_ _ ____ _ ___| | | | _ \| | / __| | | | |_) | | | (__| |_| | _ <| |___ \___|\___/|_| \_\_____| Things that could be nice to do in the future Things to do in project curl. Please tell us what you think, contribute and send us patches that improve things! Be aware that these are things that we could do, or have once been considered things we could do. If you want to work on any of these areas, please consider bringing it up for discussions first on the mailing list so that we all agree it is still a good idea for the project! All bugs documented in the KNOWN_BUGS document are subject for fixing! 1. libcurl 1.2 More data sharing 1.3 struct lifreq 1.4 signal-based resolver timeouts 1.5 get rid of PATH_MAX 1.6 Modified buffer size approach 1.7 Support HTTP/2 for HTTP(S) proxies 1.8 CURLOPT_RESOLVE for any port number 1.9 Cache negative name resolves 1.10 auto-detect proxy 1.11 minimize dependencies with dynamically loaded modules 1.12 updated DNS server while running 1.13 c-ares and CURLOPT_OPENSOCKETFUNCTION 1.14 Typesafe curl_easy_setopt() 1.15 Monitor connections in the connection pool 1.16 Try to URL encode given URL 1.17 Add support for IRIs 1.18 try next proxy if one doesn't work 1.19 Timeout idle connections from the pool 1.20 SRV and URI DNS records 1.21 API for URL parsing/splitting 1.22 CURLINFO_PAUSE_STATE 1.23 Offer API to flush the connection pool 1.24 TCP Fast Open for windows 1.25 Expose tried IP addresses that failed 1.26 CURL_REFUSE_CLEARTEXT 1.27 hardcode the "localhost" addresses 1.28 FD_CLOEXEC 2. libcurl - multi interface 2.1 More non-blocking 2.2 Better support for same name resolves 2.3 Non-blocking curl_multi_remove_handle() 2.4 Split connect and authentication process 2.5 Edge-triggered sockets should work 3. Documentation 3.2 Provide cmake config-file 4. FTP 4.1 HOST 4.2 Alter passive/active on failure and retry 4.3 Earlier bad letter detection 4.4 REST for large files 4.5 ASCII support 4.6 GSSAPI via Windows SSPI 4.7 STAT for LIST without data connection 4.8 Option to ignore private IP addresses in PASV response 5. HTTP 5.1 Better persistency for HTTP 1.0 5.2 support FF3 sqlite cookie files 5.3 Rearrange request header order 5.5 auth= in URLs 5.6 Refuse "downgrade" redirects 5.7 QUIC 5.8 Leave secure cookies alone 6. TELNET 6.1 ditch stdin 6.2 ditch telnet-specific select 6.3 feature negotiation debug data 7. SMTP 7.1 Pipelining 7.2 Enhanced capability support 7.3 Add CURLOPT_MAIL_CLIENT option 8. POP3 8.1 Pipelining 8.2 Enhanced capability support 9. IMAP 9.1 Enhanced capability support 10. LDAP 10.1 SASL based authentication mechanisms 11. SMB 11.1 File listing support 11.2 Honor file timestamps 11.3 Use NTLMv2 11.4 Create remote directories 12. New protocols 12.1 RSYNC 13. SSL 13.1 Disable specific versions 13.2 Provide mutex locking API 13.3 Support in-memory certs/ca certs/keys 13.4 Cache/share OpenSSL contexts 13.5 Export session ids 13.6 Provide callback for cert verification 13.7 improve configure --with-ssl 13.8 Support DANE 13.9 Configurable loading of OpenSSL configuration file 13.10 Support Authority Information Access certificate extension (AIA) 13.11 Support intermediate & root pinning for PINNEDPUBLICKEY 13.12 Support HSTS 13.13 Support HPKP 13.14 Support the clienthello extension 14. GnuTLS 14.1 SSL engine stuff 14.2 check connection 15. WinSSL/SChannel 15.1 Add support for client certificate authentication 15.3 Add support for the --ciphers option 16. SASL 16.1 Other authentication mechanisms 16.2 Add QOP support to GSSAPI authentication 16.3 Support binary messages (i.e.: non-base64) 17. SSH protocols 17.1 Multiplexing 17.2 SFTP performance 17.3 Support better than MD5 hostkey hash 17.4 Support CURLOPT_PREQUOTE 18. Command line tool 18.1 sync 18.2 glob posts 18.3 prevent file overwriting 18.4 simultaneous parallel transfers 18.5 UTF-8 filenames in Content-Disposition 18.6 warning when setting an option 18.7 warning if curl version is not in sync with libcurl version 18.8 offer color-coded HTTP header output 18.9 Choose the name of file in braces for complex URLs 18.10 improve how curl works in a windows console window 18.11 -w output to stderr 18.12 keep running, read instructions from pipe/socket 18.13 support metalink in http headers 18.14 --fail without --location should treat 3xx as a failure 18.15 --retry should resume 18.16 send only part of --data 18.17 consider file name from the redirected URL with -O ? 18.18 retry on network is unreachable 18.19 expand ~/ in config files 18.20 host name sections in config files 19. Build 19.1 roffit 19.2 Enable PIE and RELRO by default 20. Test suite 20.1 SSL tunnel 20.2 nicer lacking perl message 20.3 more protocols supported 20.4 more platforms supported 20.5 Add support for concurrent connections 20.6 Use the RFC6265 test suite 21. Next SONAME bump 21.1 http-style HEAD output for FTP 21.2 combine error codes 21.3 extend CURLOPT_SOCKOPTFUNCTION prototype 22. Next major release 22.1 cleanup return codes 22.2 remove obsolete defines 22.3 size_t 22.4 remove several functions 22.5 remove CURLOPT_FAILONERROR 22.6 remove CURLOPT_DNS_USE_GLOBAL_CACHE 22.7 remove progress meter from libcurl 22.8 remove 'curl_httppost' from public ============================================================================== 1. libcurl 1.2 More data sharing curl_share_* functions already exist and work, and they can be extended to share more. For example, enable sharing of the ares channel. 1.3 struct lifreq Use 'struct lifreq' and SIOCGLIFADDR instead of 'struct ifreq' and SIOCGIFADDR on newer Solaris versions as they claim the latter is obsolete. To support IPv6 interface addresses for network interfaces properly. 1.4 signal-based resolver timeouts libcurl built without an asynchronous resolver library uses alarm() to time out DNS lookups. When a timeout occurs, this causes libcurl to jump from the signal handler back into the library with a sigsetjmp, which effectively causes libcurl to continue running within the signal handler. This is non-portable and could cause problems on some platforms. A discussion on the problem is available at https://curl.haxx.se/mail/lib-2008-09/0197.html Also, alarm() provides timeout resolution only to the nearest second. alarm ought to be replaced by setitimer on systems that support it. 1.5 get rid of PATH_MAX Having code use and rely on PATH_MAX is not nice: https://insanecoding.blogspot.com/2007/11/pathmax-simply-isnt.html Currently the SSH based code uses it a bit, but to remove PATH_MAX from there we need libssh2 to properly tell us when we pass in a too small buffer and its current API (as of libssh2 1.2.7) doesn't. 1.6 Modified buffer size approach Current libcurl allocates a fixed 16K size buffer for download and an additional 16K for upload. They are always unconditionally part of the easy handle. If CRLF translations are requested, an additional 32K "scratch buffer" is allocated. A total of 64K transfer buffers in the worst case. First, while the handles are not actually in use these buffers could be freed so that lingering handles just kept in queues or whatever waste less memory. Secondly, SFTP is a protocol that needs to handle many ~30K blocks at once since each need to be individually acked and therefore libssh2 must be allowed to send (or receive) many separate ones in parallel to achieve high transfer speeds. A current libcurl build with a 16K buffer makes that impossible, but one with a 512K buffer will reach MUCH faster transfers. But allocating 512K unconditionally for all buffers just in case they would like to do fast SFTP transfers at some point is not a good solution either. Dynamically allocate buffer size depending on protocol in use in combination with freeing it after each individual transfer? Other suggestions? 1.7 Support HTTP/2 for HTTP(S) proxies Support for doing HTTP/2 to HTTP and HTTPS proxies is still missing. 1.8 CURLOPT_RESOLVE for any port number This option allows applications to set a replacement IP address for a given host + port pair. Consider making support for providing a replacement address for the host name on all port numbers. See https://github.com/curl/curl/issues/1264 1.9 Cache negative name resolves A name resolve that has failed is likely to fail when made again within a short period of time. Currently we only cache positive responses. 1.10 auto-detect proxy libcurl could be made to detect the system proxy setup automatically and use that. On Windows, macOS and Linux desktops for example. The pull-request to use libproxy for this was deferred due to doubts on the reliability of the dependency and how to use it: https://github.com/curl/curl/pull/977 libdetectproxy is a (C++) library for detecting the proxy on Windows https://github.com/paulharris/libdetectproxy 1.11 minimize dependencies with dynamically loaded modules We can create a system with loadable modules/plug-ins, where these modules would be the ones that link to 3rd party libs. That would allow us to avoid having to load ALL dependencies since only the necessary ones for this app/invoke/used protocols would be necessary to load. See https://github.com/curl/curl/issues/349 1.12 updated DNS server while running If /etc/resolv.conf gets updated while a program using libcurl is running, it is may cause name resolves to fail unless res_init() is called. We should consider calling res_init() + retry once unconditionally on all name resolve failures to mitigate against this. Firefox works like that. Note that Windows doesn't have res_init() or an alternative. https://github.com/curl/curl/issues/2251 1.13 c-ares and CURLOPT_OPENSOCKETFUNCTION curl will create most sockets via the CURLOPT_OPENSOCKETFUNCTION callback and close them with the CURLOPT_CLOSESOCKETFUNCTION callback. However, c-ares does not use those functions and instead opens and closes the sockets itself. This means that when curl passes the c-ares socket to the CURLMOPT_SOCKETFUNCTION it isn't owned by the application like other sockets. See https://github.com/curl/curl/issues/2734 1.14 Typesafe curl_easy_setopt() One of the most common problems in libcurl using applications is the lack of type checks for curl_easy_setopt() which happens because it accepts varargs and thus can take any type. One possible solution to this is to introduce a few different versions of the setopt version for the different kinds of data you can set. curl_easy_set_num() - sets a long value curl_easy_set_large() - sets a curl_off_t value curl_easy_set_ptr() - sets a pointer curl_easy_set_cb() - sets a callback PLUS its callback data 1.15 Monitor connections in the connection pool libcurl's connection cache or pool holds a number of open connections for the purpose of possible subsequent connection reuse. It may contain a few up to a significant amount of connections. Currently, libcurl leaves all connections as they are and first when a connection is iterated over for matching or reuse purpose it is verified that it is still alive. Those connections may get closed by the server side for idleness or they may get a HTTP/2 ping from the peer to verify that they're still alive. By adding monitoring of the connections while in the pool, libcurl can detect dead connections (and close them) better and earlier, and it can handle HTTP/2 pings to keep such ones alive even when not actively doing transfers on them. 1.16 Try to URL encode given URL Given a URL that for example contains spaces, libcurl could have an option that would try somewhat harder than it does now and convert spaces to %20 and perhaps URL encoded byte values over 128 etc (basically do what the redirect following code already does). https://github.com/curl/curl/issues/514 1.17 Add support for IRIs IRIs (RFC 3987) allow localized, non-ascii, names in the URL. To properly support this, curl/libcurl would need to translate/encode the given input from the input string encoding into percent encoded output "over the wire". To make that work smoothly for curl users even on Windows, curl would probably need to be able to convert from several input encodings. 1.18 try next proxy if one doesn't work Allow an application to specify a list of proxies to try, and failing to connect to the first go on and try the next instead until the list is exhausted. Browsers support this feature at least when they specify proxies using PACs. https://github.com/curl/curl/issues/896 1.19 Timeout idle connections from the pool libcurl currently keeps connections in its connection pool for an indefinite period of time, until it either gets reused, gets noticed that it has been closed by the server or gets pruned to make room for a new connection. To reduce overhead (especially for when we add monitoring of the connections in the pool), we should introduce a timeout so that connections that have been idle for N seconds get closed. 1.20 SRV and URI DNS records Offer support for resolving SRV and URI DNS records for libcurl to know which server to connect to for various protocols (including HTTP!). 1.21 API for URL parsing/splitting libcurl has always parsed URLs internally and never exposed any API or features to allow applications to do it. Still most or many applications using libcurl need that ability. In polls to users, we've learned that many libcurl users would like to see and use such an API. 1.22 CURLINFO_PAUSE_STATE Return information about the transfer's current pause state, in both directions. https://github.com/curl/curl/issues/2588 1.23 Offer API to flush the connection pool Sometimes applications want to flush all the existing connections kept alive. An API could allow a forced flush or just a forced loop that would properly close all connections that have been closed by the server already. 1.24 TCP Fast Open for windows libcurl supports the CURLOPT_TCP_FASTOPEN option since 7.49.0 for Linux and Mac OS. Windows supports TCP Fast Open starting with Windows 10, version 1607 and we should add support for it. 1.25 Expose tried IP addresses that failed When libcurl fails to connect to a host, it should be able to offer the application the list of IP addresses that were used in the attempt. https://github.com/curl/curl/issues/2126 1.26 CURL_REFUSE_CLEARTEXT An environment variable that when set will make libcurl refuse to use any cleartext network protocol. That's all non-encrypted ones (FTP, HTTP, Gopher, etc). By adding the check to libcurl and not just curl, this environment variable can then help users to block all libcurl-using programs from accessing the network using unsafe protocols. The variable could be given some sort of syntax or different levels and be used to also allow for example users to refuse libcurl to do transfers with HTTPS certificate checks disabled. It could also automatically refuse usernames in URLs when set (see CURLOPT_DISALLOW_USERNAME_IN_URL) 1.27 hardcode the "localhost" addresses There's this new spec getting adopted that says "localhost" should always and unconditionally be a local address and not get resolved by a DNS server. A fine way for curl to fix this would be to simply hard-code the response to 127.0.0.1 and/or ::1 (depending on what IP versions that are requested). This is what the browsers probably will do with this hostname. https://bugzilla.mozilla.org/show_bug.cgi?id=1220810 https://tools.ietf.org/html/draft-ietf-dnsop-let-localhost-be-localhost-02 1.28 FD_CLOEXEC It sets the close-on-exec flag for the file descriptor, which causes the file descriptor to be automatically (and atomically) closed when any of the exec-family functions succeed. Should probably be set by default? https://github.com/curl/curl/issues/2252 2. libcurl - multi interface 2.1 More non-blocking Make sure we don't ever loop because of non-blocking sockets returning EWOULDBLOCK or similar. Blocking cases include: - Name resolves on non-windows unless c-ares or the threaded resolver is used - SOCKS proxy handshakes - file:// transfers - TELNET transfers - The "DONE" operation (post transfer protocol-specific actions) for the protocols SFTP, SMTP, FTP. Fixing Curl_done() for this is a worthy task. 2.2 Better support for same name resolves If a name resolve has been initiated for name NN and a second easy handle wants to resolve that name as well, make it wait for the first resolve to end up in the cache instead of doing a second separate resolve. This is especially needed when adding many simultaneous handles using the same host name when the DNS resolver can get flooded. 2.3 Non-blocking curl_multi_remove_handle() The multi interface has a few API calls that assume a blocking behavior, like add_handle() and remove_handle() which limits what we can do internally. The multi API need to be moved even more into a single function that "drives" everything in a non-blocking manner and signals when something is done. A remove or add would then only ask for the action to get started and then multi_perform() etc still be called until the add/remove is completed. 2.4 Split connect and authentication process The multi interface treats the authentication process as part of the connect phase. As such any failures during authentication won't trigger the relevant QUIT or LOGOFF for protocols such as IMAP, POP3 and SMTP. 2.5 Edge-triggered sockets should work The multi_socket API should work with edge-triggered socket events. One of the internal actions that need to be improved for this to work perfectly is the 'maxloops' handling in transfer.c:readwrite_data(). 3. Documentation 3.2 Provide cmake config-file A config-file package is a set of files provided by us to allow applications to write cmake scripts to find and use libcurl easier. See https://github.com/curl/curl/issues/885 4. FTP 4.1 HOST HOST is a command for a client to tell which host name to use, to offer FTP servers named-based virtual hosting: https://tools.ietf.org/html/rfc7151 4.2 Alter passive/active on failure and retry When trying to connect passively to a server which only supports active connections, libcurl returns CURLE_FTP_WEIRD_PASV_REPLY and closes the connection. There could be a way to fallback to an active connection (and vice versa). https://curl.haxx.se/bug/feature.cgi?id=1754793 4.3 Earlier bad letter detection Make the detection of (bad) %0d and %0a codes in FTP URL parts earlier in the process to avoid doing a resolve and connect in vain. 4.4 REST for large files REST fix for servers not behaving well on >2GB requests. This should fail if the server doesn't set the pointer to the requested index. The tricky (impossible?) part is to figure out if the server did the right thing or not. 4.5 ASCII support FTP ASCII transfers do not follow RFC959. They don't convert the data accordingly. 4.6 GSSAPI via Windows SSPI In addition to currently supporting the SASL GSSAPI mechanism (Kerberos V5) via third-party GSS-API libraries, such as Heimdal or MIT Kerberos, also add support for GSSAPI authentication via Windows SSPI. 4.7 STAT for LIST without data connection Some FTP servers allow STAT for listing directories instead of using LIST, and the response is then sent over the control connection instead of as the otherwise usedw data connection: http://www.nsftools.com/tips/RawFTP.htm#STAT This is not detailed in any FTP specification. 4.8 Option to ignore private IP addresses in PASV response Some servers respond with and some other FTP client implementations can ignore private (RFC 1918 style) IP addresses when received in PASV responses. To consider for libcurl as well. See https://github.com/curl/curl/issues/1455 5. HTTP 5.1 Better persistency for HTTP 1.0 "Better" support for persistent connections over HTTP 1.0 https://curl.haxx.se/bug/feature.cgi?id=1089001 5.2 support FF3 sqlite cookie files Firefox 3 is changing from its former format to a a sqlite database instead. We should consider how (lib)curl can/should support this. https://curl.haxx.se/bug/feature.cgi?id=1871388 5.3 Rearrange request header order Server implementors often make an effort to detect browser and to reject clients it can detect to not match. One of the last details we cannot yet control in libcurl's HTTP requests, which also can be exploited to detect that libcurl is in fact used even when it tries to impersonate a browser, is the order of the request headers. I propose that we introduce a new option in which you give headers a value, and then when the HTTP request is built it sorts the headers based on that number. We could then have internally created headers use a default value so only headers that need to be moved have to be specified. 5.5 auth= in URLs Add the ability to specify the preferred authentication mechanism to use by using ;auth= in the login part of the URL. For example: http://test:pass;auth=NTLM@example.com would be equivalent to specifying --user test:pass;auth=NTLM or --user test:pass --ntlm from the command line. Additionally this should be implemented for proxy base URLs as well. 5.6 Refuse "downgrade" redirects See https://github.com/curl/curl/issues/226 Consider a way to tell curl to refuse to "downgrade" protocol with a redirect and/or possibly a bit that refuses redirect to change protocol completely. 5.7 QUIC The standardization process of QUIC has been taken to the IETF and can be followed on the [IETF QUIC Mailing list](https://www.ietf.org/mailman/listinfo/quic). I'd like us to get on the bandwagon. Ideally, this would be done with a separate library/project to handle the binary/framing layer in a similar fashion to how HTTP/2 is implemented. This, to allow other projects to benefit from the work and to thus broaden the interest and chance of others to participate. 5.8 Leave secure cookies alone Non-secure origins (HTTP sites) should not be allowed to set or modify cookies with the 'secure' property: https://tools.ietf.org/html/draft-ietf-httpbis-cookie-alone-01 6. TELNET 6.1 ditch stdin Reading input (to send to the remote server) on stdin is a crappy solution for library purposes. We need to invent a good way for the application to be able to provide the data to send. 6.2 ditch telnet-specific select Move the telnet support's network select() loop go away and merge the code into the main transfer loop. Until this is done, the multi interface won't work for telnet. 6.3 feature negotiation debug data Add telnet feature negotiation data to the debug callback as header data. 7. SMTP 7.1 Pipelining Add support for pipelining emails. 7.2 Enhanced capability support Add the ability, for an application that uses libcurl, to obtain the list of capabilities returned from the EHLO command. 7.3 Add CURLOPT_MAIL_CLIENT option Rather than use the URL to specify the mail client string to present in the HELO and EHLO commands, libcurl should support a new CURLOPT specifically for specifying this data as the URL is non-standard and to be honest a bit of a hack ;-) Please see the following thread for more information: https://curl.haxx.se/mail/lib-2012-05/0178.html 8. POP3 8.1 Pipelining Add support for pipelining commands. 8.2 Enhanced capability support Add the ability, for an application that uses libcurl, to obtain the list of capabilities returned from the CAPA command. 9. IMAP 9.1 Enhanced capability support Add the ability, for an application that uses libcurl, to obtain the list of capabilities returned from the CAPABILITY command. 10. LDAP 10.1 SASL based authentication mechanisms Currently the LDAP module only supports ldap_simple_bind_s() in order to bind to an LDAP server. However, this function sends username and password details using the simple authentication mechanism (as clear text). However, it should be possible to use ldap_bind_s() instead specifying the security context information ourselves. 11. SMB 11.1 File listing support Add support for listing the contents of a SMB share. The output should probably be the same as/similar to FTP. 11.2 Honor file timestamps The timestamp of the transferred file should reflect that of the original file. 11.3 Use NTLMv2 Currently the SMB authentication uses NTLMv1. 11.4 Create remote directories Support for creating remote directories when uploading a file to a directory that doesn't exist on the server, just like --ftp-create-dirs. 12. New protocols 12.1 RSYNC There's no RFC for the protocol or an URI/URL format. An implementation should most probably use an existing rsync library, such as librsync. 13. SSL 13.1 Disable specific versions Provide an option that allows for disabling specific SSL versions, such as SSLv2 https://curl.haxx.se/bug/feature.cgi?id=1767276 13.2 Provide mutex locking API Provide a libcurl API for setting mutex callbacks in the underlying SSL library, so that the same application code can use mutex-locking independently of OpenSSL or GnutTLS being used. 13.3 Support in-memory certs/ca certs/keys You can specify the private and public keys for SSH/SSL as file paths. Some programs want to avoid using files and instead just pass them as in-memory data blobs. There's probably a challenge to make this work across the plethory of different TLS and SSH backends that curl supports. https://github.com/curl/curl/issues/2310 13.4 Cache/share OpenSSL contexts "Look at SSL cafile - quick traces look to me like these are done on every request as well, when they should only be necessary once per SSL context (or once per handle)". The major improvement we can rather easily do is to make sure we don't create and kill a new SSL "context" for every request, but instead make one for every connection and re-use that SSL context in the same style connections are re-used. It will make us use slightly more memory but it will libcurl do less creations and deletions of SSL contexts. Technically, the "caching" is probably best implemented by getting added to the share interface so that easy handles who want to and can reuse the context specify that by sharing with the right properties set. https://github.com/curl/curl/issues/1110 13.5 Export session ids Add an interface to libcurl that enables "session IDs" to get exported/imported. Cris Bailiff said: "OpenSSL has functions which can serialise the current SSL state to a buffer of your choice, and recover/reset the state from such a buffer at a later date - this is used by mod_ssl for apache to implement and SSL session ID cache". 13.6 Provide callback for cert verification OpenSSL supports a callback for customised verification of the peer certificate, but this doesn't seem to be exposed in the libcurl APIs. Could it be? There's so much that could be done if it were! 13.7 improve configure --with-ssl make the configure --with-ssl option first check for OpenSSL, then GnuTLS, then NSS... 13.8 Support DANE DNS-Based Authentication of Named Entities (DANE) is a way to provide SSL keys and certs over DNS using DNSSEC as an alternative to the CA model. https://www.rfc-editor.org/rfc/rfc6698.txt An initial patch was posted by Suresh Krishnaswamy on March 7th 2013 (https://curl.haxx.se/mail/lib-2013-03/0075.html) but it was a too simple approach. See Daniel's comments: https://curl.haxx.se/mail/lib-2013-03/0103.html . libunbound may be the correct library to base this development on. Björn Stenberg wrote a separate initial take on DANE that was never completed. 13.9 Configurable loading of OpenSSL configuration file libcurl calls the OpenSSL function CONF_modules_load_file() in openssl.c, Curl_ossl_init(). "We regard any changes in the OpenSSL configuration as a security risk or at least as unnecessary." Please add a configuration switch or something similar to disable the CONF_modules_load_file() call. See https://github.com/curl/curl/issues/2724 13.10 Support Authority Information Access certificate extension (AIA) AIA can provide various things like CRLs but more importantly information about intermediate CA certificates that can allow validation path to be fullfilled when the HTTPS server doesn't itself provide them. Since AIA is about downloading certs on demand to complete a TLS handshake, it is probably a bit tricky to get done right. See https://github.com/curl/curl/issues/2793 13.11 Support intermediate & root pinning for PINNEDPUBLICKEY CURLOPT_PINNEDPUBLICKEY does not consider the hashes of intermediate & root certificates when comparing the pinned keys. Therefore it is not compatible with "HTTP Public Key Pinning" as there also intermediate and root certificates can be pinned. This is very useful as it prevents webadmins from "locking themself out of their servers". Adding this feature would make curls pinning 100% compatible to HPKP and allow more flexible pinning. 13.12 Support HSTS "HTTP Strict Transport Security" is TOFU (trust on first use), time-based features indicated by a HTTP header send by the webserver. It is widely used in browsers and it's purpose is to prevent insecure HTTP connections after a previous HTTPS connection. It protects against SSLStripping attacks. Doc: https://developer.mozilla.org/en-US/docs/Web/Security/HTTP_strict_transport_security RFC 6797: https://tools.ietf.org/html/rfc6797 13.13 Support HPKP "HTTP Public Key Pinning" is TOFU (trust on first use), time-based features indicated by a HTTP header send by the webserver. It's purpose is to prevent Man-in-the-middle attacks by trusted CAs by allowing webadmins to specify which CAs/certificates/public keys to trust when connection to their websites. It can be build based on PINNEDPUBLICKEY. Wikipedia: https://en.wikipedia.org/wiki/HTTP_Public_Key_Pinning OWASP: https://www.owasp.org/index.php/Certificate_and_Public_Key_Pinning Doc: https://developer.mozilla.org/de/docs/Web/Security/Public_Key_Pinning RFC: https://tools.ietf.org/html/draft-ietf-websec-key-pinning-21 13.14 Support the clienthello extension Certain stupid networks and middle boxes have a problem with SSL handshake pakets that are within a certain size range because how that sets some bits that previously (in older TLS version) were not set. The clienthello extension adds padding to avoid that size range. https://tools.ietf.org/html/rfc7685 https://github.com/curl/curl/issues/2299 14. GnuTLS 14.1 SSL engine stuff Is this even possible? 14.2 check connection Add a way to check if the connection seems to be alive, to correspond to the SSL_peak() way we use with OpenSSL. 15. WinSSL/SChannel 15.1 Add support for client certificate authentication WinSSL/SChannel currently makes use of the OS-level system and user certificate and private key stores. This does not allow the application or the user to supply a custom client certificate using curl or libcurl. Therefore support for the existing -E/--cert and --key options should be implemented by supplying a custom certificate to the SChannel APIs, see: - Getting a Certificate for Schannel https://msdn.microsoft.com/en-us/library/windows/desktop/aa375447.aspx 15.3 Add support for the --ciphers option The cipher suites used by WinSSL/SChannel are configured on an OS-level instead of an application-level. This does not allow the application or the user to customize the configured cipher suites using curl or libcurl. Therefore support for the existing --ciphers option should be implemented by mapping the OpenSSL/GnuTLS cipher suites to the SChannel APIs, see - Specifying Schannel Ciphers and Cipher Strengths https://msdn.microsoft.com/en-us/library/windows/desktop/aa380161.aspx 16. SASL 16.1 Other authentication mechanisms Add support for other authentication mechanisms such as OLP, GSS-SPNEGO and others. 16.2 Add QOP support to GSSAPI authentication Currently the GSSAPI authentication only supports the default QOP of auth (Authentication), whilst Kerberos V5 supports both auth-int (Authentication with integrity protection) and auth-conf (Authentication with integrity and privacy protection). 16.3 Support binary messages (i.e.: non-base64) Mandatory to support LDAP SASL authentication. 17. SSH protocols 17.1 Multiplexing SSH is a perfectly fine multiplexed protocols which would allow libcurl to do multiple parallel transfers from the same host using the same connection, much in the same spirit as HTTP/2 does. libcurl however does not take advantage of that ability but will instead always create a new connection for new transfers even if an existing connection already exists to the host. To fix this, libcurl would have to detect an existing connection and "attach" the new transfer to the existing one. 17.2 SFTP performance libcurl's SFTP transfer performance is sub par and can be improved, mostly by the approach mentioned in "1.6 Modified buffer size approach". 17.3 Support better than MD5 hostkey hash libcurl offers the CURLOPT_SSH_HOST_PUBLIC_KEY_MD5 option for verifying the server's key. MD5 is generally being deprecated so we should implement support for stronger hashing algorithms. libssh2 itself is what provides this underlying functionality and it supports at least SHA-1 as an alternative. SHA-1 is also being deprecated these days so we should consider working with libssh2 to instead offer support for SHA-256 or similar. 17.4 Support CURLOPT_PREQUOTE The two other QUOTE options are supported for SFTP, but this was left out for unknown reasons! 18. Command line tool 18.1 sync "curl --sync http://example.com/feed[1-100].rss" or "curl --sync http://example.net/{index,calendar,history}.html" Downloads a range or set of URLs using the remote name, but only if the remote file is newer than the local file. A Last-Modified HTTP date header should also be used to set the mod date on the downloaded file. 18.2 glob posts Globbing support for -d and -F, as in 'curl -d "name=foo[0-9]" URL'. This is easily scripted though. 18.3 prevent file overwriting Add an option that prevents curl from overwriting existing local files. When used, and there already is an existing file with the target file name (either -O or -o), a number should be appended (and increased if already existing). So that index.html becomes first index.html.1 and then index.html.2 etc. 18.4 simultaneous parallel transfers The client could be told to use maximum N simultaneous parallel transfers and then just make sure that happens. It should of course not make more than one connection to the same remote host. This would require the client to use the multi interface. https://curl.haxx.se/bug/feature.cgi?id=1558595 Using the multi interface would also allow properly using parallel transfers with HTTP/2 and supporting HTTP/2 server push from the command line. 18.5 UTF-8 filenames in Content-Disposition RFC 6266 documents how UTF-8 names can be passed to a client in the Content-Disposition header, and curl does not support this. https://github.com/curl/curl/issues/1888 18.6 warning when setting an option Display a warning when libcurl returns an error when setting an option. This can be useful to tell when support for a particular feature hasn't been compiled into the library. 18.7 warning if curl version is not in sync with libcurl version This is usually a sign of a funny, weird or unexpected install situations that aren't always quickly nor easily detected by users. curl and libcurl are always released in sync and should use the same version numbers unless very special situations. 18.8 offer color-coded HTTP header output By offering different color output on the header name and the header contents, they could be made more readable and thus help users working on HTTP services. 18.9 Choose the name of file in braces for complex URLs When using braces to download a list of URLs and you use complicated names in the list of alternatives, it could be handy to allow curl to use other names when saving. Consider a way to offer that. Possibly like {partURL1:name1,partURL2:name2,partURL3:name3} where the name following the colon is the output name. See https://github.com/curl/curl/issues/221 18.10 improve how curl works in a windows console window If you pull the scrollbar when transferring with curl in a Windows console window, the transfer is interrupted and can get disconnected. This can probably be improved. See https://github.com/curl/curl/issues/322 18.11 -w output to stderr -w is quite useful, but not to those of us who use curl without -o or -O (such as for scripting through a higher level language). It would be nice to have an option that is exactly like -w but sends it to stderr instead. Proposed name: --write-stderr. See https://github.com/curl/curl/issues/613 18.12 keep running, read instructions from pipe/socket Provide an option that makes curl not exit after the last URL (or even work without a given URL), and then make it read instructions passed on a pipe or over a socket to make further instructions so that a second subsequent curl invoke can talk to the still running instance and ask for transfers to get done, and thus maintain its connection pool, DNS cache and more. 18.13 support metalink in http headers Curl has support for downloading a metalink xml file, processing it, and then downloading the target of the metalink. This is done via the --metalink option. It would be nice if metalink also supported downloading via metalink information that is stored in HTTP headers (RFC 6249). Theoretically this could also be supported with the --metalink option. See https://tools.ietf.org/html/rfc6249 See also https://lists.gnu.org/archive/html/bug-wget/2015-06/msg00034.html for an implematation of this in wget. 18.14 --fail without --location should treat 3xx as a failure To allow a command line like this to detect a redirect and consider it a failure: curl -v --fail -O https://example.com/curl-7.48.0.tar.gz ... --fail must treat 3xx responses as failures too. The least problematic way to implement this is probably to add that new logic in the command line tool only and not in the underlying CURLOPT_FAILONERROR logic. 18.15 --retry should resume When --retry is used and curl actually retries transfer, it should use the already transferred data and do a resumed transfer for the rest (when possible) so that it doesn't have to transfer the same data again that was already transferred before the retry. See https://github.com/curl/curl/issues/1084 18.16 send only part of --data When the user only wants to send a small piece of the data provided with --data or --data-binary, like when that data is a huge file, consider a way to specify that curl should only send a piece of that. One suggested syntax would be: "--data-binary @largefile.zip!1073741823-2147483647". See https://github.com/curl/curl/issues/1200 18.17 consider file name from the redirected URL with -O ? When a user gives a URL and uses -O, and curl follows a redirect to a new URL, the file name is not extracted and used from the newly redirected-to URL even if the new URL may have a much more sensible file name. This is clearly documented and helps for security since there's no surprise to users which file name that might get overwritten. But maybe a new option could allow for this or maybe -J should imply such a treatment as well as -J already allows for the server to decide what file name to use so it already provides the "may overwrite any file" risk. This is extra tricky if the original URL has no file name part at all since then the current code path will error out with an error message, and we can't *know* already at that point if curl will be redirected to a URL that has a file name... See https://github.com/curl/curl/issues/1241 18.18 retry on network is unreachable The --retry option retries transfers on "transient failures". We later added --retry-connrefused to also retry for "connection refused" errors. Suggestions have been brought to also allow retry on "network is unreachable" errors and while totally reasonable, maybe we should consider a way to make this more configurable than to add a new option for every new error people want to retry for? https://github.com/curl/curl/issues/1603 18.19 expand ~/ in config files For example .curlrc could benefit from being able to do this. See https://github.com/curl/curl/issues/2317 18.20 host name sections in config files config files would be more powerful if they could set different configurations depending on used URLs, host name or possibly origin. Then a default .curlrc could a specific user-agent only when doing requests against a certain site. 19. Build 19.1 roffit Consider extending 'roffit' to produce decent ASCII output, and use that instead of (g)nroff when building src/tool_hugehelp.c 19.2 Enable PIE and RELRO by default Especially when having programs that execute curl via the command line, PIE renders the exploitation of memory corruption vulnerabilities a lot more difficult. This can be attributed to the additional information leaks being required to conduct a successful attack. RELRO, on the other hand, masks different binary sections like the GOT as read-only and thus kills a handful of techniques that come in handy when attackers are able to arbitrarily overwrite memory. A few tests showed that enabling these features had close to no impact, neither on the performance nor on the general functionality of curl. 20. Test suite 20.1 SSL tunnel Make our own version of stunnel for simple port forwarding to enable HTTPS and FTP-SSL tests without the stunnel dependency, and it could allow us to provide test tools built with either OpenSSL or GnuTLS 20.2 nicer lacking perl message If perl wasn't found by the configure script, don't attempt to run the tests but explain something nice why it doesn't. 20.3 more protocols supported Extend the test suite to include more protocols. The telnet could just do FTP or http operations (for which we have test servers). 20.4 more platforms supported Make the test suite work on more platforms. OpenBSD and Mac OS. Remove fork()s and it should become even more portable. 20.5 Add support for concurrent connections Tests 836, 882 and 938 were designed to verify that separate connections aren't used when using different login credentials in protocols that shouldn't re-use a connection under such circumstances. Unfortunately, ftpserver.pl doesn't appear to support multiple concurrent connections. The read while() loop seems to loop until it receives a disconnect from the client, where it then enters the waiting for connections loop. When the client opens a second connection to the server, the first connection hasn't been dropped (unless it has been forced - which we shouldn't do in these tests) and thus the wait for connections loop is never entered to receive the second connection. 20.6 Use the RFC6265 test suite A test suite made for HTTP cookies (RFC 6265) by Adam Barth is available at https://github.com/abarth/http-state/tree/master/tests It'd be really awesome if someone would write a script/setup that would run curl with that test suite and detect deviances. Ideally, that would even be incorporated into our regular test suite. 21. Next SONAME bump 21.1 http-style HEAD output for FTP #undef CURL_FTP_HTTPSTYLE_HEAD in lib/ftp.c to remove the HTTP-style headers from being output in NOBODY requests over FTP 21.2 combine error codes Combine some of the error codes to remove duplicates. The original numbering should not be changed, and the old identifiers would be macroed to the new ones in an CURL_NO_OLDIES section to help with backward compatibility. Candidates for removal and their replacements: CURLE_FILE_COULDNT_READ_FILE => CURLE_REMOTE_FILE_NOT_FOUND CURLE_FTP_COULDNT_RETR_FILE => CURLE_REMOTE_FILE_NOT_FOUND CURLE_FTP_COULDNT_USE_REST => CURLE_RANGE_ERROR CURLE_FUNCTION_NOT_FOUND => CURLE_FAILED_INIT CURLE_LDAP_INVALID_URL => CURLE_URL_MALFORMAT CURLE_TFTP_NOSUCHUSER => CURLE_TFTP_ILLEGAL CURLE_TFTP_NOTFOUND => CURLE_REMOTE_FILE_NOT_FOUND CURLE_TFTP_PERM => CURLE_REMOTE_ACCESS_DENIED 21.3 extend CURLOPT_SOCKOPTFUNCTION prototype The current prototype only provides 'purpose' that tells what the connection/socket is for, but not any protocol or similar. It makes it hard for applications to differentiate on TCP vs UDP and even HTTP vs FTP and similar. 22. Next major release 22.1 cleanup return codes curl_easy_cleanup() returns void, but curl_multi_cleanup() returns a CURLMcode. These should be changed to be the same. 22.2 remove obsolete defines remove obsolete defines from curl/curl.h 22.3 size_t make several functions use size_t instead of int in their APIs 22.4 remove several functions remove the following functions from the public API: curl_getenv curl_mprintf (and variations) curl_strequal curl_strnequal They will instead become curlx_ - alternatives. That makes the curl app still capable of using them, by building with them from source. These functions have no purpose anymore: curl_multi_socket curl_multi_socket_all 22.5 remove CURLOPT_FAILONERROR Remove support for CURLOPT_FAILONERROR, it has gotten too kludgy and weird internally. Let the app judge success or not for itself. 22.6 remove CURLOPT_DNS_USE_GLOBAL_CACHE Remove support for a global DNS cache. Anything global is silly, and we already offer the share interface for the same functionality but done "right". 22.7 remove progress meter from libcurl The internally provided progress meter output doesn't belong in the library. Basically no application wants it (apart from curl) but instead applications can and should do their own progress meters using the progress callback. The progress callback should then be bumped as well to get proper 64bit variable types passed to it instead of doubles so that big files work correctly. 22.8 remove 'curl_httppost' from public curl_formadd() was made to fill in a public struct, but the fact that the struct is public is never really used by application for their own advantage but instead often restricts how the form functions can or can't be modified. Changing them to return a private handle will benefit the implementation and allow us much greater freedoms while still maintaining a solid API and ABI.