1 These are problems known to exist at the time of this release. Feel free to
2 join in and help us correct one or more of these! Also be sure to check the
3 changelog of the current development status, as one or more of these problems
4 may have been fixed since this was written!
6 43. There seems to be a problem when connecting to the Microsoft telnet server.
7 http://curl.haxx.se/bug/view.cgi?id=1720605
9 41. When doing an operation over FTP that requires the ACCT command (but not
10 when logging in), the operation will fail since libcurl doesn't detect this
11 and thus fails to issue the correct command:
12 http://curl.haxx.se/bug/view.cgi?id=1693337
14 40. Mac OS X test failures (Daniel Johnson)
15 http://curl.haxx.se/mail/lib-2007-03/0095.html
17 39. Steffen Rumler's Race Condition in Curl_proxyCONNECT:
18 http://curl.haxx.se/mail/lib-2007-01/0045.html
20 38. Kumar Swamy Bhatt's problem in ftp/ssl "LIST" operation:
21 http://curl.haxx.se/mail/lib-2007-01/0103.html
23 37. Having more than one connection to the same host when doing NTLM
24 authentication (with performs multiple "passes" and authenticates a
25 connection rather than a HTTP request), and particularly when using the
26 multi interface, there's a risk that libcurl will re-use a wrong connection
27 when doing the different passes in the NTLM negotiation and thus fail to
28 negotiate (in seemingly mysterious ways).
30 35. Both SOCKS5 and SOCKS4 proxy connections are done blocking, which is very
31 bad when used with the multi interface.
33 34. The SOCKS4 connection codes don't properly acknowledge (connect) timeouts.
34 Also see #12. According to bug #1556528, even the SOCKS5 connect code does
35 not do it right: http://curl.haxx.se/bug/view.cgi?id=1556528,
37 33. Doing multi-pass HTTP authentication on a non-default port does not work.
38 This happens because the multi-pass code abuses the redirect following code
39 for doing multiple requests, and when we following redirects to an absolute
40 URL we must use the newly specified port and not the one specified in the
41 original URL. A proper fix to this would need to separate the negotiation
42 "redirect" from an actual redirect.
44 32. (At least on Windows) If libcurl is built with c-ares and there's no DNS
45 server configured in the system, the ares_init() call fails and thus
46 curl_easy_init() fails as well. This causes weird effects for people who use
47 numerical IP addresses only.
49 31. "curl-config --libs" will include details set in LDFLAGS when configure is
50 run that might be needed only for building libcurl. Similarly, it might
51 include options that perhaps aren't suitable both for static and dynamic
52 linking. Further, curl-config --cflags suffers from the same effects with
55 30. You need to use -g to the command line tool in order to use RFC2732-style
56 IPv6 numerical addresses in URLs.
58 29. IPv6 URLs with zone ID is not supported.
59 http://www.ietf.org/internet-drafts/draft-fenner-literal-zone-02.txt
60 specifies the use of a plus sign instead of a percent when specifying zone
61 IDs in URLs to get around the problem of percent signs being
62 special. According to the reporter, Firefox deals with the URL _with_ a
63 percent letter (which seems like a blatant URL spec violation).
65 See http://curl.haxx.se/bug/view.cgi?id=1371118
67 26. NTLM authentication using SSPI (on Windows) when (lib)curl is running in
68 "system context" will make it use wrong(?) user name - at least when compared
69 to what winhttp does. See http://curl.haxx.se/bug/view.cgi?id=1281867
71 23. SOCKS-related problems:
72 A) libcurl doesn't support SOCKS for IPv6.
73 B) libcurl doesn't support FTPS over a SOCKS proxy.
74 C) We don't have any test cases for SOCKS proxy.
75 E) libcurl doesn't support active FTP over a SOCKS proxy
77 We probably have even more bugs and lack of features when a SOCKS proxy is
80 22. Sending files to a FTP server using curl on VMS, might lead to curl
81 complaining on "unaligned file size" on completion. The problem is related
82 to VMS file structures and the perceived file sizes stat() returns. A
83 possible fix would involve sending a "STRU VMS" command.
84 http://curl.haxx.se/bug/view.cgi?id=1156287
86 21. FTP ASCII transfers do not follow RFC959. They don't convert the data
87 accordingly (not for sending nor for receiving). RFC 959 section 3.1.1.1
88 clearly describes how this should be done:
90 The sender converts the data from an internal character representation to
91 the standard 8-bit NVT-ASCII representation (see the Telnet
92 specification). The receiver will convert the data from the standard
93 form to his own internal form.
95 Since 7.15.4 at least line endings are converted.
97 16. FTP URLs passed to curl may contain NUL (0x00) in the RFC 1738 <user>,
98 <password>, and <fpath> components, encoded as "%00". The problem is that
99 curl_unescape does not detect this, but instead returns a shortened C
100 string. From a strict FTP protocol standpoint, NUL is a valid character
101 within RFC 959 <string>, so the way to handle this correctly in curl would
102 be to use a data structure other than a plain C string, one that can handle
103 embedded NUL characters. From a practical standpoint, most FTP servers
104 would not meaningfully support NUL characters within RFC 959 <string>,
105 anyway (e.g., UNIX pathnames may not contain NUL).
107 14. Test case 165 might fail on system which has libidn present, but with an
108 old iconv version (2.1.3 is a known bad version), since it doesn't recognize
109 the charset when named ISO8859-1. Changing the name to ISO-8859-1 makes the
110 test pass, but instead makes it fail on Solaris hosts that use its native
113 13. curl version 7.12.2 fails on AIX if compiled with --enable-ares.
114 The workaround is to combine --enable-ares with --disable-shared
116 12. When connecting to a SOCKS proxy, the (connect) timeout is not properly
117 acknowledged after the actual TCP connect (during the SOCKS "negotiate"
120 10. To get HTTP Negotiate authentication to work fine, you need to provide a
121 (fake) user name (this concerns both curl and the lib) because the code
122 wrongly only considers authentication if there's a user name provided.
123 http://curl.haxx.se/bug/view.cgi?id=1004841. How?
124 http://curl.haxx.se/mail/lib-2004-08/0182.html
126 8. Doing resumed upload over HTTP does not work with '-C -', because curl
127 doesn't do a HEAD first to get the initial size. This needs to be done
128 manually for HTTP PUT resume to work, and then '-C [index]'.
130 7. CURLOPT_USERPWD and CURLOPT_PROXYUSERPWD have no way of providing user names
131 that contain a colon. This can't be fixed easily in a backwards compatible
132 way without adding new options (and then, they should most probably allow
133 setting user name and password separately).
135 6. libcurl ignores empty path parts in FTP URLs, whereas RFC1738 states that
136 such parts should be sent to the server as 'CWD ' (without an argument).
137 The only exception to this rule, is that we knowingly break this if the
138 empty part is first in the path, as then we use the double slashes to
139 indicate that the user wants to reach the root dir (this exception SHALL
140 remain even when this bug is fixed).
142 5. libcurl doesn't treat the content-length of compressed data properly, as
143 it seems HTTP servers send the *uncompressed* length in that header and
144 libcurl thinks of it as the *compressed* length. Some explanations are here:
145 http://curl.haxx.se/mail/lib-2003-06/0146.html
147 2. If a HTTP server responds to a HEAD request and includes a body (thus
148 violating the RFC2616), curl won't wait to read the response but just stop
149 reading and return back. If a second request (let's assume a GET) is then
150 immediately made to the same server again, the connection will be re-used
151 fine of course, and the second request will be sent off but when the
152 response is to get read, the previous response-body is what curl will read
153 and havoc is what happens.
154 More details on this is found in this libcurl mailing list thread:
155 http://curl.haxx.se/mail/lib-2002-08/0000.html