5 \___|\___/|_| \_\_____|
9 Things to do in project cURL. Please tell us what you think, contribute and
10 send us patches that improve things! Also check the http://curl.haxx.se/dev
11 web section for various technical development notes.
13 All bugs documented in the KNOWN_BUGS document are subject for fixing!
17 * Introduce an interface to libcurl that allows applications to easier get to
18 know what cookies that are received. Pushing interface that calls a
19 callback on each received cookie? Querying interface that asks about
20 existing cookies? We probably need both. Enable applications to modify
21 existing cookies as well. http://curl.haxx.se/dev/COOKIES
23 * Introduce another callback interface for upload/download that makes one
24 less copy of data and thus a faster operation.
25 [http://curl.haxx.se/dev/no_copy_callbacks.txt]
27 * More data sharing. curl_share_* functions already exist and work, and they
28 can be extended to share more. For example, enable sharing of the ares
31 * Introduce a new error code indicating authentication problems (for proxy
32 CONNECT error 407 for example). This cannot be an error code, we must not
33 return informational stuff as errors, consider a new info returned by
34 curl_easy_getinfo() #845941
36 * Option to set the SO_KEEPALIVE socket option to make libcurl notice and
37 disconnect very long time idle connections.
39 LIBCURL - multi interface
41 * Add curl_multi_timeout() to make libcurl's ares-functionality better.
43 * Make sure we don't ever loop because of non-blocking sockets return
44 EWOULDBLOCK or similar. This FTP command sending, the SSL connection etc.
46 * Make transfers treated more carefully. We need a way to tell libcurl we
47 have data to write, as the current system expects us to upload data each
48 time the socket is writable and there is no way to say that we want to
49 upload data soon just not right now, without that aborting the upload. The
50 opposite situation should be possible as well, that we tell libcurl we're
51 ready to accept read data. Today libcurl feeds the data as soon as it is
52 available for reading, no matter what.
60 * Support the most common FTP proxies, Philip Newton provided a list
62 http://curl.haxx.se/mail/archive-2003-04/0126.html
64 * Make CURLOPT_FTPPORT support an additional port number on the IP/if/name,
65 like "blabla:[port]" or possibly even "blabla:[portfirst]-[portsecond]".
67 * FTP ASCII transfers do not follow RFC959. They don't convert the data
70 * Since USERPWD always override the user and password specified in URLs, we
71 might need another way to specify user+password for anonymous ftp logins.
75 * Digest and GSS-Negotiate support for HTTP proxies. They only work on
76 direct-connections to the server.
78 * Pipelining. Sending multiple requests before the previous one(s) are done.
79 This could possibly be implemented using the multi interface to queue
80 requests and the response data.
84 * Reading input (to send to the remote server) on stdin is a crappy solution
85 for library purposes. We need to invent a good way for the application to
86 be able to provide the data to send.
88 * Move the telnet support's network select() loop go away and merge the code
89 into the main transfer loop. Until this is done, the multi interface won't
94 * If you really want to improve the SSL situation, you should probably have a
95 look at SSL cafile loading as well - quick traces look to me like these are
96 done on every request as well, when they should only be necessary once per
97 ssl context (or once per handle). Even better would be to support the SSL
98 CAdir option - instead of loading all of the root CA certs for every
99 request, this option allows you to only read the CA chain that is actually
100 required (into the cache)...
102 * Add an interface to libcurl that enables "session IDs" to get
103 exported/imported. Cris Bailiff said: "OpenSSL has functions which can
104 serialise the current SSL state to a buffer of your choice, and
105 recover/reset the state from such a buffer at a later date - this is used
106 by mod_ssl for apache to implement and SSL session ID cache". This whole
107 idea might become moot if we enable the 'data sharing' as mentioned in the
110 * OpenSSL supports a callback for customised verification of the peer
111 certificate, but this doesn't seem to be exposed in the libcurl APIs. Could
112 it be? There's so much that could be done if it were! (brought by Chris
115 * Make curl's SSL layer option capable of using other free SSL libraries.
116 Such as the Mozilla Security Services
117 (http://www.mozilla.org/projects/security/pki/nss/) and GNUTLS
118 (http://gnutls.hellug.gr/)
122 * Look over the implementation. The looping will have to "go away" from the
123 lib/ldap.c source file and get moved to the main network code so that the
124 multi interface and friends will work for LDAP as well.
128 * Add an option that prevents cURL from overwiting existing local files. When
129 used, and there already is an existing file with the target file name
130 (either -O or -o), a number should be appended (and increased if already
131 existing). So that index.html becomes first index.html.1 and then
132 index.html.2 etc. Jeff Pohlmeyer suggested.
134 * "curl ftp://site.com/*.txt"
136 * The client could be told to use maximum N simultaneous transfers and then
137 just make sure that happens. It should of course not make more than one
138 connection to the same remote host. This would require the client to use
141 * Extending the capabilities of the multipart formposting. How about leaving
142 the ';type=foo' syntax as it is and adding an extra tag (headers) which
143 works like this: curl -F "coolfiles=@fil1.txt;headers=@fil1.hdr" where
144 fil1.hdr contains extra headers like
146 Content-Type: text/plain; charset=KOI8-R"
147 Content-Transfer-Encoding: base64
148 X-User-Comment: Please don't use browser specific HTML code
150 which should overwrite the program reasonable defaults (plain/text,
151 8bit...) (Idea brough to us by kromJx)
153 * ability to specify the classic computing suffixes on the range
154 specifications. For example, to download the first 500 Kilobytes of a file,
155 be able to specify the following for the -r option: "-r 0-500K" or for the
156 first 2 Megabytes of a file: "-r 0-2M". (Mark Smith suggested)
158 * --data-encode that URL encodes the data before posting
159 http://curl.haxx.se/mail/archive-2003-11/0091.html (Kevin Roth suggested)
163 * Consider extending 'roffit' to produce decent ASCII output, and use that
164 instead of (g)nroff when building src/hugehelp.c
168 * Make the test servers able to serve multiple running test suites. Like if
169 two users run 'make test' at once.
171 * Make runtests.pl capable of changing port numbers for the servers. This was
172 the intention from the start, but in practise it is now hard.
174 * If perl wasn't found by the configure script, don't attempt to run the
175 tests but explain something nice why it doesn't.
177 * Extend the test suite to include more protocols. The telnet could just do
178 ftp or http operations (for which we have test servers).
180 * Make the test suite work on more platforms. OpenBSD and Mac OS. Remove
181 fork()s and it should become even more portable.
185 * curl_easy_cleanup() returns void, but curl_multi_cleanup() returns a
186 CURLMcode. These should be changed to be the same.
188 * curl_formparse() should be removed
190 * remove obsolete defines from curl/curl.h
192 * remove the following functions from the public API:
194 curl_mprintf (and variations)
198 They will instead become curlx_ - alternatives. That makes the curl app
199 still capable of building with them from source.