5 \___|\___/|_| \_\_____|
10 - Daniel Steinberg pointed out that Curl_FormInit() in formdata.c was not
11 initializing the fread callback pointer and this triggered a compiler
12 warning, also provided a friendly suggestion on how to fix it.
14 Daniel Stenberg (8 Jun 2009)
15 - Claes Jakobsson provided a patch for libcurl-NSS that fixed a bad refcount
16 issue with client certs that caused issues like segfaults.
17 http://curl.haxx.se/mail/lib-2009-05/0316.html
19 - Triggered by bug report #2798852 and the patch in there, I fixed configure
20 to detect gnutls build options with pkg-config only and not libgnutls-config
21 anymore since GnuTLS has stopped distributing that tool. If an explicit path
22 is given to configure, we will instead guess on how to link and use that
23 lib. I did not use the patch from the bug report.
26 - Igor Novoseltsev adjusted Makefile.vxworks to get sources and headers
27 included from Makefile.inc, and provided docs\INSTALL VxWorks section.
29 - I removed buildconf.bat from release and daily snapshot archives. This
30 file is only for CVS tree checkout builds.
32 Daniel Stenberg (8 Jun 2009)
33 - Eric Wong fixed --no-buffer to actually switch off output buffering. Been
36 Bill Hoffman (6 Jun 2009)
37 - Added some cmake docs and fixed socklen_t in the build.
40 - John E. Malmberg provided VMS specific patch: "This fixes an existing bug
41 in urlglob.c where it was not converting the Curl Unix exit code to a VMS
42 DCL compatible exit code. This fix required the enhancement described next.
43 This also adds an enhancement to main.c so that when curl is run under a
44 Unix shell like Bash on VMS, it will return the standard Unix exit codes
45 and messages." And another patch for docs/examples.
47 I introduced os-specific.c and os-specific.h for use in curl tool code
48 and adjusted John E. Malmberg's patch placement to use these new files
49 as an effort to prevent main.c from growing ad infinitum. Code already
50 existing in main.c which is OS specific should be moved into these files.
52 Daniel Stenberg (4 June 2009)
53 - Setting the Content-Length: header from your app when you do a POST or PUT
54 is almost always a VERY BAD IDEA. Yet there are still apps out there doing
55 this, and now recently it triggered a bug/side-effect in libcurl as when
56 libcurl sends a POST or PUT with NTLM, it sends an empty post first when it
57 knows it will just get a 401/407 back. If the app then replaced the
58 Content-Length header, it caused the server to wait for input that libcurl
59 wouldn't send. Aaron Oneal reported this problem in bug report #2799008
60 http://curl.haxx.se/bug/view.cgi?id=2799008) and helped us verify the fix.
63 - Igor Novoseltsev provided patches and information, that after some
64 adjustments to better fit curl's way of doing things, have resulted
65 in the posibility of building libcurl for VxWorks.
67 Daniel Fandrich (2 June 2009)
68 - Checked in a Google Android make file. To use it, you must first
69 create a config.h file by running configure in the Android environment,
70 which doesn't seem to be easy to do. If no easy way can be found, a
71 static config-android.h may need to be created and checked in to the
74 Daniel Stenberg (1 June 2009)
75 - Claes Jakobsson fixed the configure script to better find and use NSS
79 - John E. Malmberg provided a VMS specific clean-up for curl.h, and pointed
80 out that the configure script was failing to detect the timeval struct on
81 VMS when building with _XOPEN_SOURCE_EXTENDED undefined due to definition
82 taking place in socket.h instead of time.h. I have adjusted configure
83 script to also include this header when checking struct timeval.
85 Daniel Stenberg (27 May 2009)
86 - Frank McGeough provided a small OpenSSL #include fix to make libcurl compile
87 fine with Nokia 5th edition 1.0 SDK for Symbian.
89 - Andre Guibert de Bruet found a call to a OpenSSL function that didn't check
90 for a failure properly.
92 - Mike Crowe pointed out that setting CURLOPT_USERPWD to NULL used to clear
93 the auth credentials back in 7.19.0 and earlier while now you have to set ""
94 to get the same effect. His patch brings back the ability to use NULL.
96 - Claes Jakobsson fixed libcurl-NSS to build fine even without the
97 PK11_CreateGenericObject() function.
99 Daniel Stenberg (25 May 2009)
100 - bug report #2796358 (http://curl.haxx.se/bug/view.cgi?id=2796358) pointed
101 out that the cookie parser would leak memory when it parses cookies that are
102 received with domain, path etc set multiple times in the same header. While
103 such a cookie is questionable, they occur in the wild and libcurl no longer
104 leaks memory for them. I added such a header to test case 8.
106 Daniel Fandrich (22 May 2009)
107 - Removed some obsolete digest code that caused a valgrind error in test 551.
109 Daniel Fandrich (20 May 2009)
110 - Added "non-existing host" test keywords to make it easy to skip those
111 tests on machines that have broken DNS configurations (such as
112 those configured to use OpenDNS).
114 Daniel Stenberg (19 May 2009)
115 - Kamil Dudka brought the patch from the Redhat bug entry
116 https://bugzilla.redhat.com/show_bug.cgi?id=427966 which was libcurl closing
117 a bad file descriptor when closing down the FTP data connection. Caolan
118 McNamara seems to be the original author of it.
120 Version 7.19.5 (18 May 2009)
122 Daniel Stenberg (17 May 2009)
123 - James Bursa posted a patch to the mailing list that fixed a problem with
124 no_proxy which made it not skip the proxy if the URL entered contained a
125 user name. I added test case 1101 to verify.
127 Daniel Stenberg (11 May 2009)
128 - Balint Szilakszi reported a memory leak when libcurl did gzip decompression
129 of streams that had some parts (legitimately) missing. We now provide and use
130 a proper cleanup function for the content encoding submodule.
131 http://curl.haxx.se/mail/lib-2009-05/0092.html
133 - Kamil Dudka provided a fix for libcurl-NSS reported by Michael Cronenworth
134 at https://bugzilla.redhat.com/show_bug.cgi?id=453612#c12
136 If an incorrect password is given while loading a private key, libcurl ends
137 up in an infinite loop consuming memory. The bug is critical.
139 - I fixed the problem with doing NTLM, POST and then following a 302 redirect,
140 as reported by Ebenezer Ikonne (on curl-users) and Laurent Rabret (on
141 curl-library). The transfer was mistakenly marked to get more data to send
142 but since it didn't actually have that, it just hung there...
144 Daniel Stenberg (10 May 2009)
145 - Andre Guibert de Bruet correctly pointed out an over-alloc with one wasted
146 byte in the digest code.
148 Yang Tse (9 May 2009)
149 - Removed DOS and TPF package's subdirectory Makefile.am, it was only used
150 to include some files in the distribution tarball serving no other purpose.
151 Files from the DOS and TPF subdirectories are now included in the EXTRA_DIST
152 of the Makefile in the parent subdirectory.
154 Yang Tse (8 May 2009)
155 - Changed host name literal in several tests to one under the haxx.se domain.
157 - Renamed vc6 workspace and project files to avoid filename clash when used
158 for conversion to later VS versions.
160 Daniel Stenberg (8 May 2009)
161 - Constantine Sapuntzakis fixed bug report #2784055
162 (http://curl.haxx.se/bug/view.cgi?id=2784055) identifying a problem to
163 connect to SOCKS proxies when using the multi interface. It turned out to
164 almost not work at all previously. We need to wait for the TCP connect to
165 be properly verified before doing the SOCKS magic.
167 There's still a flaw in the FTP code for this.
169 Daniel Stenberg (7 May 2009)
170 - Made the SO_SNDBUF setting for the data connection socket for ftp uploads as
171 well. See change 28 Apr 2009.
173 Yang Tse (7 May 2009)
174 - Fixed an issue affecting FTP transfers, introduced with the transfer.c
175 patch committed May 4.
177 Daniel Stenberg (7 May 2009)
178 - Man page *roff problems fixed thanks to input from Colin Watson. Problems
179 reported in the Debian package.
181 - Vijay G filed bug report #2723236
182 (http://curl.haxx.se/bug/view.cgi?id=2723236) identifying a problem with
183 libcurl's TFTP code and its lack of dealing with the OACK packet.
185 Yang Tse (5 May 2009)
186 - Fixed the --ftp-port address of test #251 to the CLIENTIP address, and
187 reverted the change affecting test suite harness committed 4 May.
189 Daniel Stenberg (5 May 2009)
190 - Inspired by Michael Smith's session id fix for OpenSSL, I did the
191 corresponding fix in the GnuTLS code: make sure to store the new session id
192 in case the previous re-used one is rejected.
194 Daniel Stenberg (4 May 2009)
195 - Michael Smith posted bug report #2786255
196 (http://curl.haxx.se/bug/view.cgi?id=2786255) with a patch, identifying how
197 libcurl did not deal with SSL session ids properly if the server rejected a
198 re-use of one. Starting now, it will forget the rejected one and remember
199 the new. This change was for OpenSSL only, it is likely that other SSL lib
200 code needs similar fixes.
202 Yang Tse (4 May 2009)
203 - Applied David McCreedy's "transfer.c fixes for CURL_DO_LINEEND_CONV and
204 non-ASCII platform HTTP requests" patch addressing two HTTP PUT problems:
205 1) On non-ASCII platforms not all of the protocol portions of the PUT are
206 being translated to ASCII. 2) On all platforms the line endings of part of
207 the protocol portions are mangled from CRLF to CRCRLF if data->set.crlf or
208 data->set.prefer_ascii are set (depending on CURL_DO_LINEEND_CONV).
210 - Applied David McCreedy's patch to fix test suite harness to allow test FTP
211 server and client on different machines, providing FTP client address when
212 running the FTP test server.
214 Daniel Fandrich (3 May 2009)
215 - Added and disabled test case 563 which shows KNOWN_BUGS #59. The bug
216 report failed to mention that a proxy must be used to reproduce it.
218 Yang Tse (2 May 2009)
219 - Use a build-time configured curl_socklen_t data type instead of socklen_t.
221 Yang Tse (1 May 2009)
222 - Applied David McCreedy's patches "TPF-platform specific changes to various
223 files" and "http.c fix to Curl_proxyCONNECT for non-ASCII platforms", the
224 former with minor edits.
226 Daniel Stenberg (30 Apr 2009)
227 - I was going to fix issue #59 in KNOWN_BUGS
229 If the CURLOPT_PORT option is used on an FTP URL like
230 "ftp://example.com/file;type=A" the ";type=A" is stripped off.
232 I added test case 562 to verify, only to find out that I couldn't repeat
233 this bug so I hereby consider it not a bug anymore!
235 Daniel Stenberg (29 Apr 2009)
236 - Based on bug report #2723219 (http://curl.haxx.se/bug/view.cgi?id=2723219)
237 I've now made TFTP "connections" not being kept for re-use within libcurl.
238 TFTP is UDP-based so the benefit was really low (if even existing) to begin
239 with so instead of tracking down to fix this problem we instead removed the
240 re-use. I also enabled test case 1099 that I wrote a few days ago to verify
241 that this change fixes the reported problem.
243 Daniel Stenberg (28 Apr 2009)
244 - Constantine Sapuntzakis filed bug report #2783090
245 (http://curl.haxx.se/bug/view.cgi?id=2783090) pointing out that on windows
246 we need to grow the SO_SNDBUF buffer somewhat to get really good upload
247 speeds. http://support.microsoft.com/kb/823764 has the details. Friends
248 confirmed that simply adding 32 to CURL_MAX_WRITE_SIZE is enough.
250 - Bug report #2709004 (http://curl.haxx.se/bug/view.cgi?id=2709004) by Tim
251 Chen pointed out how curl couldn't upload with resume when reading from a
254 This ended up with the introduction of a new return code for the
255 CURLOPT_SEEKFUNCTION callback that basically says that the seek failed but
256 that libcurl may try to resolve the situation anyway. In our case this means
257 libcurl will attempt to instead read that much data from the stream instead
258 of seeking and that way curl can now upload with resume when data is read
261 Daniel Stenberg (26 Apr 2009)
262 - Bug report #2779733 (http://curl.haxx.se/bug/view.cgi?id=2779733) by Sven
263 Wegener pointed out that CURLINFO_APPCONNECT_TIME didn't work with the multi
264 interface and provided a patch that fixed the problem!
266 Daniel Stenberg (24 Apr 2009)
267 - Kamil Dudka fixed another NSS-related leak when client certs were used.
269 - Bug report #2779245 (http://curl.haxx.se/bug/view.cgi?id=2779245) by Rainer
270 Koenig pointed out that the man page didn't tell that the *_proxy
271 environment variables can be specified lower case or UPPER CASE and the
272 lower case takes precedence,
274 Daniel Fandrich (21 Apr 2009)
275 - Added new libcurl source files to Amiga, RiscOS and VC6 build files.
277 Yang Tse (21 Apr 2009)
278 - Moved potential inclusion of system's malloc.h and memory.h header files to
279 setup_once.h. Inclusion of each header file is based on the definition of
280 NEED_MALLOC_H and NEED_MEMORY_H respectively.
282 Renamed libcurl's memory.h to curl_memory.h
284 Daniel Stenberg (20 Apr 2009)
285 - Leanic Lefever reported a crash and did some detailed research on why and
286 how it occurs (http://curl.haxx.se/mail/lib-2009-04/0289.html). The
287 conclusion was that if an error is detected and Curl_done() is called for
288 the connection, ftp_done() could at times return another error code that
289 then would take precedence and that new code confused existing logic that
290 works for the first error code (CURLE_SEND_ERROR) only.
292 - Gisle Vanem noticed that --libtool would produce bogus strings at times for
293 OBJECTPOINT options. Now we've introduced a new function - my_setopt_str -
294 within the app for setting plain string options to avoid the risk of this
297 Daniel Stenberg (17 Apr 2009)
298 - Pramod Sharma reported and tracked down a bug when doing FTP over a HTTP
299 proxy. libcurl would then wrongly close the connection after each
300 request. In his case it had the weird side-effect that it killed NTLM auth
301 for the proxy causing an inifinite loop!
303 I added test case 1098 to verify this fix. The test case does however not
304 properly verify that the transfers are done persistently - as I couldn't
305 think of a clever way to achieve it right now - but you need to read the
306 stderr output after a test run to see that it truly did the right thing.
308 Daniel Stenberg (13 Apr 2009)
309 - bug report #2727981 (http://curl.haxx.se/bug/view.cgi?id=2727981) by Martin
310 Storsjö pointed out how setting CURLOPT_NOBODY to 0 could be downright
311 confusing as it set the method to either GET or HEAD. The example he showed
314 curl_easy_setopt(curl, CURLOPT_PUT, 1);
315 curl_easy_setopt(curl, CURLOPT_NOBODY, 0);
317 The new way doesn't alter the method until the request is about to start. If
318 CURLOPT_NOBODY is then 1 the HTTP request will be HEAD. If CURLOPT_NOBODY is
319 0 and the request happens to have been set to HEAD, it will then instead be
320 set to GET. I believe this will be less surprising to users, and hopefully
321 not hit any existing users badly.
323 - Toshio Kuratomi reported a memory leak problem with libcurl+NSS that turned
324 out to be leaking cacerts. Kamil Dudka helped me complete the fix. The issue
325 is found in Redhat's bug tracker:
326 https://bugzilla.redhat.com/show_bug.cgi?id=453612
328 There are still memory leaks present, but they seem to have other reasons.
330 Daniel Fandrich (11 Apr 2009)
331 - Added new libcurl source files to Symbian OS build files.
332 - Improved Symbian support for SSL.
334 Yang Tse (10 Apr 2009)
335 - Daniel Johnson improved the MacOSX-Framework shell script to now perform all
336 the steps required to build a Mac OS X four way fat ppc/i386/ppc64/x86_64
337 libcurl.framework. Four way fat framework requires OS X 10.5 SDK or later.
339 Yang Tse (8 Apr 2009)
340 - Removed Sun compilers preprocessor block from curlbuild.h.dist, this also
341 removes it from the curlbuild.h file originally distributed by the cURL
342 project as this file is intended for systems not capable of running the
343 configure script. For those who have been building curl out of the source
344 code curl distribution tarball provided by curl.haxx.se the change implies
345 nothing. Previous change in this area committed 2 Apr becomes irrelevant.
347 Daniel Stenberg (6 Apr 2009)
348 - I clarified in the docs that CURLOPT_SEEKFUNCTION should return 0 on success
349 and 1 on fatal errors. Previously it only mentioned non-zero on fatal
350 errors. This is a slight change in meaning, but it follows what we've done
351 elsewhere before and it opens up for LOTS of more useful return codes
352 whenever we can think of them...
354 Yang Tse (2 Apr 2009)
355 - Fix curl_off_t definition for builds done using Sun compilers and a
356 non-configured libcurl. In this case curl_off_t data type was gated
357 to the off_t data type which depends on the _FILE_OFFSET_BITS. This
358 configuration is exactly the unwanted configuration for our curl_off_t
359 data type which must not depend on such setting. This breaks ABI for
360 libcurl libraries built with Sun compilers which were built without
361 having run the configure script with _FILE_OFFSET_BITS different than
362 64 and using the ILP32 data model.
364 Daniel Stenberg (1 Apr 2009)
365 - Andre Guibert de Bruet fixed a NULL pointer use in an infof() call if a
366 strdup() call failed.
368 Daniel Fandrich (31 Mar 2009)
369 - Properly return an error code in curl_easy_recv (reported by Jim Freeman).
371 Daniel Stenberg (18 Mar 2009)
372 - Kamil Dudka brought a patch that enables 6 additional crypto algorithms when
373 NSS is used. These ciphers were added in NSS 3.4 and require to be enabled
376 Daniel Stenberg (13 Mar 2009)
377 - Use libssh2_version() to present the libssh2 version in case the libssh2
378 library is found to support it.
380 Yang Tse (12 Mar 2009)
381 - Added missing Curl_read() return code checking in TELNET transfers.
383 - Pierre Brico found and fixed TELNET transfers not being aborted upon
384 a write callback failure.
386 Daniel Stenberg (11 Mar 2009)
387 - Kamil Dudka made the curl tool properly call curl_global_init() before any
388 other libcurl function.
390 Yang Tse (11 Mar 2009)
391 - Added missing TELNET timeout support for Windows builds. This issue was
392 reported by Pierre Brico.
394 Daniel Stenberg (9 Mar 2009)
395 - Frank Hempel found out a bug and provided the fix:
397 curl_easy_duphandle did not necessarily duplicate the CURLOPT_COOKIEFILE
398 option. It only enabled the cookie engine in the destination handle if
399 data->cookies is not NULL (where data is the source handle). In case of a
400 newly initialized handle which just had the cookie support enabled by a
401 curl_easy_setopt(handle, CURL_COOKIEFILE, "")-call, handle->cookies was
402 still NULL because the setopt-call only appends the value to
403 data->change.cookielist, hence duplicating this handle would not have the
404 cookie engine switched on.
406 We also concluded that the slist-functionality would be suitable for being
407 put in its own module rather than simply hanging out in lib/sendf.c so I
408 created lib/slist.[ch] for them.
410 - Andreas Farber made the 'buildconf' script check for the presence of m4
411 scripts to make it detect a bad checkout earlier. People with older
412 checkouts who don't do cvs update with the -d option won't get the new dirs
413 and then will get funny outputs that can be a bit hard to understand and
416 Daniel Stenberg (8 Mar 2009)
417 - Andre Guibert de Bruet found and fixed a code segment in ssluse.c where the
418 allocation of the memory BIO was not being properly checked.
420 - Andre Guibert de Bruet fixed the gnutls-using code: There are a few places
421 in the gnutls code where we were checking for negative values for errors,
422 when the man pages state that GNUTLS_E_SUCCESS is returned on success and
423 other values indicate error conditions.
425 - Bill Egert pointed out (http://curl.haxx.se/bug/view.cgi?id=2671602) that
426 curl didn't use sprintf() in a way that is documented to work in POSIX but
427 since we use our own printf() code (from libcurl) that shouldn't be a
428 problem. Nonetheless I modified the code to not rely on such particular
429 features and to not cause further raised eyebrowse with no good reason.
431 Daniel Fandrich (5 Mar 2009)
432 - Expanded the security section of the libcurl-tutorial man page to cover
433 more issues for authors to consider when writing robust libcurl-using
436 Yang Tse (5 Mar 2009)
437 - Fixed NTLM authentication memory leak on SSPI enabled Windows builds. This
438 issue was noticed by Chris Deidun.
440 Daniel Fandrich (4 Mar 2009)
441 - Fixed a problem with m4 quoting in the OpenSSL configure check reported
444 Daniel Stenberg (3 Mar 2009)
445 - David James brought a patch that make libcurl close (all) dead connections
446 whenever you attempt to open a new connection.
448 1. After cleaning up a dead connection, "continue" instead of
449 returning FALSE. This ensures that we clean up all dead connections,
450 rather than just cleaning up the first dead connection.
451 2. Move up the cleanup for dead connections so that it occurs for
452 all connections, rather than just the connections which have the same
453 preferences as our current new connection.
455 Version 7.19.4 (3 March 2009)
457 Daniel Stenberg (3 Mar 2009)
458 - David Kierznowski notified us about a security flaw
459 (http://curl.haxx.se/docs/adv_20090303.html also known as CVE-2009-0037) in
460 which previous libcurl versions (by design) can be tricked to access an
461 arbitrary local/different file instead of a remote one when
462 CURLOPT_FOLLOWLOCATION is enabled. This flaw is now fixed in this release
463 together this the addition of two new setopt options for controlling this
466 o CURLOPT_REDIR_PROTOCOLS controls what protocols libcurl is allowed to
467 follow to when CURLOPT_FOLLOWLOCATION is enabled. By default, this option
468 excludes the FILE and SCP protocols and thus you nee to explicitly allow
469 them in your app if you really want that behavior.
471 o CURLOPT_PROTOCOLS controls what protocol(s) libcurl is allowed to fetch
472 using the primary URL option. This is useful if you want to allow a user or
473 other outsiders control what URL to pass to libcurl and yet not allow all
474 protocols libcurl may have been built to support.
476 Daniel Stenberg (27 Feb 2009)
477 - Senthil Raja Velu reported a problem when CURLOPT_INTERFACE and
478 CURLOPT_LOCALPORT were used together (the local port bind failed), and
479 Markus Koetter provided the fix!
481 Daniel Stenberg (25 Feb 2009)
482 - As Daniel Fandrich figured out, we must do the GnuTLS initing in the
483 curl_global_init() function to properly maintain the performing functions
484 thread-safe. We've previously (28 April 2007) moved the init to a later time
485 just to avoid it to fail very early when libgcrypt dislikes the situation,
486 but that move was bad and the fix should rather be in libgcrypt or
489 Daniel Stenberg (24 Feb 2009)
490 - Brian J. Murrell found out that Negotiate proxy authentication didn't work.
491 It happened because the code used the struct for server-based auth all the
492 time for both proxy and server auth which of course was wrong.
494 Daniel Stenberg (23 Feb 2009)
495 - After a bug reported by James Cheng I've made curl_easy_getinfo() for
496 CURLINFO_CONTENT_LENGTH_DOWNLOAD and CURLINFO_CONTENT_LENGTH_UPLOAD return
497 -1 if the sizes aren't know. Previously these returned 0, make it impossible
498 to detect the difference between actually zero and unknown.
500 Yang Tse (23 Feb 2009)
501 - Daniel Johnson provided a shell script that will perform all the steps needed
502 to build a Mac OS X fat ppc/i386 or ppc64/x86_64 libcurl.framework
504 Daniel Stenberg (23 Feb 2009)
505 - I renamed everything in the windows builds files that used the name 'curllib'
506 to the proper 'libcurl' as clearly this caused confusion.
508 Yang Tse (20 Feb 2009)
509 - Do not halt compilation when using VS2008 to build a Windows 2000 target.
511 Daniel Stenberg (20 Feb 2009)
512 - Linus Nielsen Feltzing reported and helped me repeat and fix a problem with
513 FTP with the multi interface: when a transfer fails, like when aborted by a
514 write callback, the control connection was wrongly closed and thus not
517 This change is also an attempt to cleanup the code somewhat in this area, as
518 now the FTP code attempts to keep (better) track on pending responses
519 necessary to get read in ftp_done().
521 Daniel Stenberg (19 Feb 2009)
522 - Patrik Thunstrom reported a problem and helped me repeat it. It turned out
523 libcurl did a superfluous 1000ms wait when doing SFTP downloads!
525 We read data with libssh2 while doing the "DO" operation for SFTP and then
526 when we were about to start getting data for the actual file part, the
527 "TRANSFER" part, we waited for socket action (in 1000ms) before doing a
528 libssh2-read. But in this case libssh2 had already read and buffered the
529 data so we ended up always just waiting 1000ms before we get working on the
532 Patrick Monnerat (18 Feb 2009)
533 - FTP downloads (i.e.: RETR) ending with code 550 now return error
534 CURLE_REMOTE_FILE_NOT_FOUND instead of CURLE_FTP_COULDNT_RETR_FILE.
536 Daniel Stenberg (17 Feb 2009)
537 - Kamil Dudka made NSS-powered builds compile and run again!
539 - A second follow-up change by Andre Guibert de Bruet to fix a related memory
540 leak like that fixed on the 14th. When zlib returns failure, we need to
541 cleanup properly before returning error.
543 - CURLOPT_FTP_CREATE_MISSING_DIRS can now be set to 2 in addition to 1 for
544 plain FTP connections, and it will then allow MKD to fail once and retry the
545 CWD afterwards. This is especially useful if you're doing many simultanoes
546 connections against the same server and they all have this option enabled,
547 as then CWD may first fail but then another connection does MKD before this
548 connection and thus MKD fails but trying CWD works! The numbers can
549 (should?) now be set with the convenience enums now called
550 CURLFTP_CREATE_DIR and CURLFTP_CREATE_DIR_RETRY.
552 Tests has proven that if you're making an application that uploads a set of
553 files to an ftp server, you will get a noticable gain in speed if you're
554 using multiple connections and this option will be then be very useful.
556 Daniel Stenberg (14 Feb 2009)
557 - Andre Guibert de Bruet found and fixed a memory leak in the content encoding
558 code, which could happen on libz errors.
560 Daniel Fandrich (12 Feb 2009)
561 - Added support for Digest and NTLM authentication using GnuTLS.
563 Daniel Stenberg (11 Feb 2009)
564 - CURLINFO_CONDITION_UNMET was added to allow an application to get to know if
565 the condition in the previous request was unmet. This is typically a time
566 condition set with CURLOPT_TIMECONDITION and was previously not possible to
567 reliably figure out. From bug report #2565128
568 (http://curl.haxx.se/bug/view.cgi?id=2565128) filed by Jocelyn Jaubert.
570 Daniel Fandrich (4 Feb 2009)
571 - Don't add the standard /usr/lib or /usr/include paths to LDFLAGS and CPPFLAGS
572 (respectively) when --with-ssl=/usr is used (patch based on FreeBSD).
574 - Added an explicit buffer limit check in msdosify() (patch based on FreeBSD).
575 This couldn't ever overflow in curl, but might if the code were used
576 elsewhere or under different conditions.
578 Daniel Stenberg (3 Feb 2009)
579 - Hidemoto Nakada provided a small fix that makes it possible to get the
580 CURLINFO_CONTENT_LENGTH_DOWNLOAD size from file:// "transfers" with
581 CURLOPT_NOBODY set true.
583 Daniel Stenberg (2 Feb 2009)
584 - Patrick Scott found a rather large memory leak when using the multi
585 interface and setting CURLMOPT_MAXCONNECTS to something less than the number
586 of handles you add to the multi handle. All the connections that didn't fit
587 in the cache would not be properly disconnected nor freed!
589 - Craig A West brought us: libcurl now defaults to do CONNECT with HTTP
590 version 1.1 instead of 1.0 like before. This change also introduces the new
591 proxy type for libcurl called 'CURLPROXY_HTTP_1_0' that then allows apps to
592 switch (back) to CONNECT 1.0 requests. The curl tool also got a --proxy1.0
593 option that works exactly like --proxy but sets CURLPROXY_HTTP_1_0.
595 I updated all test cases cases that use CONNECT and I tried to do some using
596 --proxy1.0 and some updated to do CONNECT 1.1 to get both versions run.
598 Daniel Stenberg (31 Jan 2009)
599 - When building with c-ares 1.6.1 (not yet released) or later and IPv6 support
600 enabled, we can now take advantage of its brand new AF_UNSPEC support in
601 ares_gethostbyname(). This makes test case 241 finally run fine for me with
602 this setup since it now parses the "::1 ip6-localhost" line fine in my
605 Daniel Stenberg (30 Jan 2009)
606 - Scott Cantor filed bug report #2550061
607 (http://curl.haxx.se/bug/view.cgi?id=2550061) mentioning that I failed to
608 properly make sure that the VC9 makefiles got included in the latest
609 release. I've now fixed the release script and verified it so next release
610 will hopefully include them properly!
612 Daniel Fandrich (30 Jan 2009)
613 - Fixed --disable-proxy for FTP and SOCKS. Thanks to Daniel Egger for
616 Yang Tse (29 Jan 2009)
617 - Introduced curl_sspi.c and curl_sspi.h for the implementation of functions
618 Curl_sspi_global_init() and Curl_sspi_global_cleanup() which previously were
619 named Curl_ntlm_global_init() and Curl_ntlm_global_cleanup() in http_ntlm.c
620 Also adjusted socks_sspi.c to remove the link-time dependency on the Windows
621 SSPI library using it now in the same way as it was done in http_ntlm.c.
623 Daniel Stenberg (28 Jan 2009)
624 - Markus Moeller introduced two new options to libcurl:
625 CURLOPT_SOCKS5_GSSAPI_SERVICE and CURLOPT_SOCKS5_GSSAPI_NEC to allow libcurl
626 to do GSS-style authentication with SOCKS5 proxies. The curl tool got the
627 options called --socks5-gssapi-service and --socks5-gssapi-nec to enable
630 Daniel Stenberg (26 Jan 2009)
631 - Chad Monroe provided the new CURLOPT_TFTP_BLKSIZE option that allows an app
632 to set desired block size to use for TFTP transfers instead of the default
635 - The "-no_ticket" option was introduced in Openssl0.9.8j. It's a flag to
636 disable "rfc4507bis session ticket support". rfc4507bis was later turned
637 into the proper RFC5077 it seems: http://tools.ietf.org/html/rfc5077
639 The enabled extension concerns the session management. I wonder how often
640 libcurl stops a connection and then resumes a TLS session. also, sending the
641 session data is some overhead. .I suggest that you just use your proposed
642 patch (which explicitly disables TICKET).
644 If someone writes an application with libcurl and openssl who wants to
645 enable the feature, one can do this in the SSL callback.
647 Sharad Gupta brought this to my attention. Peter Sylvester helped me decide
648 on the proper action.
650 - Alexey Borzov filed bug report #2535504
651 (http://curl.haxx.se/bug/view.cgi?id=2535504) pointing out that realms with
652 quoted quotation marks in HTTP Digest headers didn't work. I've now added
653 test case 1095 that verifies my fix.
655 - Craig A West brought CURLOPT_NOPROXY and the corresponding --noproxy option.
656 They basically offer the same thing the NO_PROXY environment variable only
657 offered previously: list a set of host names that shall not use the proxy
658 even if one is specified.
660 Daniel Fandrich (20 Jan 2009)
661 - Call setlocale() for libtest tests to test the effects of locale-induced
662 libc changes on libcurl.
664 - Fixed a couple more locale-dependent toupper conversions, mainly for
665 clarity. This does fix one problem that causes ;type=i FTP URLs
666 to fail in the Turkish locale when CURLOPT_PROXY_TRANSFER_MODE is
669 - Added tests 561 and 1091 through 1094 to test various combinations
670 of ;type= and ;mode= URLs that could potentially fail in the Turkish
673 Daniel Stenberg (20 Jan 2009)
674 - Lisa Xu pointed out that the ssh.obj file was missing from the
675 lib/Makefile.vc6 file (and thus from the vc8 and vc9 ones too).
677 Version 7.19.3 (19 January 2009)
679 Daniel Stenberg (16 Jan 2009)
680 - Andrew de los Reyes fixed curlbuild.h for "generic" gcc builds on PPC, both
683 Daniel Stenberg (15 Jan 2009)
684 - Tim Ansell fixed a compiler warning in lib/cookie.c
686 Daniel Stenberg (14 Jan 2009)
687 - Grant Erickson fixed timeouts for TFTP such that specifying a
688 connect-timeout, a max-time or both options work correctly and as expected
689 by passing the correct boolean value to Curl_timeleft via the
690 'duringconnect' parameter.
692 With this small change, curl TFTP now behaves as expected (and likely as
693 originally-designed):
695 1) For non-existent or unreachable dotted IP addresses:
697 a) With no options, follows the default curl 300s timeout...
698 b) With --connect-timeout only, follows that value...
699 c) With --max-time only, follows that value...
700 d) With both --connect-timeout and --max-time, follows the smaller value...
702 and times out with a "curl: (7) Couldn't connect to server" error.
704 2) For transfers to/from a valid host:
706 a) With no options, follows default curl 300s timeout for the
707 first XRQ/DATA/ACK transaction and the default TFTP 3600s
708 timeout for the remainder of the transfer...
710 b) With --connect-time only, follows that value for the
711 first XRQ/DATA/ACK transaction and the default TFTP 3600s
712 timeout for the remainder of the transfer...
714 c) With --max-time only, follows that value for the first
715 XRQ/DATA/ACK transaction and for the remainder of the
718 d) With both --connect-timeout and --max-time, follows the former
719 for the first XRQ/DATA/ACK transaction and the latter for the
720 remainder of the transfer...
722 and times out with a "curl: (28) Timeout was reached" error as
725 Daniel Stenberg (13 Jan 2009)
726 - Michael Wallner fixed a NULL pointer deref when calling
727 curl_easy_setup(curl, CURLOPT_COOKIELIST, "SESS") on a CURL handle with no
730 - Stefan Teleman brought a patch to fix the default curlbuild.h file for the
733 Daniel Stenberg (12 Jan 2009)
734 - Based on bug report #2498665 (http://curl.haxx.se/bug/view.cgi?id=2498665)
735 by Daniel Black, I've now added magic to the configure script that makes it
736 use pkg-config to detect gnutls details as well if the existing method
737 (using libgnutls-config) fails. While doing this, I cleaned up and unified
738 the pkg-config usage when detecting openssl and nss as well.
740 Daniel Stenberg (11 Jan 2009)
741 - Karl Moerder brought the patch that creates vc9 Makefiles, and I made
742 'maketgz' now use the actual makefile targets to do the VC8 and VC9
745 Daniel Stenberg (10 Jan 2009)
746 - Emil Romanus fixed:
748 When using the multi interface over HTTP and the server returns a Location
749 header, the running easy handle will get stuck in the CURLM_STATE_PERFORM
750 state, leaving the external event loop stuck waiting for data from the
751 ingoing socket (when using the curl_multi_socket_action stuff). While this
752 bug was pretty hard to find, it seems to require only a one-line fix. The
753 break statement on line 1374 in multi.c caused the function to skip the call
756 How to reproduce this bug? Well, that's another question. evhiperfifo.c in
757 the examples directory chokes on this bug only _sometimes_, probably
758 depending on how fast the URLs are added. One way of testing the bug out is
759 writing to hiper.fifo from more than one source at the same time.
761 Daniel Fandrich (7 Jan 2009)
762 - Unified much of the SessionHandle initialization done in Curl_open() and
763 curl_easy_reset() by creating Curl_init_userdefined(). This had the side
764 effect of fixing curl_easy_reset() so it now also resets
765 CURLOPT_FTP_FILEMETHOD and CURLOPT_SSL_SESSIONID_CACHE
767 Daniel Stenberg (7 Jan 2009)
768 - Rob Crittenden did once again provide an NSS update:
770 I have to jump through a few hoops now with the NSS library initialization
771 since another part of an application may have already initialized NSS by the
772 time Curl gets invoked. This patch is more careful to only shutdown the NSS
773 library if Curl did the initialization.
775 It also adds in a bit of code to set the default ciphers if the app that
776 call NSS_Init* did not call NSS_SetDomesticPolicy() or set specific
777 ciphers. One might argue that this lets other application developers get
778 lazy and/or they aren't using the NSS API correctly, and you'd be right.
779 But still, this will avoid terribly difficult-to-trace crashes and is
782 Daniel Stenberg (1 Jan 2009)
783 - 'reconf' is removed since we rather have users use 'buildconf'
785 Daniel Stenberg (31 Dec 2008)
786 - Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
787 out that 'reconf' didn't properly point out the m4 subdirectory when running
790 Daniel Stenberg (29 Dec 2008)
791 - Phil Lisiecki filed bug report #2413067
792 (http://curl.haxx.se/bug/view.cgi?id=2413067) that identified a problem that
793 would cause libcurl to mark a DNS cache entry "in use" eternally if the
794 subsequence TCP connect failed. It would thus never get pruned and refreshed
795 as it should've been.
797 Phil provided his own patch to this problem that while it seemed to work
798 wasn't complete and thus I wrote my own fix to the problem.
800 Daniel Stenberg (28 Dec 2008)
801 - Peter Korsgaard fixed building libcurl with "configure --with-ssl
804 - Anthony Bryan fixed more language and spelling flaws in man pages.
806 Daniel Stenberg (22 Dec 2008)
807 - Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
808 on file indexes beyond 2 or 4GB.
810 - Anthony Bryan provided a set of patches that cleaned up manual language,
811 corrected spellings and more.
813 Daniel Stenberg (20 Dec 2008)
814 - Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
815 pipelining, as libcurl could then easily get confused and A) work on the
816 handle that was not "first in queue" on a pipeline, or even B) tell the app
817 to REMOVE a socket while it was in use by a second handle in a pipeline. Both
818 errors caused hanging or stalling applications.
820 Daniel Stenberg (19 Dec 2008)
821 - curl_multi_timeout() could return a timeout value of 0 even though nothing
822 was actually ready to get done, as the internal time resolution is higher
823 than the returned millisecond timer. Therefore it could cause applications
824 running on fast processors to do short bursts of busy-loops.
825 curl_multi_timeout() will now only return 0 if the timeout is actually
828 - Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
829 now has an improved ability to do right when the multi interface (both
830 "regular" and multi_socket) is used for SCP and SFTP transfers. This should
831 result in (much) less busy-loop situations and thus less CPU usage with no
834 Daniel Stenberg (17 Dec 2008)
835 - SCP and SFTP with the multi interface had the same flaw: the 'DONE'
836 operation didn't complete properly if the EAGAIN equivalent was returned but
837 libcurl would simply continue with a half-completed close operation
838 performed. This ruined persistent connection re-use and cause some
839 SSH-protocol errors in general. The correction is unfortunately adding a
840 blocking function - doing it entirely non-blocking should be considered for
843 Gisle Vanem (16 Dec 2008)
844 - Added the possibility to use the Watt-32 tcp/ip stack under Windows.
845 The change simply involved adding a USE_WATT32 section in the
846 config-win32.h files (under ./lib and ./src). This section disables
847 the use of any Winsock headers.
849 Daniel Stenberg (16 Dec 2008)
850 - libssh2_sftp_last_error() was wrongly used at some places in libcurl which
851 made libcurl sometimes not properly abort problematic SFTP transfers.
853 Daniel Stenberg (12 Dec 2008)
854 - More work with Igor Novoseltsev to first fix the remaining stuff for
855 removing easy handles from multi handles when the easy handle is/was within
856 a HTTP pipeline. His bug report #2351653
857 (http://curl.haxx.se/bug/view.cgi?id=2351653) was also related and was
858 eventually fixed by a patch by Igor himself.
860 Yang Tse (12 Dec 2008)
861 - Patrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
862 OS/400 compilations with IPv6 enabled.
864 Daniel Stenberg (12 Dec 2008)
865 - Mark Karpeles filed bug report #2416182 titled "crash in ConnectionExists
866 when using duphandle+curl_mutli"
867 (http://curl.haxx.se/bug/view.cgi?id=2416182) which showed that
868 curl_easy_duphandle() wrongly also copied the pointer to the connection
869 cache, which was plain wrong and caused a segfault if the handle would be
870 used in a different multi handle than the handle it was duplicated from.
872 Daniel Stenberg (11 Dec 2008)
873 - Keshav Krity found out that libcurl failed to deal with dotted IPv6
874 addresses if they were very long (>39 letters) due to a too strict address
875 validity parser. It now accepts addresses up to 45 bytes long.
877 Daniel Stenberg (11 Dec 2008)
878 - Internet Explorer had a broken HTTP digest authentication before v7 and
879 there are servers "out there" that relies on the client doing this broken
880 Digest authentication. Apache even comes with an option to work with such
883 The difference is only for URLs that contain a query-part (a '?'-letter and
884 text to the right of it).
886 libcurl now supports this quirk, and you enable it by setting the
887 CURLAUTH_DIGEST_IE bit in the bitmask you pass to the CURLOPT_HTTPAUTH or
888 CURLOPT_PROXYAUTH options. They are thus individually controlled to server
891 (note that there's no way to activate this with the curl tool yet)
893 Daniel Fandrich (9 Dec 2008)
894 - Added test cases 1089 and 1090 to test --write-out after a redirect to
895 test a report that the size didn't work, but these test cases pass.
897 - Documented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.
899 Daniel Stenberg (9 Dec 2008)
900 - Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
901 particular state for the control connection like it did before for implicit
902 FTPS (libcurl assumed such control connections to be encrypted while some
903 FTPS servers such as FileZilla assumes such connections to be clear
904 mode). Use the CURLOPT_USE_SSL option to set your desired level.
906 Daniel Stenberg (8 Dec 2008)
907 - Fred Machado posted about a weird FTP problem on the curl-users list and when
908 researching it, it turned out he got a 550 response back from a SIZE command
909 and then I fell over the text in RFC3659 that says:
911 The presence of the 550 error response to a SIZE command MUST NOT be taken
912 by the client as an indication that the file cannot be transferred in the
913 current MODE and TYPE.
915 In other words: the change I did on September 30th 2008 and that has been
916 included in the last two releases were a regression and a bad idea. We MUST
917 NOT take a 550 response from SIZE as a hint that the file doesn't exist.
919 - Christian Krause filed bug #2221237
920 (http://curl.haxx.se/bug/view.cgi?id=2221237) that identified an infinite
921 loop during GSS authentication given some specific conditions. With his
922 patience and great feedback I managed to narrow down the problem and
923 eventually fix it although I can't test any of this myself!
925 Daniel Fandrich (3 Dec 2008)
926 - Fixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
929 Daniel Stenberg (3 Dec 2008)
930 - Igor Novoseltsev filed bug #2351645
931 (http://curl.haxx.se/bug/view.cgi?id=2351645) that identified a problem with
932 the multi interface that occured if you removed an easy handle while in
933 progress and the handle was used in a HTTP pipeline.
935 - Pawel Kierski pointed out a mistake in the cookie code that could lead to a
936 bad fclose() after a fatal error had occured.
937 (http://curl.haxx.se/bug/view.cgi?id=2382219)
939 Daniel Fandrich (25 Nov 2008)
940 - If a HTTP request is Basic and num is already >=1000, the HTTP test
941 server adds 1 to num to get the data section to return. This allows
942 testing authentication negotiations using the Basic authentication
945 - Added tests 1087 and 1088 to test Basic authentication on a redirect
946 with and without --location-trusted
948 Daniel Stenberg (24 Nov 2008)
949 - Based on a patch by Vlad Grachov, libcurl now uses a new libssh2 0.19
950 function when built to support SCP and SFTP that helps the library to know
951 in which direction a particular libssh2 operation would return EAGAIN so
952 that libcurl knows what socket conditions to wait for before trying the
953 function call again. Previously (and still when using libssh2 0.18 or
954 earlier), libcurl will busy-loop in this situation when the easy interface
957 Daniel Fandrich (20 Nov 2008)
958 - Automatically detect OpenBSD's CA cert bundle.
960 Daniel Stenberg (19 Nov 2008)
961 - I removed the default use of "Pragma: no-cache" from libcurl when a proxy is
962 used. It has been used since forever but it was never a good idea to use
963 unless explicitly asked for.
965 - Josef Wolf's extension that allows a $TESTDIR/gdbinit$testnum file that when
966 you use runtests.pl -g, will be sourced by gdb to allow additional fancy or
969 - Christian Krause reported and fixed a memory leak that would occur with HTTP
970 GSS/kerberos authentication (http://curl.haxx.se/bug/view.cgi?id=2284386)
972 - Andreas Wurf and Markus Koetter helped me analyze a problem that Andreas got
973 when uploading files to a single FTP server using multiple easy handle
974 handles with the multi interface. Occasionally a handle would stall in
977 The problem turned out to be a side-effect of the ConnectionExists()
978 function's eagerness to re-use a handle for HTTP pipelining so it would
979 select it even if already being in use, due to an inadequate check for its
980 chances of being used for pipelnining.
982 Daniel Fandrich (17 Nov 2008)
983 - Added more compiler warning options for gcc 4.3
985 Yang Tse (17 Nov 2008)
986 - Fix a remaining problem in the inet_pton() runtime configure check. And
987 fix internal Curl_inet_pton() failures to reject certain malformed literals.
989 - Make configure script check if ioctl with the SIOCGIFADDR command can be
990 used, and define HAVE_IOCTL_SIOCGIFADDR if appropriate.
992 Daniel Stenberg (16 Nov 2008)
993 - Christian Krause fixed a build failure when building with gss support
994 enabled and FTP disabled.
996 - Added check for NULL returns from strdup() in src/main.c and lib/formdata.c
997 - reported by Jim Meyering also prevent buffer overflow on MSDOS when you do
998 for example -O on a url with a file name part longer than PATH_MAX letters
1000 - lib/nss.c fixes based on the report by Jim Meyering: I went over and added
1001 checks for return codes for all calls to malloc and strdup that were
1002 missing. I also changed a few malloc(13) to use arrays on the stack and a
1003 few malloc(PATH_MAX) to instead use aprintf() to lower memory use.
1005 - I fixed a memory leak in Curl_nss_connect() when CURLOPT_ISSUERCERT is
1008 Daniel Fandrich (14 Nov 2008)
1009 - Added .xml as one of the few common file extensions known by the multipart
1012 - Added some #ifdefs around header files and change the EAGAIN test to
1013 fix compilation on Cell (reported by Jeff Curley).
1015 Yang Tse (14 Nov 2008)
1016 - Fixed several configure script issues affecting checks for inet_ntoa_r(),
1017 inet_ntop(), inet_pton(), getifaddrs(), fcntl() and getaddrinfo().
1019 Yang Tse (13 Nov 2008)
1020 - Refactored configure script detection of functions used to set sockets into
1021 non-blocking mode, and decouple function detection from function capability.