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 64. The threaded resolver used in libcurl on Windows has some kind of race
7 problem when multiple simultanoes resolves are done, like with the multi
8 interface transferring many files in parallell:
9 http://curl.haxx.se/mail/lib-2009-04/0028.html
11 63. When CURLOPT_CONNECT_ONLY is used, the handle cannot reliably be re-used
12 for any further requests or transfers. The work-around is then to close that
13 handle with curl_easy_cleanup() and create a new. Some more details:
14 http://curl.haxx.se/mail/lib-2009-04/0300.html
16 62. CURLOPT_TIMEOUT does not work properly with the regular multi and
17 multi_socket interfaces. The work-around for apps is to simply remove the
18 easy handle once the time is up. See also:
19 http://curl.haxx.se/bug/view.cgi?id=2501457
21 61. If an upload using Expect: 100-continue receives an HTTP 417 response,
22 it ought to be automatically resent without the Expect:. A workaround is
23 for the client application to redo the transfer after disabling Expect:.
24 http://curl.haxx.se/mail/archive-2008-02/0043.html
26 60. libcurl closes the connection if an HTTP 401 reply is received while it
27 is waiting for the the 100-continue response.
28 http://curl.haxx.se/mail/lib-2008-08/0462.html
30 59. If the CURLOPT_PORT option is used on an FTP URL like
31 "ftp://example.com/file;type=A" the ";type=A" is stripped off.
33 58. It seems sensible to be able to use CURLOPT_NOBODY and
34 CURLOPT_FAILONERROR with FTP to detect if a file exists or not, but it is
35 not working: http://curl.haxx.se/mail/lib-2008-07/0295.html
37 57. On VMS-Alpha: When using an http-file-upload the file is not sent to the
38 Server with the correct content-length. Sending a file with 511 or less
39 bytes, content-length 512 is used. Sending a file with 513 - 1023 bytes,
40 content-length 1024 is used. Files with a length of a multiple of 512 Bytes
41 show the correct content-length. Only these files work for upload.
42 http://curl.haxx.se/bug/view.cgi?id=2057858
44 56. When libcurl sends CURLOPT_POSTQUOTE commands when connected to a SFTP
45 server using the multi interface, the commands are not being sent correctly
46 and instead the connection is "cancelled" (the operation is considered done)
47 prematurely. There is a half-baked (busy-looping) patch provided in the bug
48 report but it cannot be accepted as-is. See
49 http://curl.haxx.se/bug/view.cgi?id=2006544
51 55. libcurl fails to build with MIT Kerberos for Windows (KfW) due to KfW's
52 library header files exporting symbols/macros that should be kept private
53 to the KfW library. See ticket #5601 at http://krbdev.mit.edu/rt/
55 52. Gautam Kachroo's issue that identifies a problem with the multi interface
56 where a connection can be re-used without actually being properly
58 http://curl.haxx.se/mail/lib-2008-01/0277.html
60 49. If using --retry and the transfer timeouts (possibly due to using -m or
61 -y/-Y) the next attempt doesn't resume the transfer properly from what was
62 downloaded in the previous attempt but will truncate and restart at the
63 original position where it was at before the previous failed attempt. See
64 http://curl.haxx.se/mail/lib-2008-01/0080.html and Mandriva bug report
65 https://qa.mandriva.com/show_bug.cgi?id=22565
67 48. If a CONNECT response-headers are larger than BUFSIZE (16KB) when the
68 connection is meant to be kept alive (like for NTLM proxy auth), the
69 function will return prematurely and will confuse the rest of the HTTP
70 protocol code. This should be very rare.
72 43. There seems to be a problem when connecting to the Microsoft telnet server.
73 http://curl.haxx.se/bug/view.cgi?id=1720605
75 41. When doing an operation over FTP that requires the ACCT command (but not
76 when logging in), the operation will fail since libcurl doesn't detect this
77 and thus fails to issue the correct command:
78 http://curl.haxx.se/bug/view.cgi?id=1693337
80 39. Steffen Rumler's Race Condition in Curl_proxyCONNECT:
81 http://curl.haxx.se/mail/lib-2007-01/0045.html
83 38. Kumar Swamy Bhatt's problem in ftp/ssl "LIST" operation:
84 http://curl.haxx.se/mail/lib-2007-01/0103.html
86 37. Having more than one connection to the same host when doing NTLM
87 authentication (with performs multiple "passes" and authenticates a
88 connection rather than a HTTP request), and particularly when using the
89 multi interface, there's a risk that libcurl will re-use a wrong connection
90 when doing the different passes in the NTLM negotiation and thus fail to
91 negotiate (in seemingly mysterious ways).
93 35. Both SOCKS5 and SOCKS4 proxy connections are done blocking, which is very
94 bad when used with the multi interface.
96 34. The SOCKS4 connection codes don't properly acknowledge (connect) timeouts.
97 Also see #12. According to bug #1556528, even the SOCKS5 connect code does
98 not do it right: http://curl.haxx.se/bug/view.cgi?id=1556528,
100 31. "curl-config --libs" will include details set in LDFLAGS when configure is
101 run that might be needed only for building libcurl. Further, curl-config
102 --cflags suffers from the same effects with CFLAGS/CPPFLAGS.
104 30. You need to use -g to the command line tool in order to use RFC2732-style
105 IPv6 numerical addresses in URLs.
107 29. IPv6 URLs with zone ID is not supported.
108 http://www.ietf.org/internet-drafts/draft-fenner-literal-zone-02.txt (expired)
109 specifies the use of a plus sign instead of a percent when specifying zone
110 IDs in URLs to get around the problem of percent signs being
111 special. According to the reporter, Firefox deals with the URL _with_ a
112 percent letter (which seems like a blatant URL spec violation).
114 See http://curl.haxx.se/bug/view.cgi?id=1371118
116 26. NTLM authentication using SSPI (on Windows) when (lib)curl is running in
117 "system context" will make it use wrong(?) user name - at least when compared
118 to what winhttp does. See http://curl.haxx.se/bug/view.cgi?id=1281867
120 23. SOCKS-related problems:
121 A) libcurl doesn't support SOCKS for IPv6.
122 B) libcurl doesn't support FTPS over a SOCKS proxy.
123 E) libcurl doesn't support active FTP over a SOCKS proxy
125 We probably have even more bugs and lack of features when a SOCKS proxy is
128 22. Sending files to a FTP server using curl on VMS, might lead to curl
129 complaining on "unaligned file size" on completion. The problem is related
130 to VMS file structures and the perceived file sizes stat() returns. A
131 possible fix would involve sending a "STRU VMS" command.
132 http://curl.haxx.se/bug/view.cgi?id=1156287
134 21. FTP ASCII transfers do not follow RFC959. They don't convert the data
135 accordingly (not for sending nor for receiving). RFC 959 section 3.1.1.1
136 clearly describes how this should be done:
138 The sender converts the data from an internal character representation to
139 the standard 8-bit NVT-ASCII representation (see the Telnet
140 specification). The receiver will convert the data from the standard
141 form to his own internal form.
143 Since 7.15.4 at least line endings are converted.
145 16. FTP URLs passed to curl may contain NUL (0x00) in the RFC 1738 <user>,
146 <password>, and <fpath> components, encoded as "%00". The problem is that
147 curl_unescape does not detect this, but instead returns a shortened C
148 string. From a strict FTP protocol standpoint, NUL is a valid character
149 within RFC 959 <string>, so the way to handle this correctly in curl would
150 be to use a data structure other than a plain C string, one that can handle
151 embedded NUL characters. From a practical standpoint, most FTP servers
152 would not meaningfully support NUL characters within RFC 959 <string>,
153 anyway (e.g., UNIX pathnames may not contain NUL).
155 14. Test case 165 might fail on system which has libidn present, but with an
156 old iconv version (2.1.3 is a known bad version), since it doesn't recognize
157 the charset when named ISO8859-1. Changing the name to ISO-8859-1 makes the
158 test pass, but instead makes it fail on Solaris hosts that use its native
161 13. curl version 7.12.2 fails on AIX if compiled with --enable-ares.
162 The workaround is to combine --enable-ares with --disable-shared
164 12. When connecting to a SOCKS proxy, the (connect) timeout is not properly
165 acknowledged after the actual TCP connect (during the SOCKS "negotiate"
168 10. To get HTTP Negotiate authentication to work fine, you need to provide a
169 (fake) user name (this concerns both curl and the lib) because the code
170 wrongly only considers authentication if there's a user name provided.
171 http://curl.haxx.se/bug/view.cgi?id=1004841. How?
172 http://curl.haxx.se/mail/lib-2004-08/0182.html
174 8. Doing resumed upload over HTTP does not work with '-C -', because curl
175 doesn't do a HEAD first to get the initial size. This needs to be done
176 manually for HTTP PUT resume to work, and then '-C [index]'.
178 6. libcurl ignores empty path parts in FTP URLs, whereas RFC1738 states that
179 such parts should be sent to the server as 'CWD ' (without an argument).
180 The only exception to this rule, is that we knowingly break this if the
181 empty part is first in the path, as then we use the double slashes to
182 indicate that the user wants to reach the root dir (this exception SHALL
183 remain even when this bug is fixed).
185 5. libcurl doesn't treat the content-length of compressed data properly, as
186 it seems HTTP servers send the *uncompressed* length in that header and
187 libcurl thinks of it as the *compressed* length. Some explanations are here:
188 http://curl.haxx.se/mail/lib-2003-06/0146.html
190 2. If a HTTP server responds to a HEAD request and includes a body (thus
191 violating the RFC2616), curl won't wait to read the response but just stop
192 reading and return back. If a second request (let's assume a GET) is then
193 immediately made to the same server again, the connection will be re-used
194 fine of course, and the second request will be sent off but when the
195 response is to get read, the previous response-body is what curl will read
196 and havoc is what happens.
197 More details on this is found in this libcurl mailing list thread:
198 http://curl.haxx.se/mail/lib-2002-08/0000.html