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 76. The SOCKET type in Win64 is 64 bits large (and thus so is curl_socket_t on
7 that platform), and long is only 32 bits. It makes it impossible for
8 curl_easy_getinfo() to return a socket properly with the CURLINFO_LASTSOCKET
9 option as for all other operating systems.
11 75. NTLM authentication involving unicode user name or password.
12 http://curl.haxx.se/mail/lib-2009-10/0024.html
13 http://curl.haxx.se/bug/view.cgi?id=2944325
15 73. if a connection is made to a FTP server but the server then just never
16 sends the 220 response or otherwise is dead slow, libcurl will not
17 acknowledge the connection timeout during that phase but only the "real"
18 timeout - which may surprise users as it is probably considered to be the
19 connect phase to most people. Brought up (and is being misunderstood) in:
20 http://curl.haxx.se/bug/view.cgi?id=2844077
22 72. "Pausing pipeline problems."
23 http://curl.haxx.se/mail/lib-2009-07/0214.html
25 70. Problem re-using easy handle after call to curl_multi_remove_handle
26 http://curl.haxx.se/mail/lib-2009-07/0249.html
28 68. "More questions about ares behavior".
29 http://curl.haxx.se/mail/lib-2009-08/0012.html
31 67. When creating multipart formposts. The file name part can be encoded with
32 something beyond ascii but currently libcurl will only pass in the verbatim
33 string the app provides. There are several browsers that already do this
34 encoding. The key seems to be the updated draft to RFC2231:
35 http://tools.ietf.org/html/draft-reschke-rfc2231-in-http-02
37 66. When using telnet, the time limitation options don't work.
38 http://curl.haxx.se/bug/view.cgi?id=2818950
40 65. When doing FTP over a socks proxy or CONNECT through HTTP proxy and the
41 multi interface is used, libcurl will fail if the (passive) TCP connection
42 for the data transfer isn't more or less instant as the code does not
43 properly wait for the connect to be confirmed. See test case 564 for a first
46 63. When CURLOPT_CONNECT_ONLY is used, the handle cannot reliably be re-used
47 for any further requests or transfers. The work-around is then to close that
48 handle with curl_easy_cleanup() and create a new. Some more details:
49 http://curl.haxx.se/mail/lib-2009-04/0300.html
51 61. If an upload using Expect: 100-continue receives an HTTP 417 response,
52 it ought to be automatically resent without the Expect:. A workaround is
53 for the client application to redo the transfer after disabling Expect:.
54 http://curl.haxx.se/mail/archive-2008-02/0043.html
56 60. libcurl closes the connection if an HTTP 401 reply is received while it
57 is waiting for the the 100-continue response.
58 http://curl.haxx.se/mail/lib-2008-08/0462.html
60 58. It seems sensible to be able to use CURLOPT_NOBODY and
61 CURLOPT_FAILONERROR with FTP to detect if a file exists or not, but it is
62 not working: http://curl.haxx.se/mail/lib-2008-07/0295.html
64 57. On VMS-Alpha: When using an http-file-upload the file is not sent to the
65 Server with the correct content-length. Sending a file with 511 or less
66 bytes, content-length 512 is used. Sending a file with 513 - 1023 bytes,
67 content-length 1024 is used. Files with a length of a multiple of 512 Bytes
68 show the correct content-length. Only these files work for upload.
69 http://curl.haxx.se/bug/view.cgi?id=2057858
71 56. When libcurl sends CURLOPT_POSTQUOTE commands when connected to a SFTP
72 server using the multi interface, the commands are not being sent correctly
73 and instead the connection is "cancelled" (the operation is considered done)
74 prematurely. There is a half-baked (busy-looping) patch provided in the bug
75 report but it cannot be accepted as-is. See
76 http://curl.haxx.se/bug/view.cgi?id=2006544
78 55. libcurl fails to build with MIT Kerberos for Windows (KfW) due to KfW's
79 library header files exporting symbols/macros that should be kept private
80 to the KfW library. See ticket #5601 at http://krbdev.mit.edu/rt/
82 52. Gautam Kachroo's issue that identifies a problem with the multi interface
83 where a connection can be re-used without actually being properly
85 http://curl.haxx.se/mail/lib-2008-01/0277.html
87 49. If using --retry and the transfer timeouts (possibly due to using -m or
88 -y/-Y) the next attempt doesn't resume the transfer properly from what was
89 downloaded in the previous attempt but will truncate and restart at the
90 original position where it was at before the previous failed attempt. See
91 http://curl.haxx.se/mail/lib-2008-01/0080.html and Mandriva bug report
92 https://qa.mandriva.com/show_bug.cgi?id=22565
94 48. If a CONNECT response-headers are larger than BUFSIZE (16KB) when the
95 connection is meant to be kept alive (like for NTLM proxy auth), the
96 function will return prematurely and will confuse the rest of the HTTP
97 protocol code. This should be very rare.
99 43. There seems to be a problem when connecting to the Microsoft telnet server.
100 http://curl.haxx.se/bug/view.cgi?id=1720605
102 41. When doing an operation over FTP that requires the ACCT command (but not
103 when logging in), the operation will fail since libcurl doesn't detect this
104 and thus fails to issue the correct command:
105 http://curl.haxx.se/bug/view.cgi?id=1693337
107 39. Steffen Rumler's Race Condition in Curl_proxyCONNECT:
108 http://curl.haxx.se/mail/lib-2007-01/0045.html
110 38. Kumar Swamy Bhatt's problem in ftp/ssl "LIST" operation:
111 http://curl.haxx.se/mail/lib-2007-01/0103.html
113 37. Having more than one connection to the same host when doing NTLM
114 authentication (with performs multiple "passes" and authenticates a
115 connection rather than a HTTP request), and particularly when using the
116 multi interface, there's a risk that libcurl will re-use a wrong connection
117 when doing the different passes in the NTLM negotiation and thus fail to
118 negotiate (in seemingly mysterious ways).
120 35. Both SOCKS5 and SOCKS4 proxy connections are done blocking, which is very
121 bad when used with the multi interface.
123 34. The SOCKS4 connection codes don't properly acknowledge (connect) timeouts.
124 Also see #12. According to bug #1556528, even the SOCKS5 connect code does
125 not do it right: http://curl.haxx.se/bug/view.cgi?id=1556528,
127 31. "curl-config --libs" will include details set in LDFLAGS when configure is
128 run that might be needed only for building libcurl. Further, curl-config
129 --cflags suffers from the same effects with CFLAGS/CPPFLAGS.
131 30. You need to use -g to the command line tool in order to use RFC2732-style
132 IPv6 numerical addresses in URLs.
134 29. IPv6 URLs with zone ID is not nicely supported.
135 http://www.ietf.org/internet-drafts/draft-fenner-literal-zone-02.txt (expired)
136 specifies the use of a plus sign instead of a percent when specifying zone
137 IDs in URLs to get around the problem of percent signs being
138 special. According to the reporter, Firefox deals with the URL _with_ a
139 percent letter (which seems like a blatant URL spec violation).
140 libcurl supports zone IDs where the percent sign is URL-escaped (i.e. %25).
142 See http://curl.haxx.se/bug/view.cgi?id=1371118
144 26. NTLM authentication using SSPI (on Windows) when (lib)curl is running in
145 "system context" will make it use wrong(?) user name - at least when compared
146 to what winhttp does. See http://curl.haxx.se/bug/view.cgi?id=1281867
148 23. SOCKS-related problems:
149 A) libcurl doesn't support SOCKS for IPv6.
150 B) libcurl doesn't support FTPS over a SOCKS proxy.
151 E) libcurl doesn't support active FTP over a SOCKS proxy
153 We probably have even more bugs and lack of features when a SOCKS proxy is
156 22. Sending files to a FTP server using curl on VMS, might lead to curl
157 complaining on "unaligned file size" on completion. The problem is related
158 to VMS file structures and the perceived file sizes stat() returns. A
159 possible fix would involve sending a "STRU VMS" command.
160 http://curl.haxx.se/bug/view.cgi?id=1156287
162 21. FTP ASCII transfers do not follow RFC959. They don't convert the data
163 accordingly (not for sending nor for receiving). RFC 959 section 3.1.1.1
164 clearly describes how this should be done:
166 The sender converts the data from an internal character representation to
167 the standard 8-bit NVT-ASCII representation (see the Telnet
168 specification). The receiver will convert the data from the standard
169 form to his own internal form.
171 Since 7.15.4 at least line endings are converted.
173 16. FTP URLs passed to curl may contain NUL (0x00) in the RFC 1738 <user>,
174 <password>, and <fpath> components, encoded as "%00". The problem is that
175 curl_unescape does not detect this, but instead returns a shortened C
176 string. From a strict FTP protocol standpoint, NUL is a valid character
177 within RFC 959 <string>, so the way to handle this correctly in curl would
178 be to use a data structure other than a plain C string, one that can handle
179 embedded NUL characters. From a practical standpoint, most FTP servers
180 would not meaningfully support NUL characters within RFC 959 <string>,
181 anyway (e.g., UNIX pathnames may not contain NUL).
183 14. Test case 165 might fail on a system which has libidn present, but with an
184 old iconv version (2.1.3 is a known bad version), since it doesn't recognize
185 the charset when named ISO8859-1. Changing the name to ISO-8859-1 makes the
186 test pass, but instead makes it fail on Solaris hosts that use its native
189 13. curl version 7.12.2 fails on AIX if compiled with --enable-ares.
190 The workaround is to combine --enable-ares with --disable-shared
192 12. When connecting to a SOCKS proxy, the (connect) timeout is not properly
193 acknowledged after the actual TCP connect (during the SOCKS "negotiate"
196 10. To get HTTP Negotiate authentication to work fine, you need to provide a
197 (fake) user name (this concerns both curl and the lib) because the code
198 wrongly only considers authentication if there's a user name provided.
199 http://curl.haxx.se/bug/view.cgi?id=1004841. How?
200 http://curl.haxx.se/mail/lib-2004-08/0182.html
202 8. Doing resumed upload over HTTP does not work with '-C -', because curl
203 doesn't do a HEAD first to get the initial size. This needs to be done
204 manually for HTTP PUT resume to work, and then '-C [index]'.
206 6. libcurl ignores empty path parts in FTP URLs, whereas RFC1738 states that
207 such parts should be sent to the server as 'CWD ' (without an argument).
208 The only exception to this rule, is that we knowingly break this if the
209 empty part is first in the path, as then we use the double slashes to
210 indicate that the user wants to reach the root dir (this exception SHALL
211 remain even when this bug is fixed).
213 5. libcurl doesn't treat the content-length of compressed data properly, as
214 it seems HTTP servers send the *uncompressed* length in that header and
215 libcurl thinks of it as the *compressed* length. Some explanations are here:
216 http://curl.haxx.se/mail/lib-2003-06/0146.html
218 2. If a HTTP server responds to a HEAD request and includes a body (thus
219 violating the RFC2616), curl won't wait to read the response but just stop
220 reading and return back. If a second request (let's assume a GET) is then
221 immediately made to the same server again, the connection will be re-used
222 fine of course, and the second request will be sent off but when the
223 response is to get read, the previous response-body is what curl will read
224 and havoc is what happens.
225 More details on this is found in this libcurl mailing list thread:
226 http://curl.haxx.se/mail/lib-2002-08/0000.html