5 \___|\___/|_| \_\_____|
7 Things that could be nice to do in the future
9 Things to do in project cURL. Please tell us what you think, contribute and
10 send us patches that improve things!
12 Be aware that these are things that we could do, or have once been considered
13 things we could do. If you want to work on any of these areas, please
14 consider bringing it up for discussions first on the mailing list so that we
15 all agree it is still a good idea for the project!
17 All bugs documented in the KNOWN_BUGS document are subject for fixing!
22 1.4 signal-based resolver timeouts
23 1.5 get rid of PATH_MAX
24 1.6 Modified buffer size approach
25 1.7 Detect when called from within callbacks
26 1.8 Allow SSL (HTTPS) to proxy
27 1.9 Cache negative name resolves
29 1.11 minimize dependencies with dynamicly loaded modules
30 1.12 have form functions use CURL handle argument
31 1.13 Add CURLOPT_MAIL_CLIENT option
32 1.14 Typesafe curl_easy_setopt()
34 1.16 Try to URL encode given URL
36 2. libcurl - multi interface
38 2.2 Better support for same name resolves
39 2.3 Non-blocking curl_multi_remove_handle()
40 2.4 Split connect and authentication process
43 3.1 Update date and version in man pages
47 4.2 Alter passive/active on failure and retry
48 4.3 Earlier bad letter detection
49 4.4 REST for large files
51 4.6 GSSAPI via Windows SSPI
52 4.7 STAT for LIST without data connection
55 5.1 Better persistency for HTTP 1.0
56 5.2 support FF3 sqlite cookie files
57 5.3 Rearrange request header order
60 5.6 Refuse "downgrade" redirects
65 6.2 ditch telnet-specific select
66 6.3 feature negotiation debug data
67 6.4 send data in chunks
71 7.2 Enhanced capability support
75 8.2 Enhanced capability support
78 9.1 Enhanced capability support
81 10.1 SASL based authentication mechanisms
84 11.1 File listing support
85 11.2 Honor file timestamps
87 11.4 Create remote directories
93 13.1 Disable specific versions
94 13.2 Provide mutex locking API
95 13.3 Evaluate SSL patches
96 13.4 Cache OpenSSL contexts
97 13.5 Export session ids
98 13.6 Provide callback for cert verification
99 13.7 improve configure --with-ssl
103 14.1 SSL engine stuff
104 14.2 check connection
107 15.1 Add support for client certificate authentication
108 15.2 Add support for custom server certificate validation
109 15.3 Add support for the --ciphers option
112 16.1 Other authentication mechanisms
113 16.2 Add QOP support to GSSAPI authentication
115 17. Command line tool
118 17.3 prevent file overwriting
119 17.4 simultaneous parallel transfers
120 17.5 provide formpost headers
121 17.6 warning when setting an option
122 17.7 warning when sending binary output to terminal
123 17.8 offer color-coded HTTP header output
124 17.9 Choose the name of file in braces for complex URLs
125 17.10 improve how curl works in a windows console window
126 17.11 -w output to stderr
127 17.12 keep running, read instructions from pipe/socket
134 19.2 nicer lacking perl message
135 19.3 more protocols supported
136 19.4 more platforms supported
137 19.5 Add support for concurrent connections
138 19.6 Use the RFC6265 test suite
141 20.1 http-style HEAD output for FTP
142 20.2 combine error codes
143 20.3 extend CURLOPT_SOCKOPTFUNCTION prototype
145 21. Next major release
146 21.1 cleanup return codes
147 21.2 remove obsolete defines
149 21.4 remove several functions
150 21.5 remove CURLOPT_FAILONERROR
151 21.6 remove CURLOPT_DNS_USE_GLOBAL_CACHE
152 21.7 remove progress meter from libcurl
153 21.8 remove 'curl_httppost' from public
155 ==============================================================================
159 1.2 More data sharing
161 curl_share_* functions already exist and work, and they can be extended to
162 share more. For example, enable sharing of the ares channel and the
167 Use 'struct lifreq' and SIOCGLIFADDR instead of 'struct ifreq' and
168 SIOCGIFADDR on newer Solaris versions as they claim the latter is obsolete.
169 To support IPv6 interface addresses for network interfaces properly.
171 1.4 signal-based resolver timeouts
173 libcurl built without an asynchronous resolver library uses alarm() to time
174 out DNS lookups. When a timeout occurs, this causes libcurl to jump from the
175 signal handler back into the library with a sigsetjmp, which effectively
176 causes libcurl to continue running within the signal handler. This is
177 non-portable and could cause problems on some platforms. A discussion on the
178 problem is available at https://curl.haxx.se/mail/lib-2008-09/0197.html
180 Also, alarm() provides timeout resolution only to the nearest second. alarm
181 ought to be replaced by setitimer on systems that support it.
183 1.5 get rid of PATH_MAX
185 Having code use and rely on PATH_MAX is not nice:
186 http://insanecoding.blogspot.com/2007/11/pathmax-simply-isnt.html
188 Currently the SSH based code uses it a bit, but to remove PATH_MAX from there
189 we need libssh2 to properly tell us when we pass in a too small buffer and
190 its current API (as of libssh2 1.2.7) doesn't.
192 1.6 Modified buffer size approach
194 Current libcurl allocates a fixed 16K size buffer for download and an
195 additional 16K for upload. They are always unconditionally part of the easy
196 handle. If CRLF translations are requested, an additional 32K "scratch
197 buffer" is allocated. A total of 64K transfer buffers in the worst case.
199 First, while the handles are not actually in use these buffers could be freed
200 so that lingering handles just kept in queues or whatever waste less memory.
202 Secondly, SFTP is a protocol that needs to handle many ~30K blocks at once
203 since each need to be individually acked and therefore libssh2 must be
204 allowed to send (or receive) many separate ones in parallel to achieve high
205 transfer speeds. A current libcurl build with a 16K buffer makes that
206 impossible, but one with a 512K buffer will reach MUCH faster transfers. But
207 allocating 512K unconditionally for all buffers just in case they would like
208 to do fast SFTP transfers at some point is not a good solution either.
210 Dynamically allocate buffer size depending on protocol in use in combination
211 with freeing it after each individual transfer? Other suggestions?
213 1.7 Detect when called from within callbacks
215 We should set a state variable before calling callbacks, so that we
216 subsequently can add code within libcurl that returns error if called within
217 callbacks for when that's not supported.
219 1.8 Allow SSL (HTTPS) to proxy
221 To prevent local users from snooping on your traffic to the proxy. Supported
223 https://www.chromium.org/developers/design-documents/secure-web-proxy
225 ...and by Firefox soon:
226 https://bugzilla.mozilla.org/show_bug.cgi?id=378637
228 1.9 Cache negative name resolves
230 A name resolve that has failed is likely to fail when made again within a
231 short period of time. Currently we only cache positive responses.
233 1.10 Support IDNA2008
235 International Domain Names are supported in libcurl since years back, powered
236 by libidn. libidn implements IDNA2003 which has been superseded by IDNA2008.
237 libidn2 is an existing library offering support for IDNA2008.
239 1.11 minimize dependencies with dynamicly loaded modules
241 We can create a system with loadable modules/plug-ins, where these modules
242 would be the ones that link to 3rd party libs. That would allow us to avoid
243 having to load ALL dependencies since only the necessary ones for this
244 app/invoke/used protocols would be necessary to load. See
245 https://github.com/curl/curl/issues/349
247 1.12 have form functions use CURL handle argument
249 curl_formadd() and curl_formget() both currently have no CURL handle
250 argument, but both can use a callback that is set in the easy handle, and
251 thus curl_formget() with callback cannot function without first having
252 curl_easy_perform() (or similar) called - which is hard to grasp and a design
255 The curl_formadd() design can probably also be reconsidered to make it easier
256 to use and less error-prone. Probably easiest by splitting it into several
259 1.13 Add CURLOPT_MAIL_CLIENT option
261 Rather than use the URL to specify the mail client string to present in the
262 HELO and EHLO commands, libcurl should support a new CURLOPT specifically for
263 specifying this data as the URL is non-standard and to be honest a bit of a
266 Please see the following thread for more information:
267 https://curl.haxx.se/mail/lib-2012-05/0178.html
269 1.14 Typesafe curl_easy_setopt()
271 One of the most common problems in libcurl using applications is the lack of
272 type checks for curl_easy_setopt() which happens because it accepts varargs
273 and thus can take any type.
275 One possible solution to this is to introduce a few different versions of the
276 setopt version for the different kinds of data you can set.
278 curl_easy_set_num() - sets a long value
280 curl_easy_set_large() - sets a curl_off_t value
282 curl_easy_set_ptr() - sets a pointer
284 curl_easy_set_cb() - sets a callback PLUS its callback data
288 RFC 7413 defines how to include data already in the TCP SYN handshake to
291 1.16 Try to URL encode given URL
293 Given a URL that for example contains spaces, libcurl could have an option
294 that would try somewhat harder than it does now and convert spaces to %20 and
295 perhaps URL encoded byte values over 128 etc (basically do what the redirect
296 following code already does).
298 https://github.com/curl/curl/issues/514
300 2. libcurl - multi interface
302 2.1 More non-blocking
304 Make sure we don't ever loop because of non-blocking sockets returning
305 EWOULDBLOCK or similar. Blocking cases include:
307 - Name resolves on non-windows unless c-ares is used
308 - NSS SSL connections
309 - HTTP proxy CONNECT operations
310 - SOCKS proxy handshakes
313 - The "DONE" operation (post transfer protocol-specific actions) for the
314 protocols SFTP, SMTP, FTP. Fixing Curl_done() for this is a worthy task.
316 2.2 Better support for same name resolves
318 If a name resolve has been initiated for name NN and a second easy handle
319 wants to resolve that name as well, make it wait for the first resolve to end
320 up in the cache instead of doing a second separate resolve. This is
321 especially needed when adding many simultaneous handles using the same host
322 name when the DNS resolver can get flooded.
324 2.3 Non-blocking curl_multi_remove_handle()
326 The multi interface has a few API calls that assume a blocking behavior, like
327 add_handle() and remove_handle() which limits what we can do internally. The
328 multi API need to be moved even more into a single function that "drives"
329 everything in a non-blocking manner and signals when something is done. A
330 remove or add would then only ask for the action to get started and then
331 multi_perform() etc still be called until the add/remove is completed.
333 2.4 Split connect and authentication process
335 The multi interface treats the authentication process as part of the connect
336 phase. As such any failures during authentication won't trigger the relevant
337 QUIT or LOGOFF for protocols such as IMAP, POP3 and SMTP.
341 3.1 Update date and version in man pages
343 'maketgz' or another suitable script could update the .TH sections of the man
344 pages at release time to use the current date and curl/libcurl version
351 HOST is a command for a client to tell which host name to use, to offer FTP
352 servers named-based virtual hosting:
354 https://tools.ietf.org/html/rfc7151
356 4.2 Alter passive/active on failure and retry
358 When trying to connect passively to a server which only supports active
359 connections, libcurl returns CURLE_FTP_WEIRD_PASV_REPLY and closes the
360 connection. There could be a way to fallback to an active connection (and
361 vice versa). https://curl.haxx.se/bug/feature.cgi?id=1754793
363 4.3 Earlier bad letter detection
365 Make the detection of (bad) %0d and %0a codes in FTP URL parts earlier in the
366 process to avoid doing a resolve and connect in vain.
368 4.4 REST for large files
370 REST fix for servers not behaving well on >2GB requests. This should fail if
371 the server doesn't set the pointer to the requested index. The tricky
372 (impossible?) part is to figure out if the server did the right thing or not.
376 FTP ASCII transfers do not follow RFC959. They don't convert the data
379 4.6 GSSAPI via Windows SSPI
381 In addition to currently supporting the SASL GSSAPI mechanism (Kerberos V5)
382 via third-party GSS-API libraries, such as Heimdal or MIT Kerberos, also add
383 support for GSSAPI authentication via Windows SSPI.
385 4.7 STAT for LIST without data connection
387 Some FTP servers allow STAT for listing directories instead of using LIST, and
388 the response is then sent over the control connection instead of as the
389 otherwise usedw data connection: http://www.nsftools.com/tips/RawFTP.htm#STAT
391 This is not detailed in any FTP specification.
395 5.1 Better persistency for HTTP 1.0
397 "Better" support for persistent connections over HTTP 1.0
398 https://curl.haxx.se/bug/feature.cgi?id=1089001
400 5.2 support FF3 sqlite cookie files
402 Firefox 3 is changing from its former format to a a sqlite database instead.
403 We should consider how (lib)curl can/should support this.
404 https://curl.haxx.se/bug/feature.cgi?id=1871388
406 5.3 Rearrange request header order
408 Server implementors often make an effort to detect browser and to reject
409 clients it can detect to not match. One of the last details we cannot yet
410 control in libcurl's HTTP requests, which also can be exploited to detect
411 that libcurl is in fact used even when it tries to impersonate a browser, is
412 the order of the request headers. I propose that we introduce a new option in
413 which you give headers a value, and then when the HTTP request is built it
414 sorts the headers based on that number. We could then have internally created
415 headers use a default value so only headers that need to be moved have to be
420 Chrome and Firefox already support SPDY and lots of web services do. There's
421 a library for us to use for this (spdylay) that has a similar API and the
422 same author as nghttp2.
424 spdylay: https://github.com/tatsuhiro-t/spdylay
428 Add the ability to specify the preferred authentication mechanism to use by
429 using ;auth=<mech> in the login part of the URL.
433 http://test:pass;auth=NTLM@example.com would be equivalent to specifying --user
434 test:pass;auth=NTLM or --user test:pass --ntlm from the command line.
436 Additionally this should be implemented for proxy base URLs as well.
438 5.6 Refuse "downgrade" redirects
440 See https://github.com/curl/curl/issues/226
442 Consider a way to tell curl to refuse to "downgrade" protocol with a redirect
443 and/or possibly a bit that refuses redirect to change protocol completely.
445 5.7 More compressions
447 Compression algorithms that perform better than gzip are being considered for
448 use and inclusion in existing browsers. For example 'brotli'. If servers
449 follow along it is a good reason for us to also allow users to take advantage
450 of this. The algorithm: https://github.com/google/brotli The Firefox bug:
451 https://bugzilla.mozilla.org/show_bug.cgi?id=366559
458 Reading input (to send to the remote server) on stdin is a crappy solution for
459 library purposes. We need to invent a good way for the application to be able
460 to provide the data to send.
462 6.2 ditch telnet-specific select
464 Move the telnet support's network select() loop go away and merge the code
465 into the main transfer loop. Until this is done, the multi interface won't
468 6.3 feature negotiation debug data
470 Add telnet feature negotiation data to the debug callback as header data.
472 6.4 send data in chunks
474 Currently, telnet sends data one byte at a time. This is fine for interactive
475 use, but inefficient for any other. Sent data should be sent in larger
482 Add support for pipelining emails.
484 7.2 Enhanced capability support
486 Add the ability, for an application that uses libcurl, to obtain the list of
487 capabilities returned from the EHLO command.
493 Add support for pipelining commands.
495 8.2 Enhanced capability support
497 Add the ability, for an application that uses libcurl, to obtain the list of
498 capabilities returned from the CAPA command.
502 9.1 Enhanced capability support
504 Add the ability, for an application that uses libcurl, to obtain the list of
505 capabilities returned from the CAPABILITY command.
509 10.1 SASL based authentication mechanisms
511 Currently the LDAP module only supports ldap_simple_bind_s() in order to bind
512 to an LDAP server. However, this function sends username and password details
513 using the simple authentication mechanism (as clear text). However, it should
514 be possible to use ldap_bind_s() instead specifying the security context
515 information ourselves.
519 11.1 File listing support
521 Add support for listing the contents of a SMB share. The output should probably
522 be the same as/similar to FTP.
524 11.2 Honor file timestamps
526 The timestamp of the transferred file should reflect that of the original file.
530 Currently the SMB authentication uses NTLMv1.
532 11.4 Create remote directories
534 Support for creating remote directories when uploading a file to a directory
535 that doesn't exist on the server, just like --ftp-create-dirs.
541 There's no RFC for the protocol or an URI/URL format. An implementation
542 should most probably use an existing rsync library, such as librsync.
546 13.1 Disable specific versions
548 Provide an option that allows for disabling specific SSL versions, such as
549 SSLv2 https://curl.haxx.se/bug/feature.cgi?id=1767276
551 13.2 Provide mutex locking API
553 Provide a libcurl API for setting mutex callbacks in the underlying SSL
554 library, so that the same application code can use mutex-locking
555 independently of OpenSSL or GnutTLS being used.
557 13.3 Evaluate SSL patches
559 Evaluate/apply Gertjan van Wingerde's SSL patches:
560 https://curl.haxx.se/mail/lib-2004-03/0087.html
562 13.4 Cache OpenSSL contexts
564 "Look at SSL cafile - quick traces look to me like these are done on every
565 request as well, when they should only be necessary once per SSL context (or
566 once per handle)". The major improvement we can rather easily do is to make
567 sure we don't create and kill a new SSL "context" for every request, but
568 instead make one for every connection and re-use that SSL context in the same
569 style connections are re-used. It will make us use slightly more memory but
570 it will libcurl do less creations and deletions of SSL contexts.
572 13.5 Export session ids
574 Add an interface to libcurl that enables "session IDs" to get
575 exported/imported. Cris Bailiff said: "OpenSSL has functions which can
576 serialise the current SSL state to a buffer of your choice, and recover/reset
577 the state from such a buffer at a later date - this is used by mod_ssl for
578 apache to implement and SSL session ID cache".
580 13.6 Provide callback for cert verification
582 OpenSSL supports a callback for customised verification of the peer
583 certificate, but this doesn't seem to be exposed in the libcurl APIs. Could
584 it be? There's so much that could be done if it were!
586 13.7 improve configure --with-ssl
588 make the configure --with-ssl option first check for OpenSSL, then GnuTLS,
593 DNS-Based Authentication of Named Entities (DANE) is a way to provide SSL
594 keys and certs over DNS using DNSSEC as an alternative to the CA model.
595 https://www.rfc-editor.org/rfc/rfc6698.txt
597 An initial patch was posted by Suresh Krishnaswamy on March 7th 2013
598 (https://curl.haxx.se/mail/lib-2013-03/0075.html) but it was a too simple
599 approach. See Daniel's comments:
600 https://curl.haxx.se/mail/lib-2013-03/0103.html . libunbound may be the
601 correct library to base this development on.
605 14.1 SSL engine stuff
607 Is this even possible?
609 14.2 check connection
611 Add a way to check if the connection seems to be alive, to correspond to the
612 SSL_peak() way we use with OpenSSL.
616 15.1 Add support for client certificate authentication
618 WinSSL/SChannel currently makes use of the OS-level system and user
619 certificate and private key stores. This does not allow the application
620 or the user to supply a custom client certificate using curl or libcurl.
622 Therefore support for the existing -E/--cert and --key options should be
623 implemented by supplying a custom certificate to the SChannel APIs, see:
624 - Getting a Certificate for Schannel
625 https://msdn.microsoft.com/en-us/library/windows/desktop/aa375447.aspx
627 15.2 Add support for custom server certificate validation
629 WinSSL/SChannel currently makes use of the OS-level system and user
630 certificate trust store. This does not allow the application or user to
631 customize the server certificate validation process using curl or libcurl.
633 Therefore support for the existing --cacert or --capath options should be
634 implemented by supplying a custom certificate to the SChannel APIs, see:
635 - Getting a Certificate for Schannel
636 https://msdn.microsoft.com/en-us/library/windows/desktop/aa375447.aspx
638 15.3 Add support for the --ciphers option
640 The cipher suites used by WinSSL/SChannel are configured on an OS-level
641 instead of an application-level. This does not allow the application or
642 the user to customize the configured cipher suites using curl or libcurl.
644 Therefore support for the existing --ciphers option should be implemented
645 by mapping the OpenSSL/GnuTLS cipher suites to the SChannel APIs, see
646 - Specifying Schannel Ciphers and Cipher Strengths
647 https://msdn.microsoft.com/en-us/library/windows/desktop/aa380161.aspx
651 16.1 Other authentication mechanisms
653 Add support for other authentication mechanisms such as OLP,
654 GSS-SPNEGO and others.
656 16.2 Add QOP support to GSSAPI authentication
658 Currently the GSSAPI authentication only supports the default QOP of auth
659 (Authentication), whilst Kerberos V5 supports both auth-int (Authentication
660 with integrity protection) and auth-conf (Authentication with integrity and
663 17. Command line tool
667 "curl --sync http://example.com/feed[1-100].rss" or
668 "curl --sync http://example.net/{index,calendar,history}.html"
670 Downloads a range or set of URLs using the remote name, but only if the
671 remote file is newer than the local file. A Last-Modified HTTP date header
672 should also be used to set the mod date on the downloaded file.
676 Globbing support for -d and -F, as in 'curl -d "name=foo[0-9]" URL'.
677 This is easily scripted though.
679 17.3 prevent file overwriting
681 Add an option that prevents cURL from overwriting existing local files. When
682 used, and there already is an existing file with the target file name
683 (either -O or -o), a number should be appended (and increased if already
684 existing). So that index.html becomes first index.html.1 and then
687 17.4 simultaneous parallel transfers
689 The client could be told to use maximum N simultaneous parallel transfers and
690 then just make sure that happens. It should of course not make more than one
691 connection to the same remote host. This would require the client to use the
692 multi interface. https://curl.haxx.se/bug/feature.cgi?id=1558595
694 17.5 provide formpost headers
696 Extending the capabilities of the multipart formposting. How about leaving
697 the ';type=foo' syntax as it is and adding an extra tag (headers) which
698 works like this: curl -F "coolfiles=@fil1.txt;headers=@fil1.hdr" where
699 fil1.hdr contains extra headers like
701 Content-Type: text/plain; charset=KOI8-R"
702 Content-Transfer-Encoding: base64
703 X-User-Comment: Please don't use browser specific HTML code
705 which should overwrite the program reasonable defaults (plain/text,
708 17.6 warning when setting an option
710 Display a warning when libcurl returns an error when setting an option.
711 This can be useful to tell when support for a particular feature hasn't been
712 compiled into the library.
714 17.7 warning when sending binary output to terminal
716 Provide a way that prompts the user for confirmation before binary data is
717 sent to the terminal, much in the style 'less' does it.
719 17.8 offer color-coded HTTP header output
721 By offering different color output on the header name and the header
722 contents, they could be made more readable and thus help users working on
725 17.9 Choose the name of file in braces for complex URLs
727 When using braces to download a list of URLs and you use complicated names
728 in the list of alternatives, it could be handy to allow curl to use other
731 Consider a way to offer that. Possibly like
732 {partURL1:name1,partURL2:name2,partURL3:name3} where the name following the
733 colon is the output name.
735 See https://github.com/curl/curl/issues/221
737 17.10 improve how curl works in a windows console window
739 If you pull the scrollbar when transferring with curl in a Windows console
740 window, the transfer is interrupted and can get disconnected. This can
741 probably be improved. See https://github.com/curl/curl/issues/322
743 17.11 -w output to stderr
745 -w is quite useful, but not to those of us who use curl without -o or -O
746 (such as for scripting through a higher level language). It would be nice to
747 have an option that is exactly like -w but sends it to stderr
748 instead. Proposed name: --write-stderr. See
749 https://github.com/curl/curl/issues/613
751 17.12 keep running, read instructions from pipe/socket
753 Provide an option that makes curl not exit after the last URL (or even work
754 without a given URL), and then make it read instructions passed on a pipe or
755 over a socket to make further instructions so that a second subsequent curl
756 invoke can talk to the still running instance and ask for transfers to get
757 done, and thus maintain its connection pool, DNS cache and more.
763 Consider extending 'roffit' to produce decent ASCII output, and use that
764 instead of (g)nroff when building src/tool_hugehelp.c
770 Make our own version of stunnel for simple port forwarding to enable HTTPS
771 and FTP-SSL tests without the stunnel dependency, and it could allow us to
772 provide test tools built with either OpenSSL or GnuTLS
774 19.2 nicer lacking perl message
776 If perl wasn't found by the configure script, don't attempt to run the tests
777 but explain something nice why it doesn't.
779 19.3 more protocols supported
781 Extend the test suite to include more protocols. The telnet could just do FTP
782 or http operations (for which we have test servers).
784 19.4 more platforms supported
786 Make the test suite work on more platforms. OpenBSD and Mac OS. Remove
787 fork()s and it should become even more portable.
789 19.5 Add support for concurrent connections
791 Tests 836, 882 and 938 were designed to verify that separate connections aren't
792 used when using different login credentials in protocols that shouldn't re-use
793 a connection under such circumstances.
795 Unfortunately, ftpserver.pl doesn't appear to support multiple concurrent
796 connections. The read while() loop seems to loop until it receives a disconnect
797 from the client, where it then enters the waiting for connections loop. When
798 the client opens a second connection to the server, the first connection hasn't
799 been dropped (unless it has been forced - which we shouldn't do in these tests)
800 and thus the wait for connections loop is never entered to receive the second
803 19.6 Use the RFC6265 test suite
805 A test suite made for HTTP cookies (RFC 6265) by Adam Barth is available at
806 https://github.com/abarth/http-state/tree/master/tests
808 It'd be really awesome if someone would write a script/setup that would run
809 curl with that test suite and detect deviances. Ideally, that would even be
810 incorporated into our regular test suite.
815 20.1 http-style HEAD output for FTP
817 #undef CURL_FTP_HTTPSTYLE_HEAD in lib/ftp.c to remove the HTTP-style headers
818 from being output in NOBODY requests over FTP
820 20.2 combine error codes
822 Combine some of the error codes to remove duplicates. The original
823 numbering should not be changed, and the old identifiers would be
824 macroed to the new ones in an CURL_NO_OLDIES section to help with
825 backward compatibility.
827 Candidates for removal and their replacements:
829 CURLE_FILE_COULDNT_READ_FILE => CURLE_REMOTE_FILE_NOT_FOUND
831 CURLE_FTP_COULDNT_RETR_FILE => CURLE_REMOTE_FILE_NOT_FOUND
833 CURLE_FTP_COULDNT_USE_REST => CURLE_RANGE_ERROR
835 CURLE_FUNCTION_NOT_FOUND => CURLE_FAILED_INIT
837 CURLE_LDAP_INVALID_URL => CURLE_URL_MALFORMAT
839 CURLE_TFTP_NOSUCHUSER => CURLE_TFTP_ILLEGAL
841 CURLE_TFTP_NOTFOUND => CURLE_REMOTE_FILE_NOT_FOUND
843 CURLE_TFTP_PERM => CURLE_REMOTE_ACCESS_DENIED
845 20.3 extend CURLOPT_SOCKOPTFUNCTION prototype
847 The current prototype only provides 'purpose' that tells what the
848 connection/socket is for, but not any protocol or similar. It makes it hard
849 for applications to differentiate on TCP vs UDP and even HTTP vs FTP and
852 21. Next major release
854 21.1 cleanup return codes
856 curl_easy_cleanup() returns void, but curl_multi_cleanup() returns a
857 CURLMcode. These should be changed to be the same.
859 21.2 remove obsolete defines
861 remove obsolete defines from curl/curl.h
865 make several functions use size_t instead of int in their APIs
867 21.4 remove several functions
869 remove the following functions from the public API:
873 curl_mprintf (and variations)
879 They will instead become curlx_ - alternatives. That makes the curl app
880 still capable of using them, by building with them from source.
882 These functions have no purpose anymore:
886 curl_multi_socket_all
888 21.5 remove CURLOPT_FAILONERROR
890 Remove support for CURLOPT_FAILONERROR, it has gotten too kludgy and weird
891 internally. Let the app judge success or not for itself.
893 21.6 remove CURLOPT_DNS_USE_GLOBAL_CACHE
895 Remove support for a global DNS cache. Anything global is silly, and we
896 already offer the share interface for the same functionality but done
899 21.7 remove progress meter from libcurl
901 The internally provided progress meter output doesn't belong in the library.
902 Basically no application wants it (apart from curl) but instead applications
903 can and should do their own progress meters using the progress callback.
905 The progress callback should then be bumped as well to get proper 64bit
906 variable types passed to it instead of doubles so that big files work
909 21.8 remove 'curl_httppost' from public
911 curl_formadd() was made to fill in a public struct, but the fact that the
912 struct is public is never really used by application for their own advantage
913 but instead often restricts how the form functions can or can't be modified.
915 Changing them to return a private handle will benefit the implementation and
916 allow us much greater freedoms while still maintaining a solid API and ABI.