KNOWN_BUGS 3.0 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061
  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!
  5. * NTLM authentication with passwords longer than 14 letters fail. There is
  6. a known fix for this, planned to come in curl 7.11.2
  7. * Doing resumed upload over HTTP does not work with '-C -', because curl
  8. doesn't do a HEAD first to get the initial size. This needs to be done
  9. manually for HTTP PUT resume to work, and then '-C [index]'.
  10. * CURLOPT_USERPWD and CURLOPT_PROXYUSERPWD have no way of providing user names
  11. that contain a colon. This can't be fixed easily in a backwards compatible
  12. way without adding new options (and then, they should most probably allow
  13. setting user name and password separately).
  14. * libcurl ignores empty path parts in FTP URLs, whereas RFC1738 states that
  15. such parts should be sent to the server as 'CWD ' (without an argument).
  16. The only exception to this rule, is that we knowingly break this if the
  17. empty part is first in the path, as then we use the double slashes to
  18. indicate that the user wants to reach the root dir (this exception SHALL
  19. remain even when this bug is fixed).
  20. * 1) libcurl does a POST
  21. 2) receives a 100-continue
  22. 3) sends away the POST
  23. Now, if nothing else is returned from the server, libcurl MUST return
  24. CURLE_GOT_NOTHING, but it seems it returns CURLE_OK as it seems to count
  25. the 100-continue reply as a good enough reply.
  26. * libcurl doesn't treat the content-length of compressed data properly, as
  27. it seems HTTP servers send the *uncompressed* length in that header and
  28. libcurl thinks of it as the *compressed* lenght. Some explanations are here:
  29. http://curl.haxx.se/mail/lib-2003-06/0146.html
  30. * Downloading 0 (zero) bytes files over FTP will not create a zero byte file
  31. locally, which is because libcurl doesn't call the write callback with zero
  32. bytes. Explained here: http://curl.haxx.se/mail/archive-2003-04/0143.html
  33. * Using CURLOPT_FAILONERROR (-f/--fail) will make authentication to stop
  34. working if you use anything but plain Basic auth.
  35. * IPv6 support on AIX 4.3.3 doesn't work due to a missing sockaddr_storage
  36. struct. It has been reported to work on AIX 5.1 though.
  37. * GOPHER transfers seem broken
  38. * configure --disable-http is not fully supported. All other protocols seem
  39. to work to disable.
  40. * If a HTTP server responds to a HEAD request and includes a body (thus
  41. violating the RFC2616), curl won't wait to read the response but just stop
  42. reading and return back. If a second request (let's assume a GET) is then
  43. immediately made to the same server again, the connection will be re-used
  44. fine of course, and the second request will be sent off but when the
  45. response is to get read, the previous response-body is what curl will read
  46. and havoc is what happens.
  47. More details on this is found in this libcurl mailing list thread:
  48. http://curl.haxx.se/mail/lib-2002-08/0000.html