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 All bugs documented in the KNOWN_BUGS document are subject for fixing!
17 1.4 signal-based resolver timeouts
18 1.5 get rid of PATH_MAX
19 1.6 Modified buffer size approach
20 1.7 Detect when called from within callbacks
21 1.8 Allow SSL (HTTPS) to proxy
23 2. libcurl - multi interface
25 2.2 Fix HTTP Pipelining for PUT
28 3.1 Update date and version in man pages
32 4.2 Alter passive/active on failure and retry
33 4.3 Earlier bad letter detection
34 4.4 REST for large files
38 5.1 Better persistency for HTTP 1.0
39 5.2 support FF3 sqlite cookie files
40 5.3 Rearrange request header order
46 6.2 ditch telnet-specific select
47 6.3 feature negotiation debug data
48 6.4 send data in chunks
52 7.2 Enhanced capability support
56 8.2 Enhanced capability support
59 9.1 Enhanced capability support
62 10.1 SASL based authentication mechanisms
68 12.1 Disable specific versions
69 12.2 Provide mutex locking API
70 12.3 Evaluate SSL patches
71 12.4 Cache OpenSSL contexts
72 12.5 Export session ids
73 12.6 Provide callback for cert verification
74 12.7 improve configure --with-ssl
82 14.1 Other authentication mechanisms
87 15.3 prevent file overwriting
88 15.4 simultaneous parallel transfers
89 15.5 provide formpost headers
90 15.6 warning when setting an option
97 17.2 nicer lacking perl message
98 17.3 more protocols supported
99 17.4 more platforms supported
100 17.5 Add support for concurrent connections
103 18.1 http-style HEAD output for FTP
104 18.2 combine error codes
105 18.3 extend CURLOPT_SOCKOPTFUNCTION prototype
107 19. Next major release
108 19.1 cleanup return codes
109 19.2 remove obsolete defines
111 19.4 remove several functions
112 19.5 remove CURLOPT_FAILONERROR
113 19.6 remove CURLOPT_DNS_USE_GLOBAL_CACHE
114 19.7 remove progress meter from libcurl
115 19.8 remove 'curl_httppost' from public
116 19.9 have form functions use CURL handle argument
117 19.10 Add CURLOPT_MAIL_CLIENT option
119 ==============================================================================
123 1.2 More data sharing
125 curl_share_* functions already exist and work, and they can be extended to
126 share more. For example, enable sharing of the ares channel and the
131 Use 'struct lifreq' and SIOCGLIFADDR instead of 'struct ifreq' and
132 SIOCGIFADDR on newer Solaris versions as they claim the latter is obsolete.
133 To support ipv6 interface addresses for network interfaces properly.
135 1.4 signal-based resolver timeouts
137 libcurl built without an asynchronous resolver library uses alarm() to time
138 out DNS lookups. When a timeout occurs, this causes libcurl to jump from the
139 signal handler back into the library with a sigsetjmp, which effectively
140 causes libcurl to continue running within the signal handler. This is
141 non-portable and could cause problems on some platforms. A discussion on the
142 problem is available at http://curl.haxx.se/mail/lib-2008-09/0197.html
144 Also, alarm() provides timeout resolution only to the nearest second. alarm
145 ought to be replaced by setitimer on systems that support it.
147 1.5 get rid of PATH_MAX
149 Having code use and rely on PATH_MAX is not nice:
150 http://insanecoding.blogspot.com/2007/11/pathmax-simply-isnt.html
152 Currently the SSH based code uses it a bit, but to remove PATH_MAX from there
153 we need libssh2 to properly tell us when we pass in a too small buffer and
154 its current API (as of libssh2 1.2.7) doesn't.
156 1.6 Modified buffer size approach
158 Current libcurl allocates a fixed 16K size buffer for download and an
159 additional 16K for upload. They are always unconditionally part of the easy
160 handle. If CRLF translations are requested, an additional 32K "scratch
161 buffer" is allocated. A total of 64K transfer buffers in the worst case.
163 First, while the handles are not actually in use these buffers could be freed
164 so that lingering handles just kept in queues or whatever waste less memory.
166 Secondly, SFTP is a protocol that needs to handle many ~30K blocks at once
167 since each need to be individually acked and therefore libssh2 must be
168 allowed to send (or receive) many separate ones in parallel to achieve high
169 transfer speeds. A current libcurl build with a 16K buffer makes that
170 impossible, but one with a 512K buffer will reach MUCH faster transfers. But
171 allocating 512K unconditionally for all buffers just in case they would like
172 to do fast SFTP transfers at some point is not a good solution either.
174 Dynamically allocate buffer size depending on protocol in use in combination
175 with freeing it after each individual transfer? Other suggestions?
177 1.7 Detect when called from within callbacks
179 We should set a state variable before calling callbacks, so that we
180 subsequently can add code within libcurl that returns error if called within
181 callbacks for when that's not supported.
183 1.8 Allow SSL (HTTPS) to proxy
185 To prevent local users from snooping on your traffic to the proxy. Supported
187 http://www.chromium.org/developers/design-documents/secure-web-proxy
189 ...and by Firefox soon:
190 https://bugzilla.mozilla.org/show_bug.cgi?id=378637
192 2. libcurl - multi interface
194 2.1 More non-blocking
196 Make sure we don't ever loop because of non-blocking sockets returning
197 EWOULDBLOCK or similar. Blocking cases include:
199 - Name resolves on non-windows unless c-ares is used
200 - NSS SSL connections
201 - HTTP proxy CONNECT operations
202 - SOCKS proxy handshakes
205 - The "DONE" operation (post transfer protocol-specific actions) for the
206 protocols SFTP, SMTP, FTP. Fixing Curl_done() for this is a worthy task.
208 2.2 Fix HTTP Pipelining for PUT
210 HTTP Pipelining can be a way to greatly enhance performance for multiple
211 serial requests and currently libcurl only supports that for HEAD and GET
212 requests but it should also be possible for PUT.
216 3.1 Update date and version in man pages
218 'maketgz' or another suitable script could update the .TH sections of the man
219 pages at release time to use the current date and curl/libcurl version
226 HOST is a command for a client to tell which host name to use, to offer FTP
227 servers named-based virtual hosting:
229 http://tools.ietf.org/html/rfc7151
231 4.2 Alter passive/active on failure and retry
233 When trying to connect passively to a server which only supports active
234 connections, libcurl returns CURLE_FTP_WEIRD_PASV_REPLY and closes the
235 connection. There could be a way to fallback to an active connection (and
236 vice versa). http://curl.haxx.se/bug/feature.cgi?id=1754793
238 4.3 Earlier bad letter detection
240 Make the detection of (bad) %0d and %0a codes in FTP URL parts earlier in the
241 process to avoid doing a resolve and connect in vain.
243 4.4 REST for large files
245 REST fix for servers not behaving well on >2GB requests. This should fail if
246 the server doesn't set the pointer to the requested index. The tricky
247 (impossible?) part is to figure out if the server did the right thing or not.
251 FTP ASCII transfers do not follow RFC959. They don't convert the data
256 5.1 Better persistency for HTTP 1.0
258 "Better" support for persistent connections over HTTP 1.0
259 http://curl.haxx.se/bug/feature.cgi?id=1089001
261 5.2 support FF3 sqlite cookie files
263 Firefox 3 is changing from its former format to a a sqlite database instead.
264 We should consider how (lib)curl can/should support this.
265 http://curl.haxx.se/bug/feature.cgi?id=1871388
267 5.3 Rearrange request header order
269 Server implementors often make an effort to detect browser and to reject
270 clients it can detect to not match. One of the last details we cannot yet
271 control in libcurl's HTTP requests, which also can be exploited to detect
272 that libcurl is in fact used even when it tries to impersonate a browser, is
273 the order of the request headers. I propose that we introduce a new option in
274 which you give headers a value, and then when the HTTP request is built it
275 sorts the headers based on that number. We could then have internally created
276 headers use a default value so only headers that need to be moved have to be
281 Chrome and Firefox already support SPDY and lots of web services do. There's
282 a library for us to use for this (spdylay) that has a similar API and the
283 same author as nghttp2.
285 spdylay: https://github.com/tatsuhiro-t/spdylay
289 Add the ability to specify the preferred authentication mechanism to use by
290 using ;auth=<mech> in the login part of the URL.
294 http://test:pass;auth=NTLM@example.com would be equivalent to specifying --user
295 test:pass;auth=NTLM or --user test:pass --ntlm from the command line.
297 Additionally this should be implemented for proxy base URLs as well.
303 Reading input (to send to the remote server) on stdin is a crappy solution for
304 library purposes. We need to invent a good way for the application to be able
305 to provide the data to send.
307 6.2 ditch telnet-specific select
309 Move the telnet support's network select() loop go away and merge the code
310 into the main transfer loop. Until this is done, the multi interface won't
313 6.3 feature negotiation debug data
315 Add telnet feature negotiation data to the debug callback as header data.
317 6.4 send data in chunks
319 Currently, telnet sends data one byte at a time. This is fine for interactive
320 use, but inefficient for any other. Sent data should be sent in larger
327 Add support for pipelining emails.
329 7.2 Enhanced capability support
331 Add the ability, for an application that uses libcurl, to obtain the list of
332 capabilities returned from the EHLO command.
338 Add support for pipelining commands.
340 8.2 Enhanced capability support
342 Add the ability, for an application that uses libcurl, to obtain the list of
343 capabilities returned from the CAPA command.
347 9.1 Enhanced capability support
349 Add the ability, for an application that uses libcurl, to obtain the list of
350 capabilities returned from the CAPABILITY command.
354 10.1 SASL based authentication mechanisms
356 Currently the LDAP module only supports ldap_simple_bind_s() in order to bind
357 to an LDAP server. However, this function sends username and password details
358 using the simple authentication mechanism (as clear text). However, it should
359 be possible to use ldap_bind_s() instead specifying the security context
360 information ourselves.
366 There's no RFC for the protocol or an URI/URL format. An implementation
367 should most probably use an existing rsync library, such as librsync.
371 12.1 Disable specific versions
373 Provide an option that allows for disabling specific SSL versions, such as
374 SSLv2 http://curl.haxx.se/bug/feature.cgi?id=1767276
376 12.2 Provide mutex locking API
378 Provide a libcurl API for setting mutex callbacks in the underlying SSL
379 library, so that the same application code can use mutex-locking
380 independently of OpenSSL or GnutTLS being used.
382 12.3 Evaluate SSL patches
384 Evaluate/apply Gertjan van Wingerde's SSL patches:
385 http://curl.haxx.se/mail/lib-2004-03/0087.html
387 12.4 Cache OpenSSL contexts
389 "Look at SSL cafile - quick traces look to me like these are done on every
390 request as well, when they should only be necessary once per SSL context (or
391 once per handle)". The major improvement we can rather easily do is to make
392 sure we don't create and kill a new SSL "context" for every request, but
393 instead make one for every connection and re-use that SSL context in the same
394 style connections are re-used. It will make us use slightly more memory but
395 it will libcurl do less creations and deletions of SSL contexts.
397 12.5 Export session ids
399 Add an interface to libcurl that enables "session IDs" to get
400 exported/imported. Cris Bailiff said: "OpenSSL has functions which can
401 serialise the current SSL state to a buffer of your choice, and recover/reset
402 the state from such a buffer at a later date - this is used by mod_ssl for
403 apache to implement and SSL session ID cache".
405 12.6 Provide callback for cert verification
407 OpenSSL supports a callback for customised verification of the peer
408 certificate, but this doesn't seem to be exposed in the libcurl APIs. Could
409 it be? There's so much that could be done if it were!
411 12.7 improve configure --with-ssl
413 make the configure --with-ssl option first check for OpenSSL, then GnuTLS,
418 DNS-Based Authentication of Named Entities (DANE) is a way to provide SSL
419 keys and certs over DNS using DNSSEC as an alternative to the CA model.
420 http://www.rfc-editor.org/rfc/rfc6698.txt
422 An initial patch was posted by Suresh Krishnaswamy on March 7th 2013
423 (http://curl.haxx.se/mail/lib-2013-03/0075.html) but it was a too simple
424 approach. See Daniel's comments:
425 http://curl.haxx.se/mail/lib-2013-03/0103.html . libunbound may be the
426 correct library to base this development on.
430 13.1 SSL engine stuff
432 Is this even possible?
434 13.2 check connection
436 Add a way to check if the connection seems to be alive, to correspond to the
437 SSL_peak() way we use with OpenSSL.
441 14.1 Other authentication mechanisms
443 Add support for GSSAPI to SMTP, POP3 and IMAP.
449 "curl --sync http://example.com/feed[1-100].rss" or
450 "curl --sync http://example.net/{index,calendar,history}.html"
452 Downloads a range or set of URLs using the remote name, but only if the
453 remote file is newer than the local file. A Last-Modified HTTP date header
454 should also be used to set the mod date on the downloaded file.
458 Globbing support for -d and -F, as in 'curl -d "name=foo[0-9]" URL'.
459 This is easily scripted though.
461 15.3 prevent file overwriting
463 Add an option that prevents cURL from overwriting existing local files. When
464 used, and there already is an existing file with the target file name
465 (either -O or -o), a number should be appended (and increased if already
466 existing). So that index.html becomes first index.html.1 and then
469 15.4 simultaneous parallel transfers
471 The client could be told to use maximum N simultaneous parallel transfers and
472 then just make sure that happens. It should of course not make more than one
473 connection to the same remote host. This would require the client to use the
474 multi interface. http://curl.haxx.se/bug/feature.cgi?id=1558595
476 15.5 provide formpost headers
478 Extending the capabilities of the multipart formposting. How about leaving
479 the ';type=foo' syntax as it is and adding an extra tag (headers) which
480 works like this: curl -F "coolfiles=@fil1.txt;headers=@fil1.hdr" where
481 fil1.hdr contains extra headers like
483 Content-Type: text/plain; charset=KOI8-R"
484 Content-Transfer-Encoding: base64
485 X-User-Comment: Please don't use browser specific HTML code
487 which should overwrite the program reasonable defaults (plain/text,
490 15.6 warning when setting an option
492 Display a warning when libcurl returns an error when setting an option.
493 This can be useful to tell when support for a particular feature hasn't been
494 compiled into the library.
500 Consider extending 'roffit' to produce decent ASCII output, and use that
501 instead of (g)nroff when building src/tool_hugehelp.c
507 Make our own version of stunnel for simple port forwarding to enable HTTPS
508 and FTP-SSL tests without the stunnel dependency, and it could allow us to
509 provide test tools built with either OpenSSL or GnuTLS
511 17.2 nicer lacking perl message
513 If perl wasn't found by the configure script, don't attempt to run the tests
514 but explain something nice why it doesn't.
516 17.3 more protocols supported
518 Extend the test suite to include more protocols. The telnet could just do FTP
519 or http operations (for which we have test servers).
521 17.4 more platforms supported
523 Make the test suite work on more platforms. OpenBSD and Mac OS. Remove
524 fork()s and it should become even more portable.
526 17.5 Add support for concurrent connections
528 Tests 836, 882 and 938 were designed to verify that separate connections aren't
529 used when using different login credentials in protocols that shouldn't re-use
530 a connection under such circumstances.
532 Unfortunately, ftpserver.pl doesn't appear to support multiple concurrent
533 connections. The read while() loop seems to loop until it receives a disconnect
534 from the client, where it then enters the waiting for connections loop. When
535 the client opens a second connection to the server, the first connection hasn't
536 been dropped (unless it has been forced - which we shouldn't do in these tests)
537 and thus the wait for connections loop is never entered to receive the second
542 18.1 http-style HEAD output for FTP
544 #undef CURL_FTP_HTTPSTYLE_HEAD in lib/ftp.c to remove the HTTP-style headers
545 from being output in NOBODY requests over FTP
547 18.2 combine error codes
549 Combine some of the error codes to remove duplicates. The original
550 numbering should not be changed, and the old identifiers would be
551 macroed to the new ones in an CURL_NO_OLDIES section to help with
552 backward compatibility.
554 Candidates for removal and their replacements:
556 CURLE_FILE_COULDNT_READ_FILE => CURLE_REMOTE_FILE_NOT_FOUND
558 CURLE_FTP_COULDNT_RETR_FILE => CURLE_REMOTE_FILE_NOT_FOUND
560 CURLE_FTP_COULDNT_USE_REST => CURLE_RANGE_ERROR
562 CURLE_FUNCTION_NOT_FOUND => CURLE_FAILED_INIT
564 CURLE_LDAP_INVALID_URL => CURLE_URL_MALFORMAT
566 CURLE_TFTP_NOSUCHUSER => CURLE_TFTP_ILLEGAL
568 CURLE_TFTP_NOTFOUND => CURLE_REMOTE_FILE_NOT_FOUND
570 CURLE_TFTP_PERM => CURLE_REMOTE_ACCESS_DENIED
572 18.3 extend CURLOPT_SOCKOPTFUNCTION prototype
574 The current prototype only provides 'purpose' that tells what the
575 connection/socket is for, but not any protocol or similar. It makes it hard
576 for applications to differentiate on TCP vs UDP and even HTTP vs FTP and
579 19. Next major release
581 19.1 cleanup return codes
583 curl_easy_cleanup() returns void, but curl_multi_cleanup() returns a
584 CURLMcode. These should be changed to be the same.
586 19.2 remove obsolete defines
588 remove obsolete defines from curl/curl.h
592 make several functions use size_t instead of int in their APIs
594 19.4 remove several functions
596 remove the following functions from the public API:
600 curl_mprintf (and variations)
606 They will instead become curlx_ - alternatives. That makes the curl app
607 still capable of using them, by building with them from source.
609 These functions have no purpose anymore:
613 curl_multi_socket_all
615 19.5 remove CURLOPT_FAILONERROR
617 Remove support for CURLOPT_FAILONERROR, it has gotten too kludgy and weird
618 internally. Let the app judge success or not for itself.
620 19.6 remove CURLOPT_DNS_USE_GLOBAL_CACHE
622 Remove support for a global DNS cache. Anything global is silly, and we
623 already offer the share interface for the same functionality but done
626 19.7 remove progress meter from libcurl
628 The internally provided progress meter output doesn't belong in the library.
629 Basically no application wants it (apart from curl) but instead applications
630 can and should do their own progress meters using the progress callback.
632 The progress callback should then be bumped as well to get proper 64bit
633 variable types passed to it instead of doubles so that big files work
636 19.8 remove 'curl_httppost' from public
638 curl_formadd() was made to fill in a public struct, but the fact that the
639 struct is public is never really used by application for their own advantage
640 but instead often restricts how the form functions can or can't be modified.
642 Changing them to return a private handle will benefit the implementation and
643 allow us much greater freedoms while still maintaining a solid API and ABI.
645 19.9 have form functions use CURL handle argument
647 curl_formadd() and curl_formget() both currently have no CURL handle
648 argument, but both can use a callback that is set in the easy handle, and
649 thus curl_formget() with callback cannot function without first having
650 curl_easy_perform() (or similar) called - which is hard to grasp and a design
653 19.10 Add CURLOPT_MAIL_CLIENT option
655 Rather than use the URL to specify the mail client string to present in the
656 HELO and EHLO commands, libcurl should support a new CURLOPT specifically for
657 specifying this data as the URL is non-standard and to be honest a bit of a
660 Please see the following thread for more information:
661 http://curl.haxx.se/mail/lib-2012-05/0178.html