set
493 that describe the individual content-type, size etc. To enable your
494 application to handicraft this formpost even more, libcurl allows you to
495 supply your own set of custom headers to such an individual form part. You
496 can of course supply headers to as many parts you like, but this little
497 example will show how you set headers to one specific part when you add that
500 struct curl_slist *headers=NULL;
501 headers = curl_slist_append(headers, "Content-Type: text/xml");
503 curl_formadd(&post, &last,
504 CURLFORM_COPYNAME, "logotype-image",
505 CURLFORM_FILECONTENT, "curl.xml",
506 CURLFORM_CONTENTHEADER, headers,
509 curl_easy_perform(easyhandle); /* post away! */
511 curl_formfree(post); /* free post */
512 curl_slist_free_all(post); /* free custom header list */
514 Since all options on an easyhandle are "sticky", they remain the same until
515 changed even if you do call curl_easy_perform(), you may need to tell curl to
516 go back to a plain GET request if you intend to do such a one as your next
517 request. You force an easyhandle to back to GET by using the CURLOPT_HTTPGET
520 curl_easy_setopt(easyhandle, CURLOPT_HTTPGET, TRUE);
522 Just setting CURLOPT_POSTFIELDS to "" or NULL will *not* stop libcurl from
523 doing a POST. It will just make it POST without any data to send!
528 For historical and traditional reasons, libcurl has a built-in progress meter
529 that can be switched on and then makes it presents a progress meter in your
532 Switch on the progress meter by, oddly enough, set CURLOPT_NOPROGRESS to
533 FALSE. This option is set to TRUE by default.
535 For most applications however, the built-in progress meter is useless and
536 what instead is interesting is the ability to specify a progress
537 callback. The function pointer you pass to libcurl will then be called on
538 irregular intervals with information about the current transfer.
540 Set the progress callback by using CURLOPT_PROGRESSFUNCTION. And pass a
541 pointer to a function that matches this prototype:
543 int progress_callback(void *clientp,
549 If any of the input arguments is unknown, a 0 will be passed. The first
550 argument, the 'clientp' is the pointer you pass to libcurl with
551 CURLOPT_PROGRESSDATA. libcurl won't touch it.
556 There's basicly only one thing to keep in mind when using C++ instead of C
557 when interfacing libcurl:
559 "The Callbacks Must Be Plain C"
561 So if you want a write callback set in libcurl, you should put it within
562 'extern'. Similar to this:
565 size_t write_data(void *ptr, size_t size, size_t nmemb,
568 /* do what you want with the data */
572 This will of course effectively turn the callback code into C. There won't be
573 any "this" pointer available etc.
578 What "proxy" means according to Merriam-Webster: "a person authorized to act
579 for another" but also "the agency, function, or office of a deputy who acts
580 as a substitute for another".
582 Proxies are exceedingly common these days. Companies often only offer
583 internet access to employees through their HTTP proxies. Network clients or
584 user-agents ask the proxy for docuements, the proxy does the actual request
585 and then it returns them.
587 libcurl has full support for HTTP proxies, so when a given URL is wanted,
588 libcurl will ask the proxy for it instead of trying to connect to the actual
589 host identified in the URL.
591 The fact that the proxy is a HTTP proxy puts certain restrictions on what can
592 actually happen. A requested URL that might not be a HTTP URL will be still
593 be passed to the HTTP proxy to deliver back to libcurl. This happens
594 transparantly, and an application may not need to know. I say "may", because
595 at times it is very important to understand that all operations over a HTTP
596 proxy is using the HTTP protocol. For example, you can't invoke your own
597 custom FTP commands or even proper FTP directory listings.
601 To tell libcurl to use a proxy at a given port number:
603 curl_easy_setopt(easyhandle, CURLOPT_PROXY, "proxy-host.com:8080");
605 Some proxies require user authentication before allowing a request, and
606 you pass that information similar to this:
608 curl_easy_setopt(easyhandle, CURLOPT_PROXYUSERPWD, "user:password");
610 If you want to, you can specify the host name only in the CURLOPT_PROXY
611 option, and set the port number separately with CURLOPT_PROXYPORT.
613 Environment Variables
615 libcurl automaticly checks and uses a set of environment variables to know
616 what proxies to use for certain protocols. The names of the variables are
617 following an ancient de facto standard and are built up as
618 "[protocol]_proxy" (note the lower casing). Which makes the variable
619 'http_proxy' checked for a name of a proxy to use when the input URL is
620 HTTP. Following the same rule, the variable named 'ftp_proxy' is checked
621 for FTP URLs. Again, the proxies are always HTTP proxies, the different
622 names of the variables simply allows different HTTP proxies to be used.
624 The proxy environment variable contents should be in the format
625 "[protocol://]machine[:port]". Where the protocol:// part is simply
626 ignored if present (so http://proxy and bluerk://proxy will do the same)
627 and the optional port number specifies on which port the proxy operates on
628 the host. If not specified, the internal default port number will be used
629 and that is most likely *not* the one you would like it to be.
631 There are two special environment variables. 'all_proxy' is what sets
632 proxy for any URL in case the protocol specific variable wasn't set, and
633 'no_proxy' defines a list of hosts that should not use a proxy even though
634 a variable may say so. If 'no_proxy' is a plain asterisk ("*") it matches
639 SSL is for secure point-to-point connections. This involves strong
640 encryption and similar things, which effectivly makes it impossible for a
641 proxy to operate as a "man in between" which the proxy's task is, as
642 previously discussed. Instead, the only way to have SSL work over a HTTP
643 proxy is to ask the proxy to tunnel trough everything without being able
644 to check or fiddle with the traffic.
646 Opening an SSL connection over a HTTP proxy is therefor a matter of asking
647 the proxy for a straight connection to the target host on a specified
648 port. This is made with the HTTP request CONNECT. ("please mr proxy,
649 connect me to that remote host").
651 Because of the nature of this operation, where the proxy has no idea what
652 kind of data that is passed in and out through this tunnel, this breaks
653 some of the very few advantages that come from using a proxy, such as
654 caching. Many organizations prevent this kind of tunneling to other
655 destination port numbers than 443 (which is the default HTTPS port
658 Tunneling Through Proxy
660 As explained above, tunneling is required for SSL to work and often even
661 restricted to the operation intended for SSL; HTTPS.
663 This is however not the only time proxy-tunneling might offer benefits to
664 you or your application.
666 As tunneling opens a direct connection from your application to the remote
667 machine, it suddenly also re-introduces the ability to do non-HTTP
668 operations over a HTTP proxy. You can in fact use things such as FTP
669 upload or FTP custom commands this way.
671 Again, this is often prevented by the adminstrators of proxies and is
674 Tell libcurl to use proxy tunneling like this:
676 curl_easy_setopt(easyhandle, CURLOPT_HTTPPROXYTUNNEL, TRUE);
678 In fact, there might even be times when you want to do plain HTTP
679 operations using a tunnel like this, as it then enables you to operate on
680 the remote server instead of asking the proxy to do so. libcurl will not
681 stand in the way for such innovative actions either!
685 Netscape first came up with this. It is basicly a web page (usually using
686 a .pac extension) with a javascript that when executed by the browser with
687 the requested URL as input, returns information to the browser on how to
688 connect to the URL. The returned information might be "DIRECT" (which
689 means no proxy should be used), "PROXY host:port" (to tell the browser
690 where the proxy for this particular URL is) or "SOCKS host:port" (to
691 direct the brower to a SOCKS proxy).
693 libcurl has no means to interpret or evaluate javascript and thus it
694 doesn't support this. If you get yourself in a position where you face
695 this nasty invention, the following advice have been mentioned and used in
698 - Depending on the javascript complexity, write up a script that
699 translates it to another language and execute that.
701 - Read the javascript code and rewrite the same logic in another language.
703 - Implement a javascript interpreted, people have successfully used the
704 Mozilla javascript engine in the past.
706 - Ask your admins to stop this, for a static proxy setup or similar.
709 Persistancy Is The Way to Happiness
711 Re-cycling the same easy handle several times when doing multiple requests is
714 After each single curl_easy_perform() operation, libcurl will keep the
715 connection alive and open. A subsequent request using the same easy handle to
716 the same host might just be able to use the already open connection! This
717 reduces network impact a lot.
719 Even if the connection is dropped, all connections involving SSL to the same
720 host again, will benefit from libcurl's session ID cache that drasticly
721 reduces re-connection time.
723 FTP connections that are kept alive saves a lot of time, as the command-
724 response roundtrips are skipped, and also you don't risk getting blocked
725 without permission to login again like on many FTP servers only allowing N
726 persons to be logged in at the same time.
728 libcurl caches DNS name resolving results, to make lookups of a previously
729 looked up name a lot faster.
731 Other interesting details that improve performance for subsequent requests
732 may also be added in the future.
734 Each easy handle will attempt to keep the last few connections alive for a
735 while in case they are to be used again. You can set the size of this "cache"
736 with the CURLOPT_MAXCONNECTS option. Default is 5. It is very seldom any
737 point in changing this value, and if you think of changing this it is often
738 just a matter of thinking again.
740 When the connection cache gets filled, libcurl must close an existing
741 connection in order to get room for the new one. To know which connection to
742 close, libcurl uses a "close policy" that you can affect with the
743 CURLOPT_CLOSEPOLICY option. There's only two polices implemented as of this
744 writing (libcurl 7.9.4) and they are:
746 CURLCLOSEPOLICY_LEAST_RECENTLY_USED simply close the one that hasn't been
747 used for the longest time. This is the default behavior.
749 CURLCLOSEPOLICY_OLDEST closes the oldest connection, the one that was
750 createst the longest time ago.
752 There are, or at least were, plans to support a close policy that would call
753 a user-specified callback to let the user be able to decide which connection
754 to dump when this is necessary and therefor is the CURLOPT_CLOSEFUNCTION an
755 existing option still today. Nothing ever uses this though and this will not
756 be used within the forseeable future either.
758 To force your upcoming request to not use an already existing connection (it
759 will even close one first if there happens to be one alive to the same host
760 you're about to operate on), you can do that by setting CURLOPT_FRESH_CONNECT
761 to TRUE. In a similar spirit, you can also forbid the upcoming request to be
762 "lying" around and possibly get re-used after the request by setting
763 CURLOPT_FORBID_REUSE to TRUE.
766 HTTP Headers Used by libcurl
768 When you use libcurl to do HTTP requeests, it'll pass along a series of
769 headers automaticly. It might be good for you to know and understand these
774 This header is required by HTTP 1.1 and even many 1.0 servers and should
775 be the name of the server we want to talk to. This includes the port
776 number if anything but default.
780 "no-cache". Tells a possible proxy to not grap a copy from the cache but
781 to fetch a fresh one.
785 "image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, */*". Cloned from a
786 browser once a hundred years ago.
790 When doing multi-part formposts, libcurl will set this header to
791 "100-continue" to ask the server for an "OK" message before it proceeds
792 with sending the data part of the post.
795 Customizing Operations
797 There is an ongoing development today where more and more protocols are built
798 upon HTTP for transport. This has obvious benefits as HTTP is a tested and
799 reliable protocol that is widely deployed and have excellent proxy-support.
801 When you use one of these protocols, and even when doing other kinds of
802 programming you may need to change the traditional HTTP (or FTP or...)
803 manners. You may need to change words, headers or various data.
805 libcurl is your friend here too.
809 If just changing the actual HTTP request keyword is what you want, like
810 when GET, HEAD or POST is not good enough for you, CURLOPT_CUSTOMREQUEST
811 is there for you. It is very simple to use:
813 curl_easy_setopt(easyhandle, CURLOPT_CUSTOMREQUEST, "MYOWNRUQUEST");
815 When using the custom request, you change the request keyword of the
816 actual request you are performing. Thus, by default you make GET request
817 but you can also make a POST operation (as described before) and then
818 replace the POST keyword if you want to. You're the boss.
822 HTTP-like protocols pass a series of headers to the server when doing the
823 request, and you're free to pass any amount of extra headers that you
824 think fit. Adding headers are this easy:
826 struct curl_slist *headers=NULL; /* init to NULL is important */
828 headers = curl_slist_append(headers, "Hey-server-hey: how are you?");
829 headers = curl_slist_append(headers, "X-silly-content: yes");
831 /* pass our list of custom made headers */
832 curl_easy_setopt(easyhandle, CURLOPT_HTTPHEADER, headers);
834 curl_easy_perform(easyhandle); /* transfer http */
836 curl_slist_free_all(headers); /* free the header list */
838 ... and if you think some of the internally generated headers, such as
839 Accept: or Host: don't contain the data you want them to contain, you can
840 replace them by simply setting them too:
842 headers = curl_slist_append(headers, "Accept: Agent-007");
843 headers = curl_slist_append(headers, "Host: munged.host.line");
847 If you replace an existing header with one with no contents, you will
848 prevent the header from being sent. Like if you want to completely prevent
849 the "Accept:" header to be sent, you can disable it with code similar to
852 headers = curl_slist_append(headers, "Accept:");
854 Both replacing and cancelling internal headers should be done with careful
855 consideration and you should be aware that you may violate the HTTP
856 protocol when doing so.
858 Enforcing chunked transfer-encoding
860 By making sure a request uses the custom header "Transfer-Encoding:
861 chunked" when doing a non-GET HTTP operation, libcurl will switch over to
862 "chunked" upload, even though the size of the data to upload might be
863 known. By default, libcurl usually switches over to chunked upload
864 automaticly if the upload data size is unknown.
868 There's only one aspect left in the HTTP requests that we haven't yet
869 mentioned how to modify: the version field. All HTTP requests includes the
870 version number to tell the server which version we support. libcurl speak
871 HTTP 1.1 by default. Some very old servers don't like getting 1.1-requests
872 and when dealing with stubborn old things like that, you can tell libcurl
873 to use 1.0 instead by doing something like this:
875 curl_easy_setopt(easyhandle, CURLOPT_HTTP_VERSION,
876 CURLHTTP_VERSION_1_0);
880 Not all protocols are HTTP-like, and thus the above may not help you when
881 you want to make for example your FTP transfers to behave differently.
883 Sending custom commands to a FTP server means that you need to send the
884 comands exactly as the FTP server expects them (RFC959 is a good guide
885 here), and you can only use commands that work on the control-connection
886 alone. All kinds of commands that requires data interchange and thus needs
887 a data-connection must be left to libcurl's own judgement. Also be aware
888 that libcurl will do its very best to change directory to the target
889 directory before doing any transfer, so if you change directory (with CWD
890 or similar) you might confuse libcurl and then it might not attempt to
891 transfer the file in the correct remote directory.
893 A little example that deletes a given file before an operation:
895 headers = curl_slist_append(headers, "DELE file-to-remove");
897 /* pass the list of custom commands to the handle */
898 curl_easy_setopt(easyhandle, CURLOPT_QUOTE, headers);
900 curl_easy_perform(easyhandle); /* transfer ftp data! */
902 curl_slist_free_all(headers); /* free the header list */
904 If you would instead want this operation (or chain of operations) to
905 happen _after_ the data transfer took place the option to
906 curl_easy_setopt() would instead be called CURLOPT_POSTQUOTE and used the
909 The custom FTP command will be issued to the server in the same order they
910 are added to the list, and if a command gets an error code returned back
911 from the server, no more commands will be issued and libcurl will bail out
912 with an error code (CURLE_FTP_QUOTE_ERROR). Note that if you use
913 CURLOPT_QUOTE to send commands before a transfer, no transfer will
914 actually take place when a quote command has failed.
916 If you set the CURLOPT_HEADER to true, you will tell libcurl to get
917 information about the target file and output "headers" about it. The
918 headers will be in "HTTP-style", looking like they do in HTTP.
920 The option to enable headers or to run custom FTP commands may be useful
921 to combine with CURLOPT_NOBODY. If this option is set, no actual file
922 content transfer will be performed.
924 FTP Custom CUSTOMREQUEST
926 If you do what list the contents of a FTP directory using your own defined
927 FTP command, CURLOPT_CUSTOMREQUEST will do just that. "NLST" is the
928 default one for listing directories but you're free to pass in your idea
929 of a good alternative.
932 Cookies Without Chocolate Chips
934 In the HTTP sense, a cookie is a name with an associated value. A server
935 sends the name and value to the client, and expects it to get sent back on
936 every subsequent request to the server that matches the particular conditions
937 set. The conditions include that the domain name and path match and that the
938 cookie hasn't become too old.
940 In real-world cases, servers send new cookies to replace existing one to
941 update them. Server use cookies to "track" users and to keep "sessions".
943 Cookies are sent from server to clients with the header Set-Cookie: and
944 they're sent from clients to servers with the Cookie: header.
946 To just send whatever cookie you want to a server, you can use CURLOPT_COOKIE
947 to set a cookie string like this:
949 curl_easy_setopt(easyhandle, CURLOPT_COOKIE, "name1=var1; name2=var2;");
951 In many cases, that is not enough. You might want to dynamicly save whatever
952 cookies the remote server passes to you, and make sure those cookies are then
953 use accordingly on later requests.
955 One way to do this, is to save all headers you receive in a plain file and
956 when you make a request, you tell libcurl to read the previous headers to
957 figure out which cookies to use. Set header file to read cookies from with
960 The CURLOPT_COOKIEFILE option also automaticly enables the cookie parser in
961 libcurl. Until the cookie parser is enabled, libcurl will not parse or
962 understand incoming cookies and they will just be ignored. However, when the
963 parser is enabled the cookies will be understood and the cookies will be kept
964 in memory and used properly in subsequent requests when the same handle is
965 used. Many times this is enough, and you may not have to save the cookies to
966 disk at all. Note that the file you specify to CURLOPT_COOKIEFILE doesn't
967 have to exist to enable the parser, so a common way to just enable the parser
968 and not read able might be to use a file name you know doesn't exist.
970 If you rather use existing cookies that you've previously received with your
971 Netscape or Mozilla browsers, you can make libcurl use that cookie file as
972 input. The CURLOPT_COOKIEFILE is used for that too, as libcurl will
973 automaticly find out what kind of file it is and act accordingly.
975 The perhaps most advanced cookie operation libcurl offers, is saving the
976 entire internal cookie state back into a Netscape/Mozilla formatted cookie
977 file. We call that the cookie-jar. When you set a file name with
978 CURLOPT_COOKIEJAR, that file name will be created and all received cookies
979 will be stored in it when curl_easy_cleanup() is called. This enabled cookies
980 to get passed on properly between multiple handles without any information
984 FTP Peculiarities We Need
986 FTP transfers use a second TCP/IP connection for the data transfer. This is
987 usually a fact you can forget and ignore but at times this fact will come
988 back to haunt you. libcurl offers several different ways to custom how the
989 second connection is being made.
991 libcurl can either connect to the server a second time or tell the server to
992 connect back to it. The first option is the default and it is also what works
993 best for all the people behind firewalls, NATs or IP-masquarading setups.
994 libcurl then tells the server to open up a new port and wait for a second
995 connection. This is by default attempted with EPSV first, and if that doesn't
996 work it tries PASV instead. (EPSV is an extension to the original FTP spec
997 and does not exist nor work on all FTP servers.)
999 You can prevent libcurl from first trying the EPSV command by setting
1000 CURLOPT_FTP_USE_EPSV to FALSE.
1002 In some cases, you will prefer to have the server connect back to you for the
1003 second connection. This might be when the server is perhaps behind a firewall
1004 or something and only allows connections on a single port. libcurl then
1005 informs the remote server which IP address and port number to connect to.
1006 This is made with the CURLOPT_FTPPORT option. If you set it to "-", libcurl
1007 will use your system's "default IP address". If you want to use a particular
1008 IP, you can set the full IP address, a host name to resolve to an IP address
1009 or even a local network interface name that libcurl will get the IP address
1012 When doing the "PORT" approach, libcurl will attempt to use the EPRT and the
1013 LPRT before trying PORT, as they work with more protocols. You can disable
1014 this behavior by setting CURLOPT_FTP_USE_EPRT to FALSE.
1019 Some protocols provide "headers", meta-data separated from the normal
1020 data. These headers are by default not included in the normal data stream,
1021 but you can make them appear in the data stream by setting CURLOPT_HEADER to
1024 What might be even more useful, is libcurl's ability to separate the headers
1025 from the data and thus make the callbacks differ. You can for example set a
1026 different pointer to pass to the ordinary write callback by setting
1027 CURLOPT_WRITEHEADER.
1029 Or, you can set an entirely separate function to receive the headers, by
1030 using CURLOPT_HEADERFUNCTION.
1032 The headers are passed to the callback function one by one, and you can
1033 depend on that fact. It makes it easier for you to add custom header parsers
1036 "Headers" for FTP transfers equal all the FTP server responses. They aren't
1037 actually true headers, but in this case we pretend they are! ;-)
1040 Post Transfer Information
1042 [ curl_easy_getinfo ]
1045 Security Considerations
1047 libcurl is in itself not insecure. If used the right way, you can use libcurl
1048 to transfer data pretty safely.
1050 There are of course many things to consider that may loosen up this
1055 If you use a command line tool (such as curl) that uses libcurl, and you
1056 give option to the tool on the command line those options can very likely
1057 get read by other users of your system when they use 'ps' or other tools
1058 to list currently running processes.
1060 To avoid this problem, never feed sensitive things to programs using
1061 command line options.
1065 .netrc is a pretty handy file/feature that allows you to login quickly and
1066 automaticly to frequently visited sites. The file contains passwords in
1067 clear text and is a real security risk. In some cases, your .netrc is also
1068 stored in a home directory that is NFS mounted or used on another network
1069 based file system, so the clear text password will fly through your
1070 network every time anyone reads that file!
1072 To avoid this problem, don't use .netrc files and never store passwords in
1073 plain text anywhere.
1075 Clear Text Passwords
1077 Many of the protocols libcurl supports send name and password unencrypted
1078 as clear text (HTTP Basic authentication, FTP, TELNET etc). It is very
1079 easy for anyone on your network or a network nearby yours, to just fire up
1080 a network analyzer tool and evesdrop on your passwords. Don't let the fact
1081 that HTTP uses base64 encoded passwords fool you. They may not look
1082 readable at a first glance, but they very easily "deciphered" by anyone
1085 To avoid this problem, use protocols that don't let snoopers see your
1086 password: HTTPS, FTPS and FTP-kerberos are a few examples. HTTP Digest
1087 authentication allows this too, but isn't supported by libcurl as of this
1092 On a related issue, be aware that even in situations like when you have
1093 problems with libcurl and ask somone for help, everything you reveal in
1094 order to get best possible help might also impose certain security related
1095 risks. Host names, user names, paths, operating system specifics etc (not
1096 to mention passwords of course) may in fact be used by intruders to gain
1097 additional information of a potential target.
1099 To avoid this problem, you must of course use your common sense. Often,
1100 you can just edit out the senstive data or just rearch/replace your true
1101 information with faked data.
1104 SSL, Certificates and Other Tricks
1106 [ seeding, passwords, keys, certificates, ENGINE, ca certs ]
1108 Multiple Transfers Using the multi Interface
1110 The easy interface as described in detail in this document is a synchronous
1111 interface that transfers one file at a time and doesn't return until its
1114 The multi interface on the other hand, allows your program to transfer
1115 multiple files in both directions at the same time, without forcing you to
1116 use multiple threads.
1118 [fill in lots of more multi stuff here]
1122 [ sharing between handles, mutexes, pipelining ]
1128 [1] = libcurl 7.10.3 and later have the ability to switch over to chunked
1129 Tranfer-Encoding in cases were HTTP uploads are done with data of an
1132 [2] = This happens on Windows machines when libcurl is built and used as a
1133 DLL. However, you can still do this on Windows if you link with a static
1136 [3] = The curl-config tool is generated at build-time (on unix-like systems)
1137 and should be installed with the 'make install' or similar instruction
1138 that installs the library, header files, man pages etc.