| blob | 2c263e72cd1630f9500759c78ae4f84b77d061cf |
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 58. It seems sensible to be able to use CURLOPT_NOBODY and
7 CURLOPT_FAILONERROR with FTP to detect if a file exists or not, but it is
8 not working: http://curl.haxx.se/mail/lib-2008-07/0295.html
10 57. On VMS-Alpha: When using an http-file-upload the file is not sent to the
11 Server with the correct content-length. Sending a file with 511 or less
12 bytes, content-length 512 is used. Sending a file with 513 - 1023 bytes,
13 content-length 1024 is used. Files with a length of a multiple of 512 Bytes
14 show the correct content-length. Only these files work for upload.
15 http://curl.haxx.se/bug/view.cgi?id=2057858
17 56. When libcurl sends CURLOPT_POSTQUOTE commands when connected to a SFTP
18 server using the multi interface, the commands are not being sent correctly
19 and instead the connection is "cancelled" (the operation is considered done)
20 prematurely. There is a half-baked (busy-looping) patch provided in the bug
21 report but it cannot be accepted as-is. See
22 http://curl.haxx.se/bug/view.cgi?id=2006544
24 55. libcurl fails to build with MIT Kerberos for Windows (KfW) due to KfW's
25 library header files exporting symbols/macros that should be kept private
26 to the KfW library. See ticket #5601 at http://krbdev.mit.edu/rt/
28 53. SFTP busy-loop problem. When doing SFTP uploads, we can see that libcurl
29 occasionally will busy-loop while waiting for certain network conditions.
30 Reported by Pavel Shalagin, explained somewhat by Daniel Stenberg here:
31 http://curl.haxx.se/mail/lib-2008-04/0439.html
33 52. Gautam Kachroo's issue that identifies a problem with the multi interface
34 where a connection can be re-used without actually being properly
35 SSL-negoatiated:
36 http://curl.haxx.se/mail/lib-2008-01/0277.html
38 49. If using --retry and the transfer timeouts (possibly due to using -m or
39 -y/-Y) the next attempt doesn't resume the transfer properly from what was
40 downloaded in the previous attempt but will truncate and restart at the
41 original position where it was at before the previous failed attempt. See
42 http://curl.haxx.se/mail/lib-2008-01/0080.html and Mandriva bug report
43 https://qa.mandriva.com/show_bug.cgi?id=22565
45 48. If a CONNECT response-headers are larger than BUFSIZE (16KB) when the
46 connection is meant to be kept alive (like for NTLM proxy auth), the
47 function will return prematurely and will confuse the rest of the HTTP
48 protocol code. This should be very rare.
50 45. libcurl built to support ipv6 uses getaddrinfo() to resolve host names.
51 getaddrinfo() sorts the response list which effectively kills how libcurl
52 deals with round-robin DNS entries. All details:
53 http://curl.haxx.se/mail/lib-2007-07/0168.html
54 initial suggested function to use for randomizing the response:
55 http://curl.haxx.se/mail/lib-2007-07/0178.html
57 43. There seems to be a problem when connecting to the Microsoft telnet server.
58 http://curl.haxx.se/bug/view.cgi?id=1720605
60 41. When doing an operation over FTP that requires the ACCT command (but not
61 when logging in), the operation will fail since libcurl doesn't detect this
62 and thus fails to issue the correct command:
63 http://curl.haxx.se/bug/view.cgi?id=1693337
65 39. Steffen Rumler's Race Condition in Curl_proxyCONNECT:
66 http://curl.haxx.se/mail/lib-2007-01/0045.html
68 38. Kumar Swamy Bhatt's problem in ftp/ssl "LIST" operation:
69 http://curl.haxx.se/mail/lib-2007-01/0103.html
71 37. Having more than one connection to the same host when doing NTLM
72 authentication (with performs multiple "passes" and authenticates a
73 connection rather than a HTTP request), and particularly when using the
74 multi interface, there's a risk that libcurl will re-use a wrong connection
75 when doing the different passes in the NTLM negotiation and thus fail to
76 negotiate (in seemingly mysterious ways).
78 35. Both SOCKS5 and SOCKS4 proxy connections are done blocking, which is very
79 bad when used with the multi interface.
81 34. The SOCKS4 connection codes don't properly acknowledge (connect) timeouts.
82 Also see #12. According to bug #1556528, even the SOCKS5 connect code does
83 not do it right: http://curl.haxx.se/bug/view.cgi?id=1556528,
85 31. "curl-config --libs" will include details set in LDFLAGS when configure is
86 run that might be needed only for building libcurl. Further, curl-config
87 --cflags suffers from the same effects with CFLAGS/CPPFLAGS.
89 30. You need to use -g to the command line tool in order to use RFC2732-style
90 IPv6 numerical addresses in URLs.
92 29. IPv6 URLs with zone ID is not supported.
93 http://www.ietf.org/internet-drafts/draft-fenner-literal-zone-02.txt (expired)
94 specifies the use of a plus sign instead of a percent when specifying zone
95 IDs in URLs to get around the problem of percent signs being
96 special. According to the reporter, Firefox deals with the URL _with_ a
97 percent letter (which seems like a blatant URL spec violation).
99 See http://curl.haxx.se/bug/view.cgi?id=1371118
101 26. NTLM authentication using SSPI (on Windows) when (lib)curl is running in
102 "system context" will make it use wrong(?) user name - at least when compared
103 to what winhttp does. See http://curl.haxx.se/bug/view.cgi?id=1281867
105 23. SOCKS-related problems:
106 A) libcurl doesn't support SOCKS for IPv6.
107 B) libcurl doesn't support FTPS over a SOCKS proxy.
108 E) libcurl doesn't support active FTP over a SOCKS proxy
110 We probably have even more bugs and lack of features when a SOCKS proxy is
111 used.
113 22. Sending files to a FTP server using curl on VMS, might lead to curl
114 complaining on "unaligned file size" on completion. The problem is related
115 to VMS file structures and the perceived file sizes stat() returns. A
116 possible fix would involve sending a "STRU VMS" command.
117 http://curl.haxx.se/bug/view.cgi?id=1156287
119 21. FTP ASCII transfers do not follow RFC959. They don't convert the data
120 accordingly (not for sending nor for receiving). RFC 959 section 3.1.1.1
121 clearly describes how this should be done:
123 The sender converts the data from an internal character representation to
124 the standard 8-bit NVT-ASCII representation (see the Telnet
125 specification). The receiver will convert the data from the standard
126 form to his own internal form.
128 Since 7.15.4 at least line endings are converted.
130 16. FTP URLs passed to curl may contain NUL (0x00) in the RFC 1738 <user>,
131 <password>, and <fpath> components, encoded as "%00". The problem is that
132 curl_unescape does not detect this, but instead returns a shortened C
133 string. From a strict FTP protocol standpoint, NUL is a valid character
134 within RFC 959 <string>, so the way to handle this correctly in curl would
135 be to use a data structure other than a plain C string, one that can handle
136 embedded NUL characters. From a practical standpoint, most FTP servers
137 would not meaningfully support NUL characters within RFC 959 <string>,
138 anyway (e.g., UNIX pathnames may not contain NUL).
140 14. Test case 165 might fail on system which has libidn present, but with an
141 old iconv version (2.1.3 is a known bad version), since it doesn't recognize
142 the charset when named ISO8859-1. Changing the name to ISO-8859-1 makes the
143 test pass, but instead makes it fail on Solaris hosts that use its native
144 iconv.
146 13. curl version 7.12.2 fails on AIX if compiled with --enable-ares.
147 The workaround is to combine --enable-ares with --disable-shared
149 12. When connecting to a SOCKS proxy, the (connect) timeout is not properly
150 acknowledged after the actual TCP connect (during the SOCKS "negotiate"
151 phase).
153 10. To get HTTP Negotiate authentication to work fine, you need to provide a
154 (fake) user name (this concerns both curl and the lib) because the code
155 wrongly only considers authentication if there's a user name provided.
156 http://curl.haxx.se/bug/view.cgi?id=1004841. How?
157 http://curl.haxx.se/mail/lib-2004-08/0182.html
159 8. Doing resumed upload over HTTP does not work with '-C -', because curl
160 doesn't do a HEAD first to get the initial size. This needs to be done
161 manually for HTTP PUT resume to work, and then '-C [index]'.
163 7. CURLOPT_USERPWD and CURLOPT_PROXYUSERPWD have no way of providing user names
164 that contain a colon. This can't be fixed easily in a backwards compatible
165 way without adding new options (and then, they should most probably allow
166 setting user name and password separately).
168 6. libcurl ignores empty path parts in FTP URLs, whereas RFC1738 states that
169 such parts should be sent to the server as 'CWD ' (without an argument).
170 The only exception to this rule, is that we knowingly break this if the
171 empty part is first in the path, as then we use the double slashes to
172 indicate that the user wants to reach the root dir (this exception SHALL
173 remain even when this bug is fixed).
175 5. libcurl doesn't treat the content-length of compressed data properly, as
176 it seems HTTP servers send the *uncompressed* length in that header and
177 libcurl thinks of it as the *compressed* length. Some explanations are here:
178 http://curl.haxx.se/mail/lib-2003-06/0146.html
180 2. If a HTTP server responds to a HEAD request and includes a body (thus
181 violating the RFC2616), curl won't wait to read the response but just stop
182 reading and return back. If a second request (let's assume a GET) is then
183 immediately made to the same server again, the connection will be re-used
184 fine of course, and the second request will be sent off but when the
185 response is to get read, the previous response-body is what curl will read
186 and havoc is what happens.
187 More details on this is found in this libcurl mailing list thread:
188 http://curl.haxx.se/mail/lib-2002-08/0000.html