5 \___|\___/|_| \_\_____|
9 These are problems and bugs known to exist at the time of this release. Feel
10 free to join in and help us correct one or more of these! Also be sure to
11 check the changelog of the current development status, as one or more of these
12 problems may have been fixed or changed somewhat since this was written!
15 1.1 CURLFORM_CONTENTLEN in an array
16 1.2 Disabling HTTP Pipelining
17 1.3 STARTTRANSFER time is wrong for HTTP POSTs
18 1.4 multipart formposts file name encoding
19 1.5 Expect-100 meets 417
20 1.6 Unnecessary close when 401 received waiting for 100
21 1.7 CONNECT response larger than 16KB
22 1.8 DNS timing is wrong for HTTP redirects
23 1.9 HTTP/2 frames while in the connection pool kill reuse
24 1.10 Strips trailing dot from host name
25 1.11 transfer-encoding: chunked in HTTP/2
26 1.12 CURLOPT_SEEKFUNCTION not called with CURLFORM_STREAM
29 2.1 Hangs with PolarSSL
30 2.2 CURLINFO_SSL_VERIFYRESULT has limited support
32 2.4 GnuTLS backend skips really long certificate fields
35 3.1 IMAP SEARCH ALL truncated response
36 3.2 No disconnect command
37 3.3 SMTP to multiple recipients
38 3.4 POP3 expects "CRLF.CRLF" eob for some single-line responses
41 4.1 -J with %-encoded file nameas
42 4.2 -J with -C - fails
43 4.3 --retry and transfer timeouts
45 5. Build and portability issues
46 5.1 Windows Borland compiler
47 5.2 curl-config --libs contains private details
48 5.3 libidn and old iconv
49 5.4 AIX shared build with c-ares fails
50 5.5 can't handle Unicode arguments in Windows
51 5.6 cmake support gaps
52 5.7 Visual Studio project gaps
55 6.1 NTLM authentication and unicode
56 6.2 MIT Kerberos for Windows build
57 6.3 NTLM in system context uses wrong name
58 6.4 Negotiate and Kerberos V5 need a fake user name
61 7.1 FTP without or slow 220 response
62 7.2 FTP with CONNECT and slow server
63 7.3 FTP with NOBODY and FAILONERROR
66 7.6 FTP with NULs in URL parts
67 7.7 FTP and empty path parts in the URL
68 7.8 Premature transfer end but healthy control channel
71 8.1 TELNET and time limtiations don't work
72 8.2 Microsoft telnet server
75 9.1 SFTP doesn't do CURLOPT_POSTQUOTE correct
78 10.1 SOCKS proxy connections are done blocking
79 10.2 SOCKS don't support timeouts
81 10.4 active FTP over a SOCKS
84 11.1 Curl leaks .onion hostnames in DNS
85 11.2 error buffer not set if connection to multiple addresses fails
88 12.1 OpenLDAP hangs after returning results
91 13.1 --interface for ipv6 binds to unusable IP address
94 ==============================================================================
98 1.1 CURLFORM_CONTENTLEN in an array
100 It is not possible to pass a 64-bit value using CURLFORM_CONTENTLEN with
101 CURLFORM_ARRAY, when compiled on 32-bit platforms that support 64-bit
102 integers. This is because the underlying structure 'curl_forms' uses a dual
103 purpose char* for storing these values in via casting. For more information
104 see the now closed related issue:
105 https://github.com/curl/curl/issues/608
107 1.2 Disabling HTTP Pipelining
109 Disabling HTTP Pipelining when there are ongoing transfers can lead to
110 heap corruption and crash. https://curl.haxx.se/bug/view.cgi?id=1411
112 1.3 STARTTRANSFER time is wrong for HTTP POSTs
114 Wrong STARTTRANSFER timer accounting for POST requests Timer works fine with
115 GET requests, but while using POST the time for CURLINFO_STARTTRANSFER_TIME
116 is wrong. While using POST CURLINFO_STARTTRANSFER_TIME minus
117 CURLINFO_PRETRANSFER_TIME is near to zero every time.
119 https://github.com/curl/curl/issues/218
120 https://curl.haxx.se/bug/view.cgi?id=1213
122 1.4 multipart formposts file name encoding
124 When creating multipart formposts. The file name part can be encoded with
125 something beyond ascii but currently libcurl will only pass in the verbatim
126 string the app provides. There are several browsers that already do this
127 encoding. The key seems to be the updated draft to RFC2231:
128 https://tools.ietf.org/html/draft-reschke-rfc2231-in-http-02
130 1.5 Expect-100 meets 417
132 If an upload using Expect: 100-continue receives an HTTP 417 response, it
133 ought to be automatically resent without the Expect:. A workaround is for
134 the client application to redo the transfer after disabling Expect:.
135 https://curl.haxx.se/mail/archive-2008-02/0043.html
137 1.6 Unnecessary close when 401 received waiting for 100
139 libcurl closes the connection if an HTTP 401 reply is received while it is
140 waiting for the the 100-continue response.
141 https://curl.haxx.se/mail/lib-2008-08/0462.html
143 1.7 CONNECT response larger than 16KB
145 If a CONNECT response-headers are larger than BUFSIZE (16KB) when the
146 connection is meant to be kept alive (like for NTLM proxy auth), the function
147 will return prematurely and will confuse the rest of the HTTP protocol
148 code. This should be very rare.
150 1.8 DNS timing is wrong for HTTP redirects
152 When extracting timing information after HTTP redirects, only the last
153 transfer's results are returned and not the totals:
154 https://github.com/curl/curl/issues/522
156 1.9 HTTP/2 frames while in the connection pool kill reuse
158 If the server sends HTTP/2 frames (like for example an HTTP/2 PING frame) to
159 curl while the connection is held in curl's connection pool, the socket will
160 be found readable when considered for reuse and that makes curl think it is
161 dead and then it will be closed and a new connection gets created instead.
163 This is *best* fixed by adding monitoring to connections while they are kept
164 in the pool so that pings can be responded to appropriately.
166 1.10 Strips trailing dot from host name
168 When given a URL wit a trailing dot for the host name part:
169 "https://example.com./", libcurl will strip off the dot and use the name
170 without a dot internally and send it dot-less in HTTP Host: headers and in
173 The HTTP part violates RFC 7230 section 5.4 but the SNI part is accordance
174 with RFC 6066 section 3.
176 URLs using these trailing dots are very rare in the wild and we have not seen
177 or gotten any real-world problems with such URLs reported. The popular
178 browsers seem to have stayed with not stripping the dot for both uses (thus
179 they violate RFC 6066 instead of RFC 7230).
181 Daniel took the discussion to the HTTPbis mailing list in March 2016:
182 https://lists.w3.org/Archives/Public/ietf-http-wg/2016JanMar/0430.html but
183 there was not major rush or interest to fix this. The impression I get is
184 that most HTTP people rather not rock the boat now and instead prioritize web
185 compatibility rather than to strictly adhere to these RFCs.
187 Our current approach allows a knowing client to send a custom HTTP header
190 It can also be noted that while adding a trailing dot to the host name in
191 most (all?) cases will make the name resolve to the same set of IP addresses,
192 many HTTP servers will not happily accept the trailing dot there unless that
193 has been specificly configured to be a fine virtual host.
195 If URLs with trailing dots for host names become more popular or even just
196 used more than for just plain fun experiments, I'm sure we will have reason
197 to go back and reconsider.
199 See https://github.com/curl/curl/issues/716 for the discussion.
201 1.11 transfer-encoding: chunked in HTTP/2
203 For HTTP/1, when -H transfer-encoding:chunked option is given, curl encodes
204 the request using chunked encoding. But when HTTP/2 is being used, the
205 command wrongly sends a request with both content-length and
206 transfer-encoding: chunked headers being set (and the request body is not
207 chunked-encoded). See https://github.com/curl/curl/issues/662
209 1.12 CURLOPT_SEEKFUNCTION not called with CURLFORM_STREAM
211 I'm using libcurl to POST form data using a FILE* with the CURLFORM_STREAM
212 option of curl_formadd(). I've noticed that if the connection drops at just
213 the right time, the POST is reattempted without the data from the file. It
214 seems like the file stream position isn't getting reset to the beginning of
215 the file. I found the CURLOPT_SEEKFUNCTION option and set that with a
216 function that performs an fseek() on the FILE*. However, setting that didn't
217 seem to fix the issue or even get called. See
218 https://github.com/curl/curl/issues/768
223 2.1 Hangs with PolarSSL
225 "curl_easy_perform hangs with imap and PolarSSL"
226 https://github.com/curl/curl/issues/334
228 Most likely, a fix similar to commit c111178bd4 (for mbedTLS) is
229 necessary. Or if we just wait a little longer we'll rip out all support for
232 2.2 CURLINFO_SSL_VERIFYRESULT has limited support
234 CURLINFO_SSL_VERIFYRESULT is only implemented for the OpenSSL and NSS
235 backends, so relying on this information in a generic app is flaky.
239 Curl doesn't recognize certificates in DER format in keychain, but it works
240 with PEM. https://curl.haxx.se/bug/view.cgi?id=1065
242 2.4 GnuTLS backend skips really long certificate fields
244 libcurl calls gnutls_x509_crt_get_dn() with a fixed buffer size and if the
245 field is too long in the cert, it'll just return an error and the field will
251 3.1 IMAP SEARCH ALL truncated response
253 IMAP "SEARCH ALL" truncates output on large boxes. "A quick search of the
254 code reveals that pingpong.c contains some truncation code, at line 408, when
255 it deems the server response to be too large truncating it to 40 characters"
256 https://curl.haxx.se/bug/view.cgi?id=1366
258 3.2 No disconnect command
260 The disconnect commands (LOGOUT and QUIT) may not be sent by IMAP, POP3 and
261 SMTP if a failure occurs during the authentication phase of a connection.
263 3.3 SMTP to multiple recipients
265 When sending data to multiple recipients, curl will abort and return failure
266 if one of the recipients indicate failure (on the "RCPT TO"
267 command). Ordinary mail programs would proceed and still send to the ones
268 that can receive data. This is subject for change in the future.
269 https://curl.haxx.se/bug/view.cgi?id=1116
271 3.4 POP3 expects "CRLF.CRLF" eob for some single-line responses
273 You have to tell libcurl not to expect a body, when dealing with one line
274 response commands. Please see the POP3 examples and test cases which show
275 this for the NOOP and DELE commands. https://curl.haxx.se/bug/?i=740
280 4.1 -J with %-encoded file nameas
282 -J/--remote-header-name doesn't decode %-encoded file names. RFC6266 details
283 how it should be done. The can of worm is basically that we have no charset
284 handling in curl and ascii >=128 is a challenge for us. Not to mention that
285 decoding also means that we need to check for nastiness that is attempted,
286 like "../" sequences and the like. Probably everything to the left of any
287 embedded slashes should be cut off.
288 https://curl.haxx.se/bug/view.cgi?id=1294
290 4.2 -J with -C - fails
292 When using -J (with -O), automatically resumed downloading together with "-C
293 -" fails. Without -J the same command line works! This happens because the
294 resume logic is worked out before the target file name (and thus its
295 pre-transfer size) has been figured out!
296 https://curl.haxx.se/bug/view.cgi?id=1169
298 4.3 --retry and transfer timeouts
300 If using --retry and the transfer timeouts (possibly due to using -m or
301 -y/-Y) the next attempt doesn't resume the transfer properly from what was
302 downloaded in the previous attempt but will truncate and restart at the
303 original position where it was at before the previous failed attempt. See
304 https://curl.haxx.se/mail/lib-2008-01/0080.html and Mandriva bug report
305 https://qa.mandriva.com/show_bug.cgi?id=22565
308 5. Build and portability issues
310 5.1 Windows Borland compiler
312 When building with the Windows Borland compiler, it fails because the "tlib"
313 tool doesn't support hyphens (minus signs) in file names and we have such in
314 the build. https://curl.haxx.se/bug/view.cgi?id=1222
316 5.2 curl-config --libs contains private details
318 "curl-config --libs" will include details set in LDFLAGS when configure is
319 run that might be needed only for building libcurl. Further, curl-config
320 --cflags suffers from the same effects with CFLAGS/CPPFLAGS.
322 5.3 libidn and old iconv
324 Test case 165 might fail on a system which has libidn present, but with an
325 old iconv version (2.1.3 is a known bad version), since it doesn't recognize
326 the charset when named ISO8859-1. Changing the name to ISO-8859-1 makes the
327 test pass, but instead makes it fail on Solaris hosts that use its native
330 5.4 AIX shared build with c-ares fails
332 curl version 7.12.2 fails on AIX if compiled with --enable-ares. The
333 workaround is to combine --enable-ares with --disable-shared
335 5.5 can't handle Unicode arguments in Windows
337 If a URL or filename can't be encoded using the user's current codepage then
338 it can only be encoded properly in the Unicode character set. Windows uses
339 UTF-16 encoding for Unicode and stores it in wide characters, however curl
340 and libcurl are not equipped for that at the moment. And, except for Cygwin,
341 Windows can't use UTF-8 as a locale.
343 https://curl.haxx.se/bug/?i=345
344 https://curl.haxx.se/bug/?i=731
346 5.6 cmake support gaps
348 The cmake build setup lacks several features that the autoconf build
349 offers. This includes:
351 - symbol hiding when the shared library is built
352 - use of correct soname for the shared library build
353 - support for several TLS backends are missing
354 - the unit tests cause link failures in regular non-static builds
357 5.7 Visual Studio project gaps
359 The Visual Studio projects lack some features that the autoconf and nmake
360 builds offer, such as the following:
362 - support for zlib and nghttp2
363 - use of static runtime libraries
364 - add the test suite components
366 In addition to this the following could be implemented:
368 - support for other development IDEs
369 - add PATH environment variables for third-party DLLs
373 6.1 NTLM authentication and unicode
375 NTLM authentication involving unicode user name or password only works
376 properly if built with UNICODE defined together with the WinSSL/schannel
377 backend. The original problem was mentioned in:
378 https://curl.haxx.se/mail/lib-2009-10/0024.html
379 https://curl.haxx.se/bug/view.cgi?id=896
381 The WinSSL/schannel version verified to work as mentioned in
382 https://curl.haxx.se/mail/lib-2012-07/0073.html
384 6.2 MIT Kerberos for Windows build
386 libcurl fails to build with MIT Kerberos for Windows (KfW) due to KfW's
387 library header files exporting symbols/macros that should be kept private to
388 the KfW library. See ticket #5601 at http://krbdev.mit.edu/rt/
390 6.3 NTLM in system context uses wrong name
392 NTLM authentication using SSPI (on Windows) when (lib)curl is running in
393 "system context" will make it use wrong(?) user name - at least when compared
394 to what winhttp does. See https://curl.haxx.se/bug/view.cgi?id=535
396 6.4 Negotiate and Kerberos V5 need a fake user name
398 In order to get Negotiate (SPNEGO) authentication to work in HTTP or Kerberos
399 V5 in the e-mail protocols, you need to provide a (fake) user name (this
400 concerns both curl and the lib) because the code wrongly only considers
401 authentication if there's a user name provided by setting
402 conn->bits.user_passwd in url.c https://curl.haxx.se/bug/view.cgi?id=440 How?
403 https://curl.haxx.se/mail/lib-2004-08/0182.html A possible solution is to
404 either modify this variable to be set or introduce a variable such as
405 new conn->bits.want_authentication which is set when any of the authentication
411 7.1 FTP without or slow 220 response
413 If a connection is made to a FTP server but the server then just never sends
414 the 220 response or otherwise is dead slow, libcurl will not acknowledge the
415 connection timeout during that phase but only the "real" timeout - which may
416 surprise users as it is probably considered to be the connect phase to most
417 people. Brought up (and is being misunderstood) in:
418 https://curl.haxx.se/bug/view.cgi?id=856
420 7.2 FTP with CONNECT and slow server
422 When doing FTP over a socks proxy or CONNECT through HTTP proxy and the multi
423 interface is used, libcurl will fail if the (passive) TCP connection for the
424 data transfer isn't more or less instant as the code does not properly wait
425 for the connect to be confirmed. See test case 564 for a first shot at a test
428 7.3 FTP with NOBODY and FAILONERROR
430 It seems sensible to be able to use CURLOPT_NOBODY and CURLOPT_FAILONERROR
431 with FTP to detect if a file exists or not, but it is not working:
432 https://curl.haxx.se/mail/lib-2008-07/0295.html
436 When doing an operation over FTP that requires the ACCT command (but not when
437 logging in), the operation will fail since libcurl doesn't detect this and
438 thus fails to issue the correct command:
439 https://curl.haxx.se/bug/view.cgi?id=635
443 FTP ASCII transfers do not follow RFC959. They don't convert the data
444 accordingly (not for sending nor for receiving). RFC 959 section 3.1.1.1
445 clearly describes how this should be done:
447 The sender converts the data from an internal character representation to
448 the standard 8-bit NVT-ASCII representation (see the Telnet
449 specification). The receiver will convert the data from the standard
450 form to his own internal form.
452 Since 7.15.4 at least line endings are converted.
454 7.6 FTP with NULs in URL parts
456 FTP URLs passed to curl may contain NUL (0x00) in the RFC 1738 <user>,
457 <password>, and <fpath> components, encoded as "%00". The problem is that
458 curl_unescape does not detect this, but instead returns a shortened C string.
459 From a strict FTP protocol standpoint, NUL is a valid character within RFC
460 959 <string>, so the way to handle this correctly in curl would be to use a
461 data structure other than a plain C string, one that can handle embedded NUL
462 characters. From a practical standpoint, most FTP servers would not
463 meaningfully support NUL characters within RFC 959 <string>, anyway (e.g.,
464 Unix pathnames may not contain NUL).
466 7.7 FTP and empty path parts in the URL
468 libcurl ignores empty path parts in FTP URLs, whereas RFC1738 states that
469 such parts should be sent to the server as 'CWD ' (without an argument). The
470 only exception to this rule, is that we knowingly break this if the empty
471 part is first in the path, as then we use the double slashes to indicate that
472 the user wants to reach the root dir (this exception SHALL remain even when
475 7.8 Premature transfer end but healthy control channel
477 When 'multi_done' is called before the transfer has been completed the normal
478 way, it is considered a "premature" transfer end. In this situation, libcurl
479 closes the connection assuming it doesn't know the state of the connection so
480 it can't be reused for subsequent requests.
482 With FTP however, this isn't necessarily true but there are a bunch of
483 situations (listed in the ftp_done code) where it *could* keep the connection
484 alive even in this situation - but the current code doesn't. Fixing this would
485 allow libcurl to reuse FTP connections better.
489 8.1 TELNET and time limtiations don't work
491 When using telnet, the time limitation options don't work.
492 https://curl.haxx.se/bug/view.cgi?id=846
494 8.2 Microsoft telnet server
496 There seems to be a problem when connecting to the Microsoft telnet server.
497 https://curl.haxx.se/bug/view.cgi?id=649
502 9.1 SFTP doesn't do CURLOPT_POSTQUOTE correct
504 When libcurl sends CURLOPT_POSTQUOTE commands when connected to a SFTP server
505 using the multi interface, the commands are not being sent correctly and
506 instead the connection is "cancelled" (the operation is considered done)
507 prematurely. There is a half-baked (busy-looping) patch provided in the bug
508 report but it cannot be accepted as-is. See
509 https://curl.haxx.se/bug/view.cgi?id=748
514 10.1 SOCKS proxy connections are done blocking
516 Both SOCKS5 and SOCKS4 proxy connections are done blocking, which is very bad
517 when used with the multi interface.
519 10.2 SOCKS don't support timeouts
521 The SOCKS4 connection codes don't properly acknowledge (connect) timeouts.
522 According to bug #1556528, even the SOCKS5 connect code does not do it right:
523 https://curl.haxx.se/bug/view.cgi?id=604
525 When connecting to a SOCK proxy, the (connect) timeout is not properly
526 acknowledged after the actual TCP connect (during the SOCKS "negotiate"
531 libcurl doesn't support FTPS over a SOCKS proxy.
533 10.4 active FTP over a SOCKS
535 libcurl doesn't support active FTP over a SOCKS proxy
540 11.1 Curl leaks .onion hostnames in DNS
542 Curl sends DNS requests for hostnames with a .onion TLD. This leaks
543 information about what the user is attempting to access, and violates this
544 requirement of RFC7686: https://tools.ietf.org/html/rfc7686
546 Issue: https://github.com/curl/curl/issues/543
548 11.2 error buffer not set if connection to multiple addresses fails
550 If you ask libcurl to resolve a hostname like example.com to IPv6 addresses
551 only. But you only have IPv4 connectivity. libcurl will correctly fail with
552 CURLE_COULDNT_CONNECT. But the error buffer set by CURLOPT_ERRORBUFFER
553 remains empty. Issue: https://github.com/curl/curl/issues/544
556 12. LDAP and OpenLDAP
558 12.1 OpenLDAP hangs after returning results
560 By configuration defaults, openldap automatically chase referrals on
561 secondary socket descriptors. The OpenLDAP backend is asynchronous and thus
562 should monitor all socket descriptors involved. Currently, these secondary
563 descriptors are not monitored, causing openldap library to never receive
566 As a temporary workaround, disable referrals chasing by configuration.
568 The fix is not easy: proper automatic referrals chasing requires a
569 synchronous bind callback and monitoring an arbitrary number of socket
570 descriptors for a single easy handle (currently limited to 5).
572 Generic LDAP is synchronous: OK.
574 See https://github.com/curl/curl/issues/622 and
575 https://curl.haxx.se/mail/lib-2016-01/0101.html
580 13.1 --interface for ipv6 binds to unusable IP address
582 Since IPv6 provides a lot of addresses with different scope, binding to an
583 IPv6 address needs to take the proper care so that it doesn't bind to a
584 locally scoped address as that is bound to fail.
586 https://github.com/curl/curl/issues/686