5 \___|\___/|_| \_\_____|
10 - Moved potential inclusion of system's malloc.h and memory.h header files to
11 setup_once.h. Inclusion of each header file is based on the definition of
12 NEED_MALLOC_H and NEED_MEMORY_H respectively.
14 Daniel Stenberg (20 Apr 2009)
15 - Leanic Lefever reported a crash and did some detailed research on why and
16 how it occurs (http://curl.haxx.se/mail/lib-2009-04/0289.html). The
17 conclusion was that if an error is detected and Curl_done() is called for
18 the connection, ftp_done() could at times return another error code that
19 then would take precedence and that new code confused existing logic that
20 works for the first error code (CURLE_SEND_ERROR) only.
22 - Gisle Vanem noticed that --libtool would produce bogus strings at times for
23 OBJECTPOINT options. Now we've introduced a new function - my_setopt_str -
24 within the app for setting plain string options to avoid the risk of this
27 Daniel Stenberg (17 Apr 2009)
28 - Pramod Sharma reported and tracked down a bug when doing FTP over a HTTP
29 proxy. libcurl would then wrongly close the connection after each
30 request. In his case it had the weird side-effect that it killed NTLM auth
31 for the proxy causing an inifinite loop!
33 I added test case 1098 to verify this fix. The test case does however not
34 properly verify that the transfers are done persistently - as I couldn't
35 think of a clever way to achieve it right now - but you need to read the
36 stderr output after a test run to see that it truly did the right thing.
38 Daniel Stenberg (13 Apr 2009)
39 - bug report #2727981 (http://curl.haxx.se/bug/view.cgi?id=2727981) by Martin
40 Storsjö pointed out how setting CURLOPT_NOBODY to 0 could be downright
41 confusing as it set the method to either GET or HEAD. The example he showed
44 curl_easy_setopt(curl, CURLOPT_PUT, 1);
45 curl_easy_setopt(curl, CURLOPT_NOBODY, 0);
47 The new way doesn't alter the method until the request is about to start. If
48 CURLOPT_NOBODY is then 1 the HTTP request will be HEAD. If CURLOPT_NOBODY is
49 0 and the request happens to have been set to HEAD, it will then instead be
50 set to GET. I believe this will be less surprising to users, and hopefully
51 not hit any existing users badly.
53 - Toshio Kuratomi reported a memory leak problem with libcurl+NSS that turned
54 out to be leaking cacerts. Kamil Dudka helped me complete the fix. The issue
55 is found in Redhat's bug tracker:
56 https://bugzilla.redhat.com/show_bug.cgi?id=453612
58 There are still memory leaks present, but they seem to have other reasons.
60 Daniel Fandrich (11 Apr 2009)
61 - Added new libcurl source files to Symbian OS build files.
62 - Improved Symbian support for SSL.
64 Yang Tse (10 Apr 2009)
65 - Daniel Johnson improved the MacOSX-Framework shell script to now perform all
66 the steps required to build a Mac OS X four way fat ppc/i386/ppc64/x86_64
67 libcurl.framework. Four way fat framework requires OS X 10.5 SDK or later.
70 - Removed Sun compilers preprocessor block from curlbuild.h.dist, this also
71 removes it from the curlbuild.h file originally distributed by the cURL
72 project as this file is intended for systems not capable of running the
73 configure script. For those who have been building curl out of the source
74 code curl distribution tarball provided by curl.haxx.se the change implies
75 nothing. Previous change in this area committed 2 Apr becomes irrelevant.
77 Daniel Stenberg (6 Apr 2009)
78 - I clarified in the docs that CURLOPT_SEEKFUNCTION should return 0 on success
79 and 1 on fatal errors. Previously it only mentioned non-zero on fatal
80 errors. This is a slight change in meaning, but it follows what we've done
81 elsewhere before and it opens up for LOTS of more useful return codes
82 whenever we can think of them...
85 - Fix curl_off_t definition for builds done using Sun compilers and a
86 non-configured libcurl. In this case curl_off_t data type was gated
87 to the off_t data type which depends on the _FILE_OFFSET_BITS. This
88 configuration is exactly the unwanted configuration for our curl_off_t
89 data type which must not depend on such setting. This breaks ABI for
90 libcurl libraries built with Sun compilers which were built without
91 having run the configure script with _FILE_OFFSET_BITS different than
92 64 and using the ILP32 data model.
94 Daniel Stenberg (1 Apr 2009)
95 - Andre Guibert de Bruet fixed a NULL pointer use in an infof() call if a
98 Daniel Fandrich (31 Mar 2009)
99 - Properly return an error code in curl_easy_recv (reported by Jim Freeman).
101 Daniel Stenberg (18 Mar 2009)
102 - Kamil Dudka brought a patch that enables 6 additional crypto algorithms when
103 NSS is used. These ciphers were added in NSS 3.4 and require to be enabled
106 Daniel Stenberg (13 Mar 2009)
107 - Use libssh2_version() to present the libssh2 version in case the libssh2
108 library is found to support it.
110 Yang Tse (12 Mar 2009)
111 - Added missing Curl_read() return code checking in TELNET transfers.
113 - Pierre Brico found and fixed TELNET transfers not being aborted upon
114 a write callback failure.
116 Daniel Stenberg (11 Mar 2009)
117 - Kamil Dudka made the curl tool properly call curl_global_init() before any
118 other libcurl function.
120 Yang Tse (11 Mar 2009)
121 - Added missing TELNET timeout support for Windows builds. This issue was
122 reported by Pierre Brico.
124 Daniel Stenberg (9 Mar 2009)
125 - Frank Hempel found out a bug and provided the fix:
127 curl_easy_duphandle did not necessarily duplicate the CURLOPT_COOKIEFILE
128 option. It only enabled the cookie engine in the destination handle if
129 data->cookies is not NULL (where data is the source handle). In case of a
130 newly initialized handle which just had the cookie support enabled by a
131 curl_easy_setopt(handle, CURL_COOKIEFILE, "")-call, handle->cookies was
132 still NULL because the setopt-call only appends the value to
133 data->change.cookielist, hence duplicating this handle would not have the
134 cookie engine switched on.
136 We also concluded that the slist-functionality would be suitable for being
137 put in its own module rather than simply hanging out in lib/sendf.c so I
138 created lib/slist.[ch] for them.
140 - Andreas Farber made the 'buildconf' script check for the presence of m4
141 scripts to make it detect a bad checkout earlier. People with older
142 checkouts who don't do cvs update with the -d option won't get the new dirs
143 and then will get funny outputs that can be a bit hard to understand and
146 Daniel Stenberg (8 Mar 2009)
147 - Andre Guibert de Bruet found and fixed a code segment in ssluse.c where the
148 allocation of the memory BIO was not being properly checked.
150 - Andre Guibert de Bruet fixed the gnutls-using code: There are a few places
151 in the gnutls code where we were checking for negative values for errors,
152 when the man pages state that GNUTLS_E_SUCCESS is returned on success and
153 other values indicate error conditions.
155 - Bill Egert pointed out (http://curl.haxx.se/bug/view.cgi?id=2671602) that
156 curl didn't use sprintf() in a way that is documented to work in POSIX but
157 since we use our own printf() code (from libcurl) that shouldn't be a
158 problem. Nonetheless I modified the code to not rely on such particular
159 features and to not cause further raised eyebrowse with no good reason.
161 Daniel Fandrich (5 Mar 2009)
162 - Expanded the security section of the libcurl-tutorial man page to cover
163 more issues for authors to consider when writing robust libcurl-using
166 Yang Tse (5 Mar 2009)
167 - Fixed NTLM authentication memory leak on SSPI enabled Windows builds. This
168 issue was noticed by Chris Deidun.
170 Daniel Fandrich (4 Mar 2009)
171 - Fixed a problem with m4 quoting in the OpenSSL configure check reported
174 Daniel Stenberg (3 Mar 2009)
175 - David James brought a patch that make libcurl close (all) dead connections
176 whenever you attempt to open a new connection.
178 1. After cleaning up a dead connection, "continue" instead of
179 returning FALSE. This ensures that we clean up all dead connections,
180 rather than just cleaning up the first dead connection.
181 2. Move up the cleanup for dead connections so that it occurs for
182 all connections, rather than just the connections which have the same
183 preferences as our current new connection.
185 Version 7.19.4 (3 March 2009)
187 Daniel Stenberg (3 Mar 2009)
188 - David Kierznowski notified us about a security flaw
189 (http://curl.haxx.se/docs/adv_20090303.html also known as CVE-2009-0037) in
190 which previous libcurl versions (by design) can be tricked to access an
191 arbitrary local/different file instead of a remote one when
192 CURLOPT_FOLLOWLOCATION is enabled. This flaw is now fixed in this release
193 together this the addition of two new setopt options for controlling this
196 o CURLOPT_REDIR_PROTOCOLS controls what protocols libcurl is allowed to
197 follow to when CURLOPT_FOLLOWLOCATION is enabled. By default, this option
198 excludes the FILE and SCP protocols and thus you nee to explicitly allow
199 them in your app if you really want that behavior.
201 o CURLOPT_PROTOCOLS controls what protocol(s) libcurl is allowed to fetch
202 using the primary URL option. This is useful if you want to allow a user or
203 other outsiders control what URL to pass to libcurl and yet not allow all
204 protocols libcurl may have been built to support.
206 Daniel Stenberg (27 Feb 2009)
207 - Senthil Raja Velu reported a problem when CURLOPT_INTERFACE and
208 CURLOPT_LOCALPORT were used together (the local port bind failed), and
209 Markus Koetter provided the fix!
211 Daniel Stenberg (25 Feb 2009)
212 - As Daniel Fandrich figured out, we must do the GnuTLS initing in the
213 curl_global_init() function to properly maintain the performing functions
214 thread-safe. We've previously (28 April 2007) moved the init to a later time
215 just to avoid it to fail very early when libgcrypt dislikes the situation,
216 but that move was bad and the fix should rather be in libgcrypt or
219 Daniel Stenberg (24 Feb 2009)
220 - Brian J. Murrell found out that Negotiate proxy authentication didn't work.
221 It happened because the code used the struct for server-based auth all the
222 time for both proxy and server auth which of course was wrong.
224 Daniel Stenberg (23 Feb 2009)
225 - After a bug reported by James Cheng I've made curl_easy_getinfo() for
226 CURLINFO_CONTENT_LENGTH_DOWNLOAD and CURLINFO_CONTENT_LENGTH_UPLOAD return
227 -1 if the sizes aren't know. Previously these returned 0, make it impossible
228 to detect the difference between actually zero and unknown.
230 Yang Tse (23 Feb 2009)
231 - Daniel Johnson provided a shell script that will perform all the steps needed
232 to build a Mac OS X fat ppc/i386 or ppc64/x86_64 libcurl.framework
234 Daniel Stenberg (23 Feb 2009)
235 - I renamed everything in the windows builds files that used the name 'curllib'
236 to the proper 'libcurl' as clearly this caused confusion.
238 Yang Tse (20 Feb 2009)
239 - Do not halt compilation when using VS2008 to build a Windows 2000 target.
241 Daniel Stenberg (20 Feb 2009)
242 - Linus Nielsen Feltzing reported and helped me repeat and fix a problem with
243 FTP with the multi interface: when a transfer fails, like when aborted by a
244 write callback, the control connection was wrongly closed and thus not
247 This change is also an attempt to cleanup the code somewhat in this area, as
248 now the FTP code attempts to keep (better) track on pending responses
249 necessary to get read in ftp_done().
251 Daniel Stenberg (19 Feb 2009)
252 - Patrik Thunstrom reported a problem and helped me repeat it. It turned out
253 libcurl did a superfluous 1000ms wait when doing SFTP downloads!
255 We read data with libssh2 while doing the "DO" operation for SFTP and then
256 when we were about to start getting data for the actual file part, the
257 "TRANSFER" part, we waited for socket action (in 1000ms) before doing a
258 libssh2-read. But in this case libssh2 had already read and buffered the
259 data so we ended up always just waiting 1000ms before we get working on the
262 Patrick Monnerat (18 Feb 2009)
263 - FTP downloads (i.e.: RETR) ending with code 550 now return error
264 CURLE_REMOTE_FILE_NOT_FOUND instead of CURLE_FTP_COULDNT_RETR_FILE.
266 Daniel Stenberg (17 Feb 2009)
267 - Kamil Dudka made NSS-powered builds compile and run again!
269 - A second follow-up change by Andre Guibert de Bruet to fix a related memory
270 leak like that fixed on the 14th. When zlib returns failure, we need to
271 cleanup properly before returning error.
273 - CURLOPT_FTP_CREATE_MISSING_DIRS can now be set to 2 in addition to 1 for
274 plain FTP connections, and it will then allow MKD to fail once and retry the
275 CWD afterwards. This is especially useful if you're doing many simultanoes
276 connections against the same server and they all have this option enabled,
277 as then CWD may first fail but then another connection does MKD before this
278 connection and thus MKD fails but trying CWD works! The numbers can
279 (should?) now be set with the convenience enums now called
280 CURLFTP_CREATE_DIR and CURLFTP_CREATE_DIR_RETRY.
282 Tests has proven that if you're making an application that uploads a set of
283 files to an ftp server, you will get a noticable gain in speed if you're
284 using multiple connections and this option will be then be very useful.
286 Daniel Stenberg (14 Feb 2009)
287 - Andre Guibert de Bruet found and fixed a memory leak in the content encoding
288 code, which could happen on libz errors.
290 Daniel Fandrich (12 Feb 2009)
291 - Added support for Digest and NTLM authentication using GnuTLS.
293 Daniel Stenberg (11 Feb 2009)
294 - CURLINFO_CONDITION_UNMET was added to allow an application to get to know if
295 the condition in the previous request was unmet. This is typically a time
296 condition set with CURLOPT_TIMECONDITION and was previously not possible to
297 reliably figure out. From bug report #2565128
298 (http://curl.haxx.se/bug/view.cgi?id=2565128) filed by Jocelyn Jaubert.
300 Daniel Fandrich (4 Feb 2009)
301 - Don't add the standard /usr/lib or /usr/include paths to LDFLAGS and CPPFLAGS
302 (respectively) when --with-ssl=/usr is used (patch based on FreeBSD).
304 - Added an explicit buffer limit check in msdosify() (patch based on FreeBSD).
305 This couldn't ever overflow in curl, but might if the code were used
306 elsewhere or under different conditions.
308 Daniel Stenberg (3 Feb 2009)
309 - Hidemoto Nakada provided a small fix that makes it possible to get the
310 CURLINFO_CONTENT_LENGTH_DOWNLOAD size from file:// "transfers" with
311 CURLOPT_NOBODY set true.
313 Daniel Stenberg (2 Feb 2009)
314 - Patrick Scott found a rather large memory leak when using the multi
315 interface and setting CURLMOPT_MAXCONNECTS to something less than the number
316 of handles you add to the multi handle. All the connections that didn't fit
317 in the cache would not be properly disconnected nor freed!
319 - Craig A West brought us: libcurl now defaults to do CONNECT with HTTP
320 version 1.1 instead of 1.0 like before. This change also introduces the new
321 proxy type for libcurl called 'CURLPROXY_HTTP_1_0' that then allows apps to
322 switch (back) to CONNECT 1.0 requests. The curl tool also got a --proxy1.0
323 option that works exactly like --proxy but sets CURLPROXY_HTTP_1_0.
325 I updated all test cases cases that use CONNECT and I tried to do some using
326 --proxy1.0 and some updated to do CONNECT 1.1 to get both versions run.
328 Daniel Stenberg (31 Jan 2009)
329 - When building with c-ares 1.6.1 (not yet released) or later and IPv6 support
330 enabled, we can now take advantage of its brand new AF_UNSPEC support in
331 ares_gethostbyname(). This makes test case 241 finally run fine for me with
332 this setup since it now parses the "::1 ip6-localhost" line fine in my
335 Daniel Stenberg (30 Jan 2009)
336 - Scott Cantor filed bug report #2550061
337 (http://curl.haxx.se/bug/view.cgi?id=2550061) mentioning that I failed to
338 properly make sure that the VC9 makefiles got included in the latest
339 release. I've now fixed the release script and verified it so next release
340 will hopefully include them properly!
342 Daniel Fandrich (30 Jan 2009)
343 - Fixed --disable-proxy for FTP and SOCKS. Thanks to Daniel Egger for
346 Yang Tse (29 Jan 2009)
347 - Introduced curl_sspi.c and curl_sspi.h for the implementation of functions
348 Curl_sspi_global_init() and Curl_sspi_global_cleanup() which previously were
349 named Curl_ntlm_global_init() and Curl_ntlm_global_cleanup() in http_ntlm.c
350 Also adjusted socks_sspi.c to remove the link-time dependency on the Windows
351 SSPI library using it now in the same way as it was done in http_ntlm.c.
353 Daniel Stenberg (28 Jan 2009)
354 - Markus Moeller introduced two new options to libcurl:
355 CURLOPT_SOCKS5_GSSAPI_SERVICE and CURLOPT_SOCKS5_GSSAPI_NEC to allow libcurl
356 to do GSS-style authentication with SOCKS5 proxies. The curl tool got the
357 options called --socks5-gssapi-service and --socks5-gssapi-nec to enable
360 Daniel Stenberg (26 Jan 2009)
361 - Chad Monroe provided the new CURLOPT_TFTP_BLKSIZE option that allows an app
362 to set desired block size to use for TFTP transfers instead of the default
365 - The "-no_ticket" option was introduced in Openssl0.9.8j. It's a flag to
366 disable "rfc4507bis session ticket support". rfc4507bis was later turned
367 into the proper RFC5077 it seems: http://tools.ietf.org/html/rfc5077
369 The enabled extension concerns the session management. I wonder how often
370 libcurl stops a connection and then resumes a TLS session. also, sending the
371 session data is some overhead. .I suggest that you just use your proposed
372 patch (which explicitly disables TICKET).
374 If someone writes an application with libcurl and openssl who wants to
375 enable the feature, one can do this in the SSL callback.
377 Sharad Gupta brought this to my attention. Peter Sylvester helped me decide
378 on the proper action.
380 - Alexey Borzov filed bug report #2535504
381 (http://curl.haxx.se/bug/view.cgi?id=2535504) pointing out that realms with
382 quoted quotation marks in HTTP Digest headers didn't work. I've now added
383 test case 1095 that verifies my fix.
385 - Craig A West brought CURLOPT_NOPROXY and the corresponding --noproxy option.
386 They basically offer the same thing the NO_PROXY environment variable only
387 offered previously: list a set of host names that shall not use the proxy
388 even if one is specified.
390 Daniel Fandrich (20 Jan 2009)
391 - Call setlocale() for libtest tests to test the effects of locale-induced
392 libc changes on libcurl.
394 - Fixed a couple more locale-dependent toupper conversions, mainly for
395 clarity. This does fix one problem that causes ;type=i FTP URLs
396 to fail in the Turkish locale when CURLOPT_PROXY_TRANSFER_MODE is
399 - Added tests 561 and 1091 through 1094 to test various combinations
400 of ;type= and ;mode= URLs that could potentially fail in the Turkish
403 Daniel Stenberg (20 Jan 2009)
404 - Lisa Xu pointed out that the ssh.obj file was missing from the
405 lib/Makefile.vc6 file (and thus from the vc8 and vc9 ones too).
407 Version 7.19.3 (19 January 2009)
409 Daniel Stenberg (16 Jan 2009)
410 - Andrew de los Reyes fixed curlbuild.h for "generic" gcc builds on PPC, both
413 Daniel Stenberg (15 Jan 2009)
414 - Tim Ansell fixed a compiler warning in lib/cookie.c
416 Daniel Stenberg (14 Jan 2009)
417 - Grant Erickson fixed timeouts for TFTP such that specifying a
418 connect-timeout, a max-time or both options work correctly and as expected
419 by passing the correct boolean value to Curl_timeleft via the
420 'duringconnect' parameter.
422 With this small change, curl TFTP now behaves as expected (and likely as
423 originally-designed):
425 1) For non-existent or unreachable dotted IP addresses:
427 a) With no options, follows the default curl 300s timeout...
428 b) With --connect-timeout only, follows that value...
429 c) With --max-time only, follows that value...
430 d) With both --connect-timeout and --max-time, follows the smaller value...
432 and times out with a "curl: (7) Couldn't connect to server" error.
434 2) For transfers to/from a valid host:
436 a) With no options, follows default curl 300s timeout for the
437 first XRQ/DATA/ACK transaction and the default TFTP 3600s
438 timeout for the remainder of the transfer...
440 b) With --connect-time only, follows that value for the
441 first XRQ/DATA/ACK transaction and the default TFTP 3600s
442 timeout for the remainder of the transfer...
444 c) With --max-time only, follows that value for the first
445 XRQ/DATA/ACK transaction and for the remainder of the
448 d) With both --connect-timeout and --max-time, follows the former
449 for the first XRQ/DATA/ACK transaction and the latter for the
450 remainder of the transfer...
452 and times out with a "curl: (28) Timeout was reached" error as
455 Daniel Stenberg (13 Jan 2009)
456 - Michael Wallner fixed a NULL pointer deref when calling
457 curl_easy_setup(curl, CURLOPT_COOKIELIST, "SESS") on a CURL handle with no
460 - Stefan Teleman brought a patch to fix the default curlbuild.h file for the
463 Daniel Stenberg (12 Jan 2009)
464 - Based on bug report #2498665 (http://curl.haxx.se/bug/view.cgi?id=2498665)
465 by Daniel Black, I've now added magic to the configure script that makes it
466 use pkg-config to detect gnutls details as well if the existing method
467 (using libgnutls-config) fails. While doing this, I cleaned up and unified
468 the pkg-config usage when detecting openssl and nss as well.
470 Daniel Stenberg (11 Jan 2009)
471 - Karl Moerder brought the patch that creates vc9 Makefiles, and I made
472 'maketgz' now use the actual makefile targets to do the VC8 and VC9
475 Daniel Stenberg (10 Jan 2009)
476 - Emil Romanus fixed:
478 When using the multi interface over HTTP and the server returns a Location
479 header, the running easy handle will get stuck in the CURLM_STATE_PERFORM
480 state, leaving the external event loop stuck waiting for data from the
481 ingoing socket (when using the curl_multi_socket_action stuff). While this
482 bug was pretty hard to find, it seems to require only a one-line fix. The
483 break statement on line 1374 in multi.c caused the function to skip the call
486 How to reproduce this bug? Well, that's another question. evhiperfifo.c in
487 the examples directory chokes on this bug only _sometimes_, probably
488 depending on how fast the URLs are added. One way of testing the bug out is
489 writing to hiper.fifo from more than one source at the same time.
491 Daniel Fandrich (7 Jan 2009)
492 - Unified much of the SessionHandle initialization done in Curl_open() and
493 curl_easy_reset() by creating Curl_init_userdefined(). This had the side
494 effect of fixing curl_easy_reset() so it now also resets
495 CURLOPT_FTP_FILEMETHOD and CURLOPT_SSL_SESSIONID_CACHE
497 Daniel Stenberg (7 Jan 2009)
498 - Rob Crittenden did once again provide an NSS update:
500 I have to jump through a few hoops now with the NSS library initialization
501 since another part of an application may have already initialized NSS by the
502 time Curl gets invoked. This patch is more careful to only shutdown the NSS
503 library if Curl did the initialization.
505 It also adds in a bit of code to set the default ciphers if the app that
506 call NSS_Init* did not call NSS_SetDomesticPolicy() or set specific
507 ciphers. One might argue that this lets other application developers get
508 lazy and/or they aren't using the NSS API correctly, and you'd be right.
509 But still, this will avoid terribly difficult-to-trace crashes and is
512 Daniel Stenberg (1 Jan 2009)
513 - 'reconf' is removed since we rather have users use 'buildconf'
515 Daniel Stenberg (31 Dec 2008)
516 - Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
517 out that 'reconf' didn't properly point out the m4 subdirectory when running
520 Daniel Stenberg (29 Dec 2008)
521 - Phil Lisiecki filed bug report #2413067
522 (http://curl.haxx.se/bug/view.cgi?id=2413067) that identified a problem that
523 would cause libcurl to mark a DNS cache entry "in use" eternally if the
524 subsequence TCP connect failed. It would thus never get pruned and refreshed
525 as it should've been.
527 Phil provided his own patch to this problem that while it seemed to work
528 wasn't complete and thus I wrote my own fix to the problem.
530 Daniel Stenberg (28 Dec 2008)
531 - Peter Korsgaard fixed building libcurl with "configure --with-ssl
534 - Anthony Bryan fixed more language and spelling flaws in man pages.
536 Daniel Stenberg (22 Dec 2008)
537 - Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
538 on file indexes beyond 2 or 4GB.
540 - Anthony Bryan provided a set of patches that cleaned up manual language,
541 corrected spellings and more.
543 Daniel Stenberg (20 Dec 2008)
544 - Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
545 pipelining, as libcurl could then easily get confused and A) work on the
546 handle that was not "first in queue" on a pipeline, or even B) tell the app
547 to REMOVE a socket while it was in use by a second handle in a pipeline. Both
548 errors caused hanging or stalling applications.
550 Daniel Stenberg (19 Dec 2008)
551 - curl_multi_timeout() could return a timeout value of 0 even though nothing
552 was actually ready to get done, as the internal time resolution is higher
553 than the returned millisecond timer. Therefore it could cause applications
554 running on fast processors to do short bursts of busy-loops.
555 curl_multi_timeout() will now only return 0 if the timeout is actually
558 - Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
559 now has an improved ability to do right when the multi interface (both
560 "regular" and multi_socket) is used for SCP and SFTP transfers. This should
561 result in (much) less busy-loop situations and thus less CPU usage with no
564 Daniel Stenberg (17 Dec 2008)
565 - SCP and SFTP with the multi interface had the same flaw: the 'DONE'
566 operation didn't complete properly if the EAGAIN equivalent was returned but
567 libcurl would simply continue with a half-completed close operation
568 performed. This ruined persistent connection re-use and cause some
569 SSH-protocol errors in general. The correction is unfortunately adding a
570 blocking function - doing it entirely non-blocking should be considered for
573 Gisle Vanem (16 Dec 2008)
574 - Added the possibility to use the Watt-32 tcp/ip stack under Windows.
575 The change simply involved adding a USE_WATT32 section in the
576 config-win32.h files (under ./lib and ./src). This section disables
577 the use of any Winsock headers.
579 Daniel Stenberg (16 Dec 2008)
580 - libssh2_sftp_last_error() was wrongly used at some places in libcurl which
581 made libcurl sometimes not properly abort problematic SFTP transfers.
583 Daniel Stenberg (12 Dec 2008)
584 - More work with Igor Novoseltsev to first fix the remaining stuff for
585 removing easy handles from multi handles when the easy handle is/was within
586 a HTTP pipeline. His bug report #2351653
587 (http://curl.haxx.se/bug/view.cgi?id=2351653) was also related and was
588 eventually fixed by a patch by Igor himself.
590 Yang Tse (12 Dec 2008)
591 - Patrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
592 OS/400 compilations with IPv6 enabled.
594 Daniel Stenberg (12 Dec 2008)
595 - Mark Karpeles filed bug report #2416182 titled "crash in ConnectionExists
596 when using duphandle+curl_mutli"
597 (http://curl.haxx.se/bug/view.cgi?id=2416182) which showed that
598 curl_easy_duphandle() wrongly also copied the pointer to the connection
599 cache, which was plain wrong and caused a segfault if the handle would be
600 used in a different multi handle than the handle it was duplicated from.
602 Daniel Stenberg (11 Dec 2008)
603 - Keshav Krity found out that libcurl failed to deal with dotted IPv6
604 addresses if they were very long (>39 letters) due to a too strict address
605 validity parser. It now accepts addresses up to 45 bytes long.
607 Daniel Stenberg (11 Dec 2008)
608 - Internet Explorer had a broken HTTP digest authentication before v7 and
609 there are servers "out there" that relies on the client doing this broken
610 Digest authentication. Apache even comes with an option to work with such
613 The difference is only for URLs that contain a query-part (a '?'-letter and
614 text to the right of it).
616 libcurl now supports this quirk, and you enable it by setting the
617 CURLAUTH_DIGEST_IE bit in the bitmask you pass to the CURLOPT_HTTPAUTH or
618 CURLOPT_PROXYAUTH options. They are thus individually controlled to server
621 (note that there's no way to activate this with the curl tool yet)
623 Daniel Fandrich (9 Dec 2008)
624 - Added test cases 1089 and 1090 to test --write-out after a redirect to
625 test a report that the size didn't work, but these test cases pass.
627 - Documented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.
629 Daniel Stenberg (9 Dec 2008)
630 - Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
631 particular state for the control connection like it did before for implicit
632 FTPS (libcurl assumed such control connections to be encrypted while some
633 FTPS servers such as FileZilla assumes such connections to be clear
634 mode). Use the CURLOPT_USE_SSL option to set your desired level.
636 Daniel Stenberg (8 Dec 2008)
637 - Fred Machado posted about a weird FTP problem on the curl-users list and when
638 researching it, it turned out he got a 550 response back from a SIZE command
639 and then I fell over the text in RFC3659 that says:
641 The presence of the 550 error response to a SIZE command MUST NOT be taken
642 by the client as an indication that the file cannot be transferred in the
643 current MODE and TYPE.
645 In other words: the change I did on September 30th 2008 and that has been
646 included in the last two releases were a regression and a bad idea. We MUST
647 NOT take a 550 response from SIZE as a hint that the file doesn't exist.
649 - Christian Krause filed bug #2221237
650 (http://curl.haxx.se/bug/view.cgi?id=2221237) that identified an infinite
651 loop during GSS authentication given some specific conditions. With his
652 patience and great feedback I managed to narrow down the problem and
653 eventually fix it although I can't test any of this myself!
655 Daniel Fandrich (3 Dec 2008)
656 - Fixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
659 Daniel Stenberg (3 Dec 2008)
660 - Igor Novoseltsev filed bug #2351645
661 (http://curl.haxx.se/bug/view.cgi?id=2351645) that identified a problem with
662 the multi interface that occured if you removed an easy handle while in
663 progress and the handle was used in a HTTP pipeline.
665 - Pawel Kierski pointed out a mistake in the cookie code that could lead to a
666 bad fclose() after a fatal error had occured.
667 (http://curl.haxx.se/bug/view.cgi?id=2382219)
669 Daniel Fandrich (25 Nov 2008)
670 - If a HTTP request is Basic and num is already >=1000, the HTTP test
671 server adds 1 to num to get the data section to return. This allows
672 testing authentication negotiations using the Basic authentication
675 - Added tests 1087 and 1088 to test Basic authentication on a redirect
676 with and without --location-trusted
678 Daniel Stenberg (24 Nov 2008)
679 - Based on a patch by Vlad Grachov, libcurl now uses a new libssh2 0.19
680 function when built to support SCP and SFTP that helps the library to know
681 in which direction a particular libssh2 operation would return EAGAIN so
682 that libcurl knows what socket conditions to wait for before trying the
683 function call again. Previously (and still when using libssh2 0.18 or
684 earlier), libcurl will busy-loop in this situation when the easy interface
687 Daniel Fandrich (20 Nov 2008)
688 - Automatically detect OpenBSD's CA cert bundle.
690 Daniel Stenberg (19 Nov 2008)
691 - I removed the default use of "Pragma: no-cache" from libcurl when a proxy is
692 used. It has been used since forever but it was never a good idea to use
693 unless explicitly asked for.
695 - Josef Wolf's extension that allows a $TESTDIR/gdbinit$testnum file that when
696 you use runtests.pl -g, will be sourced by gdb to allow additional fancy or
699 - Christian Krause reported and fixed a memory leak that would occur with HTTP
700 GSS/kerberos authentication (http://curl.haxx.se/bug/view.cgi?id=2284386)
702 - Andreas Wurf and Markus Koetter helped me analyze a problem that Andreas got
703 when uploading files to a single FTP server using multiple easy handle
704 handles with the multi interface. Occasionally a handle would stall in
707 The problem turned out to be a side-effect of the ConnectionExists()
708 function's eagerness to re-use a handle for HTTP pipelining so it would
709 select it even if already being in use, due to an inadequate check for its
710 chances of being used for pipelnining.
712 Daniel Fandrich (17 Nov 2008)
713 - Added more compiler warning options for gcc 4.3
715 Yang Tse (17 Nov 2008)
716 - Fix a remaining problem in the inet_pton() runtime configure check. And
717 fix internal Curl_inet_pton() failures to reject certain malformed literals.
719 - Make configure script check if ioctl with the SIOCGIFADDR command can be
720 used, and define HAVE_IOCTL_SIOCGIFADDR if appropriate.
722 Daniel Stenberg (16 Nov 2008)
723 - Christian Krause fixed a build failure when building with gss support
724 enabled and FTP disabled.
726 - Added check for NULL returns from strdup() in src/main.c and lib/formdata.c
727 - reported by Jim Meyering also prevent buffer overflow on MSDOS when you do
728 for example -O on a url with a file name part longer than PATH_MAX letters
730 - lib/nss.c fixes based on the report by Jim Meyering: I went over and added
731 checks for return codes for all calls to malloc and strdup that were
732 missing. I also changed a few malloc(13) to use arrays on the stack and a
733 few malloc(PATH_MAX) to instead use aprintf() to lower memory use.
735 - I fixed a memory leak in Curl_nss_connect() when CURLOPT_ISSUERCERT is
738 Daniel Fandrich (14 Nov 2008)
739 - Added .xml as one of the few common file extensions known by the multipart
742 - Added some #ifdefs around header files and change the EAGAIN test to
743 fix compilation on Cell (reported by Jeff Curley).
745 Yang Tse (14 Nov 2008)
746 - Fixed several configure script issues affecting checks for inet_ntoa_r(),
747 inet_ntop(), inet_pton(), getifaddrs(), fcntl() and getaddrinfo().
749 Yang Tse (13 Nov 2008)
750 - Refactored configure script detection of functions used to set sockets into
751 non-blocking mode, and decouple function detection from function capability.