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 87. -J/--remote-header-name doesn't decode %-encoded file names. RFC6266
7 details how it should be done. The can of worm is basically that we have no
8 charset handling in curl and ascii >=128 is a challenge for us. Not to
9 mention that decoding also means that we need to check for nastiness that is
10 attempted, like "../" sequences and the like. Probably everything to the left
11 of any embedded slashes should be cut off.
12 http://curl.haxx.se/bug/view.cgi?id=1294
14 86. The disconnect commands (LOGOUT and QUIT) may not be sent by IMAP, POP3
15 and SMTP if a failure occurs during the authentication phase of a
18 85. Wrong STARTTRANSFER timer accounting for POST requests
19 Timer works fine with GET requests, but while using POST the time for
20 CURLINFO_STARTTRANSFER_TIME is wrong. While using POST
21 CURLINFO_STARTTRANSFER_TIME minus CURLINFO_PRETRANSFER_TIME is near to zero
23 http://curl.haxx.se/bug/view.cgi?id=1213
25 84. CURLINFO_SSL_VERIFYRESULT is only implemented for the OpenSSL and NSS
26 backends, so relying on this information in a generic app is flaky.
28 82. When building with the Windows Borland compiler, it fails because the
29 "tlib" tool doesn't support hyphens (minus signs) in file names and we have
31 http://curl.haxx.se/bug/view.cgi?id=1222
33 81. When using -J (with -O), automatically resumed downloading together with
34 "-C -" fails. Without -J the same command line works! This happens because
35 the resume logic is worked out before the target file name (and thus its
36 pre-transfer size) has been figured out!
37 http://curl.haxx.se/bug/view.cgi?id=1169
39 80. Curl doesn't recognize certificates in DER format in keychain, but it
41 http://curl.haxx.se/bug/view.cgi?id=1065
43 79. SMTP. When sending data to multiple recipients, curl will abort and return
44 failure if one of the recipients indicate failure (on the "RCPT TO"
45 command). Ordinary mail programs would proceed and still send to the ones
46 that can receive data. This is subject for change in the future.
47 http://curl.haxx.se/bug/view.cgi?id=1116
49 78. curl and libcurl don't always signal the client properly when "sending"
50 zero bytes files - it makes for example the command line client not creating
51 any file at all. Like when using FTP.
52 http://curl.haxx.se/bug/view.cgi?id=1063
54 77. CURLOPT_FORBID_REUSE on a handle prevents NTLM from working since it
55 "abuses" the underlying connection re-use system and if connections are
56 forced to close they break the NTLM support.
58 76. The SOCKET type in Win64 is 64 bits large (and thus so is curl_socket_t on
59 that platform), and long is only 32 bits. It makes it impossible for
60 curl_easy_getinfo() to return a socket properly with the CURLINFO_LASTSOCKET
61 option as for all other operating systems.
63 75. NTLM authentication involving unicode user name or password only works
64 properly if built with UNICODE defined together with the WinSSL/schannel
65 backend. The original problem was mentioned in:
66 http://curl.haxx.se/mail/lib-2009-10/0024.html
67 http://curl.haxx.se/bug/view.cgi?id=896
69 The WinSSL/schannel version verified to work as mentioned in
70 http://curl.haxx.se/mail/lib-2012-07/0073.html
72 73. if a connection is made to a FTP server but the server then just never
73 sends the 220 response or otherwise is dead slow, libcurl will not
74 acknowledge the connection timeout during that phase but only the "real"
75 timeout - which may surprise users as it is probably considered to be the
76 connect phase to most people. Brought up (and is being misunderstood) in:
77 http://curl.haxx.se/bug/view.cgi?id=856
79 72. "Pausing pipeline problems."
80 http://curl.haxx.se/mail/lib-2009-07/0214.html
82 70. Problem re-using easy handle after call to curl_multi_remove_handle
83 http://curl.haxx.se/mail/lib-2009-07/0249.html
85 68. "More questions about ares behavior".
86 http://curl.haxx.se/mail/lib-2009-08/0012.html
88 67. When creating multipart formposts. The file name part can be encoded with
89 something beyond ascii but currently libcurl will only pass in the verbatim
90 string the app provides. There are several browsers that already do this
91 encoding. The key seems to be the updated draft to RFC2231:
92 http://tools.ietf.org/html/draft-reschke-rfc2231-in-http-02
94 66. When using telnet, the time limitation options don't work.
95 http://curl.haxx.se/bug/view.cgi?id=846
97 65. When doing FTP over a socks proxy or CONNECT through HTTP proxy and the
98 multi interface is used, libcurl will fail if the (passive) TCP connection
99 for the data transfer isn't more or less instant as the code does not
100 properly wait for the connect to be confirmed. See test case 564 for a first
103 63. When CURLOPT_CONNECT_ONLY is used, the handle cannot reliably be re-used
104 for any further requests or transfers. The work-around is then to close that
105 handle with curl_easy_cleanup() and create a new. Some more details:
106 http://curl.haxx.se/mail/lib-2009-04/0300.html
108 61. If an upload using Expect: 100-continue receives an HTTP 417 response,
109 it ought to be automatically resent without the Expect:. A workaround is
110 for the client application to redo the transfer after disabling Expect:.
111 http://curl.haxx.se/mail/archive-2008-02/0043.html
113 60. libcurl closes the connection if an HTTP 401 reply is received while it
114 is waiting for the the 100-continue response.
115 http://curl.haxx.se/mail/lib-2008-08/0462.html
117 58. It seems sensible to be able to use CURLOPT_NOBODY and
118 CURLOPT_FAILONERROR with FTP to detect if a file exists or not, but it is
119 not working: http://curl.haxx.se/mail/lib-2008-07/0295.html
121 56. When libcurl sends CURLOPT_POSTQUOTE commands when connected to a SFTP
122 server using the multi interface, the commands are not being sent correctly
123 and instead the connection is "cancelled" (the operation is considered done)
124 prematurely. There is a half-baked (busy-looping) patch provided in the bug
125 report but it cannot be accepted as-is. See
126 http://curl.haxx.se/bug/view.cgi?id=748
128 55. libcurl fails to build with MIT Kerberos for Windows (KfW) due to KfW's
129 library header files exporting symbols/macros that should be kept private
130 to the KfW library. See ticket #5601 at http://krbdev.mit.edu/rt/
132 52. Gautam Kachroo's issue that identifies a problem with the multi interface
133 where a connection can be re-used without actually being properly
135 http://curl.haxx.se/mail/lib-2008-01/0277.html
137 49. If using --retry and the transfer timeouts (possibly due to using -m or
138 -y/-Y) the next attempt doesn't resume the transfer properly from what was
139 downloaded in the previous attempt but will truncate and restart at the
140 original position where it was at before the previous failed attempt. See
141 http://curl.haxx.se/mail/lib-2008-01/0080.html and Mandriva bug report
142 https://qa.mandriva.com/show_bug.cgi?id=22565
144 48. If a CONNECT response-headers are larger than BUFSIZE (16KB) when the
145 connection is meant to be kept alive (like for NTLM proxy auth), the
146 function will return prematurely and will confuse the rest of the HTTP
147 protocol code. This should be very rare.
149 43. There seems to be a problem when connecting to the Microsoft telnet server.
150 http://curl.haxx.se/bug/view.cgi?id=649
152 41. When doing an operation over FTP that requires the ACCT command (but not
153 when logging in), the operation will fail since libcurl doesn't detect this
154 and thus fails to issue the correct command:
155 http://curl.haxx.se/bug/view.cgi?id=635
157 39. Steffen Rumler's Race Condition in Curl_proxyCONNECT:
158 http://curl.haxx.se/mail/lib-2007-01/0045.html
160 38. Kumar Swamy Bhatt's problem in ftp/ssl "LIST" operation:
161 http://curl.haxx.se/mail/lib-2007-01/0103.html
163 35. Both SOCKS5 and SOCKS4 proxy connections are done blocking, which is very
164 bad when used with the multi interface.
166 34. The SOCKS4 connection codes don't properly acknowledge (connect) timeouts.
167 Also see #12. According to bug #1556528, even the SOCKS5 connect code does
168 not do it right: http://curl.haxx.se/bug/view.cgi?id=604
170 31. "curl-config --libs" will include details set in LDFLAGS when configure is
171 run that might be needed only for building libcurl. Further, curl-config
172 --cflags suffers from the same effects with CFLAGS/CPPFLAGS.
174 26. NTLM authentication using SSPI (on Windows) when (lib)curl is running in
175 "system context" will make it use wrong(?) user name - at least when compared
176 to what winhttp does. See http://curl.haxx.se/bug/view.cgi?id=535
178 23. SOCKS-related problems:
179 B) libcurl doesn't support FTPS over a SOCKS proxy.
180 E) libcurl doesn't support active FTP over a SOCKS proxy
182 We probably have even more bugs and lack of features when a SOCKS proxy is
185 21. FTP ASCII transfers do not follow RFC959. They don't convert the data
186 accordingly (not for sending nor for receiving). RFC 959 section 3.1.1.1
187 clearly describes how this should be done:
189 The sender converts the data from an internal character representation to
190 the standard 8-bit NVT-ASCII representation (see the Telnet
191 specification). The receiver will convert the data from the standard
192 form to his own internal form.
194 Since 7.15.4 at least line endings are converted.
196 16. FTP URLs passed to curl may contain NUL (0x00) in the RFC 1738 <user>,
197 <password>, and <fpath> components, encoded as "%00". The problem is that
198 curl_unescape does not detect this, but instead returns a shortened C
199 string. From a strict FTP protocol standpoint, NUL is a valid character
200 within RFC 959 <string>, so the way to handle this correctly in curl would
201 be to use a data structure other than a plain C string, one that can handle
202 embedded NUL characters. From a practical standpoint, most FTP servers
203 would not meaningfully support NUL characters within RFC 959 <string>,
204 anyway (e.g., UNIX pathnames may not contain NUL).
206 14. Test case 165 might fail on a system which has libidn present, but with an
207 old iconv version (2.1.3 is a known bad version), since it doesn't recognize
208 the charset when named ISO8859-1. Changing the name to ISO-8859-1 makes the
209 test pass, but instead makes it fail on Solaris hosts that use its native
212 13. curl version 7.12.2 fails on AIX if compiled with --enable-ares.
213 The workaround is to combine --enable-ares with --disable-shared
215 12. When connecting to a SOCKS proxy, the (connect) timeout is not properly
216 acknowledged after the actual TCP connect (during the SOCKS "negotiate"
219 10. To get HTTP Negotiate authentication to work fine, you need to provide a
220 (fake) user name (this concerns both curl and the lib) because the code
221 wrongly only considers authentication if there's a user name provided.
222 http://curl.haxx.se/bug/view.cgi?id=440 How?
223 http://curl.haxx.se/mail/lib-2004-08/0182.html
225 8. Doing resumed upload over HTTP does not work with '-C -', because curl
226 doesn't do a HEAD first to get the initial size. This needs to be done
227 manually for HTTP PUT resume to work, and then '-C [index]'.
229 6. libcurl ignores empty path parts in FTP URLs, whereas RFC1738 states that
230 such parts should be sent to the server as 'CWD ' (without an argument).
231 The only exception to this rule, is that we knowingly break this if the
232 empty part is first in the path, as then we use the double slashes to
233 indicate that the user wants to reach the root dir (this exception SHALL
234 remain even when this bug is fixed).
236 5. libcurl doesn't treat the content-length of compressed data properly, as
237 it seems HTTP servers send the *uncompressed* length in that header and
238 libcurl thinks of it as the *compressed* length. Some explanations are here:
239 http://curl.haxx.se/mail/lib-2003-06/0146.html
241 2. If a HTTP server responds to a HEAD request and includes a body (thus
242 violating the RFC2616), curl won't wait to read the response but just stop
243 reading and return back. If a second request (let's assume a GET) is then
244 immediately made to the same server again, the connection will be re-used
245 fine of course, and the second request will be sent off but when the
246 response is to get read, the previous response-body is what curl will read
247 and havoc is what happens.
248 More details on this is found in this libcurl mailing list thread:
249 http://curl.haxx.se/mail/lib-2002-08/0000.html