1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
|
_ _ ____ _
___| | | | _ \| |
/ __| | | | |_) | |
| (__| |_| | _ <| |___
\___|\___/|_| \_\_____|
TODO
Things to do in project cURL. Please tell us what you think, contribute and
send us patches that improve things! Also check the http://curl.haxx.se/dev
web section for various technical development notes.
LIBCURL
* Introduce an interface to libcurl that allows applications to easier get to
know what cookies that are received. Pushing interface that calls a
callback on each received cookie? Querying interface that asks about
existing cookies? We probably need both. Enable applications to modify
existing cookies as well.
* Make content encoding/decoding internally be made using a filter system.
* Introduce another callback interface for upload/download that makes one
less copy of data and thus a faster operation.
[http://curl.haxx.se/dev/no_copy_callbacks.txt]
* Add asynchronous name resolving (http://libdenise.sf.net/). This should be
made to work on most of the supported platforms, or otherwise it isn't
really interesting.
* Data sharing. Tell which easy handles within a multi handle that should
share cookies, connection cache, dns cache, ssl session cache. Full
suggestion found here: http://curl.haxx.se/dev/sharing.txt
* Mutexes. By adding mutex callback support, the 'data sharing' mentioned
above can be made between several easy handles running in different threads
too. The actual mutex implementations will be left for the application to
implement, libcurl will merely call 'getmutex' and 'leavemutex' callbacks.
Part of the sharing suggestion at: http://curl.haxx.se/dev/sharing.txt
* Set the SO_KEEPALIVE socket option to make libcurl notice and disconnect
very long time idle connections.
* Go through the code and verify that libcurl deals with big files >2GB and
>4GB all over. Bug reports (and source reviews) indicate that it doesn't
currently work properly.
* CURLOPT_MAXFILESIZE. Prevent downloads that are larger than the specified
size. CURLE_FILESIZE_EXCEEDED would then be returned. Gautam Mani
requested. That is, the download should not even begin but be aborted
immediately.
* Allow the http_proxy (and other) environment variables to contain user and
password as well in the style: http://proxyuser:proxypasswd@proxy:port
Berend Reitsma suggested.
LIBCURL - multi interface
* Make sure we don't ever loop because of non-blocking sockets return
EWOULDBLOCK or similar. This FTP command sending, the SSL connection etc.
* Make transfers treated more carefully. We need a way to tell libcurl we
have data to write, as the current system expects us to upload data each
time the socket is writable and there is no way to say that we want to
upload data soon just not right now, without that aborting the upload. The
opposite situation should be possible as well, that we tell libcurl we're
ready to accept read data. Today libcurl feeds the data as soon as it is
available for reading, no matter what.
DOCUMENTATION
* More and better
FTP
* FTP ASCII upload does not follow RFC959 section 3.1.1.1: "The sender
converts the data from an internal character representation to the standard
8-bit NVT-ASCII representation (see the Telnet specification). The
receiver will convert the data from the standard form to his own internal
form."
* Since USERPWD always override the user and password specified in URLs, we
might need another way to specify user+password for anonymous ftp logins.
* An option to only download remote FTP files if they're newer than the local
one is a good idea, and it would fit right into the same syntax as the
already working http dito works (-z). It of course requires that 'MDTM'
works, and it isn't a standard FTP command.
* Add FTPS support with SSL for the data connection too. This should be made
according to the specs written in draft-murray-auth-ftp-ssl-08.txt,
"Securing FTP with TLS"
* --disable-epsv exists, but for active connections we have no --disable-eprt
(or even --disable-lprt).
HTTP
* If the "body" of the POST is < MSS it really aught to be sent along with
the headers. More generally, if the last chunk of the POST body is < MSS,
it should be sent with the previous chunk (which may be the POST headers).
So long as any one send is larger than MSS (or there is only one send when
< MSS :), the Nagle Algorithm will not be a problem on any stack where
Nagle is implemented correctly. (pointed out by Rick Jones)
* Authentication: NTLM. Support for that MS crap called NTLM
authentication. MS proxies and servers sometime require that. Since that
protocol is a proprietary one, it involves reverse engineering and network
sniffing. This should however be a library-based functionality. There are a
few different efforts "out there" to make open source HTTP clients support
this and it should be possible to take advantage of other people's hard
work. http://modntlm.sourceforge.net/ is one. There's a web page at
http://www.innovation.ch/java/ntlm.html that contains detailed reverse-
engineered info.
* RFC2617 compliance, "Digest Access Authentication" A valid test page seem
to exist at: http://hopf.math.nwu.edu/testpage/digest/ And some friendly
person's server source code is available at
http://hopf.math.nwu.edu/digestauth/index.html Then there's the Apache
mod_digest source code too of course. It seems as if Netscape doesn't
support this, and not many servers do. Although this is a lot better
authentication method than the more common "Basic". Basic sends the
password in cleartext over the network, this "Digest" method uses a
challange-response protocol which increases security quite a lot.
* Pipelining. Sending multiple requests before the previous one(s) are done.
This could possibly be implemented using the multi interface to queue
requests and the response data.
TELNET
* Make TELNET work on windows98!
* 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.
* 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.
SSL
* If you really want to improve the SSL situation, you should probably have a
look at SSL cafile loading as well - 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). Even better would be to support the SSL
CAdir option - instead of loading all of the root CA certs for every
request, this option allows you to only read the CA chain that is actually
required (into the cache)...
* 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". This whole
idea might become moot if we enable the 'data sharing' as mentioned in the
LIBCURL label above.
* 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! (brought by Chris
Clark)
* Make curl's SSL layer option capable of using other free SSL libraries.
Such as the Mozilla Security Services
(http://www.mozilla.org/projects/security/pki/nss/) and GNUTLS
(http://gnutls.hellug.gr/)
LDAP
* Look over the implementation. The looping will have to "go away" from the
lib/ldap.c source file and get moved to the main network code so that the
multi interface and friends will work for LDAP as well.
CLIENT
* Add an option that prevents cURL from overwiting 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. Jeff Pohlmeyer suggested.
* "curl ftp://site.com/*.txt"
* Several URLs can be specified to get downloaded. We should be able to use
the same syntax to specify several files to get uploaded (using the same
persistant connection), using -T.
* When the multi interface has been implemented and proved to work, the
client could be told to use maximum N simultaneous transfers and then just
make sure that happens. It should of course not make more than one
connection to the same remote host.
* Extending the capabilities of the multipart formposting. How about leaving
the ';type=foo' syntax as it is and adding an extra tag (headers) which
works like this: curl -F "coolfiles=@fil1.txt;headers=@fil1.hdr" where
fil1.hdr contains extra headers like
Content-Type: text/plain; charset=KOI8-R"
Content-Transfer-Encoding: base64
X-User-Comment: Please don't use browser specific HTML code
which should overwrite the program reasonable defaults (plain/text,
8bit...) (Idea brough to us by kromJx)
TEST SUITE
* If perl wasn't found by the configure script, don't attempt to run the
tests but explain something nice why it doesn't.
* Extend the test suite to include more protocols. The telnet could just do
ftp or http operations (for which we have test servers).
* Make the test suite work on more platforms. OpenBSD and Mac OS. Remove
fork()s and it should become even more portable.
* Introduce a test suite that tests libcurl better and more explicitly.
NEXT MAJOR RELEASE
* curl_easy_cleanup() returns void, but curl_multi_cleanup() returns a
CURLMcode. These should be changed to be the same.
* curl_formparse() should be removed
|
