5 \___|\___/|_| \_\_____|
9 Kamil Dudka (27 May 2010)
10 - Tor Arntsen changed the alternative definition of bool to use enum instead
13 Daniel Stenberg (25 May 2010)
14 - Julien Chaffraix fixed the warning seen when compiling lib/rtmp.c: one
15 unused variables, several unused arguments and some missing #include.
17 - Julien Chaffraix fixed 2 OOM errors: a missing NULL-check in
18 lib/http_negociate.c and a potential NULL dereferencing in lib/splay.c
20 - Howard Chu brought a patch that makes the LDAP code much cleaner, nicer and
21 in general being a better libcurl citizen. If a new enough OpenLDAP version
22 is detect, the new and shiny lib/openldap.c code is then used instead of the
25 Daniel Stenberg (21 May 2010)
26 - Eric Mertens posted bug #3003705: when we made TFTP use the correct timeout
27 option when sent to the server (fixed May 18th 2010) it became obvious that
28 libcurl used invalid timeout values (300 by default while the RFC allows
29 nothing above 255). While of course it is obvious that as TFTP has worked
30 thus far without being able to set timeout at all, just removing the setting
31 wouldn't make any difference in behavior. I decided to still keep it (but
32 fix the problem) as it now actually allows for easier (future) customization
35 (http://curl.haxx.se/bug/view.cgi?id=3003705)
37 - Douglas Kilpatrick filed bug report #3004787 and pointed out that the TFTP
38 code didn't handle block id wraps correctly. His suggested fix inspired the
41 (http://curl.haxx.se/bug/view.cgi?id=3004787)
43 Daniel Stenberg (20 May 2010)
44 - Tanguy Fautre brought a fix to allow curl to build with Microsoft VC10.
46 Daniel Stenberg (18 May 2010)
47 - Eric Mertens posted bug report #3003005 pointing out that the libcurl TFTP
48 code was not sending the timeout option properly to the server, and
51 (http://curl.haxx.se/bug/view.cgi?id=3003005)
53 Kamil Dudka (16 May 2010)
54 - Pavel Raiskup introduced a new option CURLOPT_FNMATCH_DATA in order to pass
55 a custom data pointer to the callback specified by CURLOPT_FNMATCH_FUNCTION.
57 Daniel Stenberg (14 May 2010)
58 - John-Mark Bell filed bug #3000052 that identified a problem (with an
59 associated patch) with the OpenSSL handshake state machine when the multi
62 Performing an https request using a curl multi handle and using select or
63 epoll to wait for events results in a hang. It appears that the cause is the
64 fix for bug #2958179, which makes ossl_connect_common unconditionally return
65 from the step 2 loop when fetching from a multi handle.
67 When ossl_connect_step2 has completed, it updates connssl->connecting_state
68 to ssl_connect_3. ossl_connect_common will then return to the caller, as a
69 multi handle is in use. Eventually, the client code will call
70 curl_multi_fdset to obtain an updated fdset to select or epoll on. For https
71 requests, curl_multi_fdset will cause https_getsock to be called.
72 https_getsock will only return a socket handle if the connecting_state is
73 ssl_connect_2_reading or ssl_connect_2_writing. Therefore, the client will
74 never obtain a valid fdset, and thus not drive the multi handle, resulting
77 (http://curl.haxx.se/bug/view.cgi?id=3000052)
79 - Sebastian V reported bug #3000056 identifying a problem with redirect
80 following. It showed that when curl followed redirects it didn't properly
81 ignore the response body of the 30X response if that response was using
82 compressed Content-Encoding!
84 (http://curl.haxx.se/bug/view.cgi?id=3000056)
86 Daniel Stenberg (12 May 2010)
87 - Howard Chu brought support for RTMP. This is powered by the underlying
88 librtmp library. It supports a range of variations and "sub-protocols"
89 within the RTMP family.
91 - Pavel Raiskup brought support for FTP directory wildcard matching to allow
92 selective downloading. To provide that, a set of new options were added:
95 CURLOPT_CHUNK_BGN_FUNCTION
96 CURLOPT_CHUNK_END_FUNCTION
98 CURLOPT_FNMATCH_FUNCTION
100 There were also a set of new tests added (574 - 577) to verify this.
102 Kamil Dudka (11 May 2010)
103 - CRL support in libcurl-NSS has been completely broken. Now it works. Original
104 bug report: https://bugzilla.redhat.com/581926
106 Daniel Stenberg (7 May 2010)
107 - Dirk Manske reported a regression. When connecting with the multi interface,
108 there were situations where libcurl wouldn't store connect time correctly as
109 it used to (and is documented to) do.
111 Using his fine sample program we could repeat it, and I wrote up test case
112 573 using that code. The problem does not easily show itself using the local
115 The fix, also as suggested by Dirk, is a bit on the ugly side as it adds yet
116 another call to Curl_verboseconnect() and setting the TIMER_CONNECT time.
117 That situation is subject for some closer inspection in the future.
119 - Howard Chu split the I/O handling functions into private handlers.
121 Howard Chu brought the bulk work of this patch that properly moves out the
122 sending and recving of data to the parts of the code that are properly
123 responsible for the various ways of doing so.
125 Daniel Stenberg assisted with polishing a few bits and fixed some minor
126 flaws in the original patch.
128 Another upside of this patch is that we now abuse CURLcodes less with the
129 "magic" -1 return codes and instead use CURLE_AGAIN more consistently.
131 Daniel Stenberg (5 May 2010)
132 - Hoi-Ho Chan introduced support for using the PolarSSL library. You control
133 this with the new configure option --with-polarssl.
135 Daniel Stenberg (29 Apr 2010)
136 - Ben Greear made telnet a lot better/easier to use by an application:
138 The main change is to allow input from user-specified methods, when they are
139 specified with CURLOPT_READFUNCTION. All calls to fflush(stdout) in
140 telnet.c were removed, which makes using 'curl telnet://foo.com' painful
141 since prompts and other data are not always returned to the user promptly.
142 Use 'curl --no-buffer telnet://foo.com' instead. In general, the user
143 should have their CURLOPT_WRITEFUNCTION do a fflush for interactive use.
145 Also fix assumption that reading from stdin never returns < 0.
146 Old code could crash in that case.
148 Call progress functions in telnet main loop.
150 Daniel Stenberg (26 Apr 2010)
151 - Make use of the libssh2_init/exit functions that libssh2 added in version
152 1.2.5. Using them will improve how libcurl works in threaded situations when
153 SCP and SFTP are transfered.
155 Daniel Stenberg (25 Apr 2010)
156 - Based on work by Kamil Dudka, I've introduced the new configure option
157 --enable-threaded-resolver. When used, the configure script will check for
158 pthreads and if around, it will build libcurl to use pthreads to do name
159 resolving in a threaded manner. Note that this is just a fix to offer an
160 option that can enable the code that already included. The threader resolver
161 code was mostly added on Jan 26 2010.
163 Daniel Stenberg (24 Apr 2010)
164 - Alex Bligh introduced the --proto and -proto-redir options that limit what
165 protocols curl accepts for the requests and when following redirects.
167 Kamil Dudka (24 Apr 2010)
168 - Fixed test536 in order to not fail with threaded DNS resolver and tweaked
169 comments in certain examples using curl_multi_fdset().
171 - Fixed SSL handshake timeout underflow in libcurl-NSS, which caused test405
172 to hang on a slow machine.
174 Daniel Stenberg (21 Apr 2010)
175 - The -O option caused curl to crash on windows and DOS due to the tool
176 writing out of boundary memory.
178 Yang Tse (20 Apr 2010)
179 - Ruslan Gazizov detected that MSVC makefiles were using wsock32.lib instead
180 of ws2_32.lib, this generated linking issues on MSVC IPv6 enabled builds
181 that were done using those makefiles.
183 Daniel Stenberg (19 Apr 2010)
184 - -J/--remote-header-name didn't strip trailing carriage returns or linefeeds
185 properly, so they could be used in the file name.
187 Daniel Stenberg (16 Apr 2010)
188 - Jerome Vouillon made the GnuTLS SSL handshake phase non-blocking.
190 - The recent overhaul of the SSL recv function made the GnuTLS specific code
191 treat a zero returned from gnutls_record_recv() as an error, and this caused
192 our HTTPS test cases to fail. We leave it to upper layer code to detect if
193 an EOF is a problem or not.
195 - I reverted the resolver fix from yesterday and instead removed all uses of
196 AI_CANONNAME all over libcurl and made the only user of that info (krb5.c)
197 use the host name from the URL instead. No reverse resolving is a good
200 - Paul Howarth made configure properly detect GSS "on ancient Linux distros"
201 by editing in which order we use headers to detect GSS.
203 Daniel Stenberg (15 Apr 2010)
204 - Rainer Canavan filed bug report #2987196 that identified libcurl doing
205 unnecesary reverse name lookups in many cases when built to use IPv4 and
206 getaddrinfo(). The logic for ipv6 is now used for ipv4 too.
208 (http://curl.haxx.se/bug/view.cgi?id=2963679)
210 Version 7.20.1 (14 April 2010)
212 Daniel Stenberg (9 Apr 2010)
213 - Prefixing the FTP quote commands with an asterisk really only worked for the
214 postquote actions. This is now fixed and test case 227 has been extended to
217 Kamil Dudka (4 Apr 2010)
218 - Eliminated a race condition in Curl_resolv_timeout().
220 - Refactorized interface of Curl_ssl_recv()/Curl_ssl_send().
222 - libcurl-NSS now provides more accurate messages and error codes in case of
223 client certificate problem. Either during connection, or transfer phase.
225 Daniel Stenberg (1 Apr 2010)
226 - Matt Wixson found and fixed a bug in the SCP/SFTP area where the code
227 treated a 0 return code from libssh2 to be the same as EAGAIN while in
228 reality it isn't. The problem caused a hang in SFTP transfers from a
231 Daniel Stenberg (28 Mar 2010)
232 - Ben Greear: If you pass a URL to pop3 that does not contain a message ID as
233 part of the URL, it would previously ask for 'INBOX' which just causes the
234 pop3 server to return an error.
236 Now libcurl treats en empty message ID as a request for LIST (list of pop3
237 message IDs). User's code could then parse this and download individual
240 Daniel Stenberg (27 Mar 2010)
241 - Ben Greear brought a patch that from now on allows all protocols to specify
242 name and user within the URL, in the same manner HTTP and FTP have been
243 allowed to in the past - although far from all of the libcurl supported
244 protocls actually have that feature in their URL definition spec.
246 Daniel Stenberg (26 Mar 2010)
247 - Ben Greear brought code that makes the rate limiting code for the easy
248 interface a bit smoother as it introduces sub-second sleeps during it and it
249 also takes the buffer sizes into account.
251 Daniel Stenberg (24 Mar 2010)
252 - Bob Richmond: There's an annoying situation where libcurl will read new HTTP
253 response data from a socket, then check if it's a timeout if one is set. If
254 the last packet received constitutes the end of the response body, libcurl
255 still treats it as a timeout condition and reports a message like:
257 "Operation timed out after 3000 milliseconds with 876 out of 876 bytes
260 It should only a timeout if the timer lapsed and we DIDN'T receive the end
261 of the response body yet.
263 - Christopher Conroy fixed a problem with RTSP and GET_PARAMETER reported
264 to us by Massimo Callegari. There's a new test case 572 that verifies this
267 - The 'ares' subtree has been removed from the source repository. It was
268 always a separate project that sort of piggybacked on the curl project since
269 the dawn of times and now the time has come for it to go stand on its own
270 legs and continue living its own life. All details on c-ares and its new
271 source code repository is found at http://c-ares.haxx.se/
273 Daniel Stenberg (23 Mar 2010)
274 - Kenny To filed the bug report #2963679 with patch to fix a problem he
275 experienced with doing multi interface HTTP POST over a proxy using
276 PROXYTUNNEL. He found a case where it would connect fine but bits.tcpconnect
277 was not set correct so libcurl didn't work properly.
279 (http://curl.haxx.se/bug/view.cgi?id=2963679)
281 - Akos Pasztory filed debian bug report #572276
282 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=572276 mentioning a problem
283 with a resource that returns chunked-encoded _and_ with a Content-Length
284 and libcurl failed to properly ignore the latter information.
286 - Hauke Duden provided an example program that made the multi interface crash.
287 His example simply used the multi interface and did first one FTP transfer
288 and after completion it used a second easy handle and did another FTP
289 transfer on the same FTP server.
291 This triggered a bug in the "delayed easy handle kill" system that curl
292 uses: when an FTP connection is left alive it must keep an easy handle
293 around internally - only for the purpose of having an easy handle when it
294 later disconnects it. The code assumed that when the easy handle was removed
295 and an internal reference was made, that version could be killed later on
296 when a new easy handle came using the same connection. This was wrong as
297 Hauke's example showed that the removed handle wasn't killed for real until
298 later. This caused a double close attempt => segfault.
300 Daniel Stenberg (22 Mar 2010)
301 - Thomas Lopatic fixed the alarm()-based DNS timeout:
303 Looking at the code of Curl_resolv_timeout() in hostip.c, I think that in
304 case of a timeout, the signal handler for SIGALRM never gets removed. I
305 think that in my case it gets executed at some point later on when execution
306 has long left Curl_resolv_timeout() or even the cURL library.
308 The code that is jumped to with siglongjmp() simply sets the error message
309 to "name lookup timed out" and then returns with CURLRESOLV_ERROR. I guess
310 that instead of simply returning without cleaning up, the code should have a
311 goto that jumps to the spot right after the call to Curl_resolv().
313 Kamil Dudka (22 Mar 2010)
314 - Douglas Steinwand contributed a patch fixing insufficient initialization in
315 Curl_clone_ssl_config()
317 Daniel Stenberg (21 Mar 2010)
318 - Ben Greear improved TFTP: the error code returning and the treatment
319 of TSIZE == 0 when uploading.
321 - We've switched from CVS to git. See http://curl.haxx.se/source.html
323 Kamil Dudka (19 Mar 2010)
324 - Improved Curl_read() to not ignore the error returned from Curl_ssl_recv().
326 Daniel Stenberg (15 Mar 2010)
327 - Constantine Sapuntzakis brought a patch:
329 The problem mentioned on Dec 10 2009
330 (http://curl.haxx.se/bug/view.cgi?id=2905220) was only partially fixed.
331 Partially because an easy handle can be associated with many connections in
332 the cache (e.g. if there is a redirect during the lifetime of the easy
333 handle). The previous patch only cleaned up the first one. The new fix now
334 removes the easy handle from all connections, not just the first one.
336 Daniel Stenberg (6 Mar 2010)
337 - Ben Greear brought a patch that fixed the rate limiting logic for TFTP when
338 the easy interface was used.
340 Daniel Stenberg (5 Mar 2010)
341 - Daniel Johnson provided fixes for building curl with the clang compiler.
343 Yang Tse (5 Mar 2010)
344 - Constantine Sapuntzakis detected and fixed a double free in builds done
345 with threaded resolver enabled (Windows default configuration) that would
346 get triggered when a curl handle is closed while doing DNS resolution.
348 Daniel Stenberg (2 Mar 2010)
349 - [Daniel Johnson] I've been trying to build libcurl with clang on Darwin and
350 ran into some issues with the GSSAPI tests in configure.ac. The tests first
351 try to determine the include dirs and libs and set CPPFLAGS and LIBS
352 accordingly. It then checks for the headers and finally sets LIBS a second
353 time, causing the libs to be included twice. The first setting of LIBS seems
354 redundant and should be left out, since the first part is otherwise just
355 about finding headers.
357 My second issue is that 'krb5-config --libs gssapi' on Darwin is less than
358 useless and returns junk that, while it happens to work with gcc, causes
359 clang to choke. For example, --libs returns $CFLAGS along with the libs,
360 which is really retarded. Simply setting 'LIBS="$LIBS -lgssapi_krb5
361 -lresolv"' on Darwin is sufficient.
363 - Based on patch provided by Jacob Moshenko, the transfer logic now properly
364 makes sure that when using sub-second timeouts, there's no final bad 1000ms
365 wait. Previously, a sub-second timeout would often make the elapsed time end
366 up the time rounded up to the nearest second (e.g. 1s for 200ms timeout)
368 - Andrei Benea filed bug report #2956698 and pointed out that the
369 CURLOPT_CERTINFO feature leaked memory due to a missing OpenSSL function
370 call. He provided the patch to fix it too.
372 http://curl.haxx.se/bug/view.cgi?id=2956698
374 - Markus Duft pointed out in bug #2961796 that even though Interix has a
375 poll() function it doesn't quite work the way we want it so we must disable
376 it, and he also provided a patch for it.
378 http://curl.haxx.se/bug/view.cgi?id=2961796
380 - Made the pingpong timeout code properly deal with the response timeout AND
381 the global timeout if set. Also, as was reported in the bug report #2956437
382 by Ryan Chan, the time stamp to use as basis for the per command timeout was
383 not set properly in the DONE phase for FTP (and not for SMTP) so I fixed
384 that just now. This was a regression compared to 7.19.7 due to the
385 conversion of FTP code over to the generic pingpong concepts.
387 http://curl.haxx.se/bug/view.cgi?id=2956437
389 Daniel Stenberg (1 Mar 2010)
390 - Ben Greear provided an update for TFTP that fixes upload.
392 - Wesley Miaw reported bug #2958179 which identified a case of looping during
393 OpenSSL based SSL handshaking even though the multi interface was used and
394 there was no good reason for it.
396 http://curl.haxx.se/bug/view.cgi?id=2958179
398 Daniel Stenberg (26 Feb 2010)
399 - Pat Ray in bug #2958474 pointed out an off-by-one case when receiving a
400 chunked-encoding trailer.
402 http://curl.haxx.se/bug/view.cgi?id=2958474
404 Daniel Fandrich (25 Feb 2010)
405 - Fixed a couple of out of memory leaks and a segfault in the SMTP & IMAP code.
407 Yang Tse (25 Feb 2010)
408 - I fixed bug report #2958074 indicating
409 (http://curl.haxx.se/bug/view.cgi?id=2958074) that curl on Windows with
410 option --trace-time did not use local time when timestamping trace lines.
411 This could also happen on other systems depending on time souurce.
413 Patrick Monnerat (22 Feb 2010)
414 - Proper handling of STARTTLS on SMTP, taking CURLUSESSL_TRY into account.
415 - SMTP falls back to RFC821 HELO when EHLO fails (and SSL is not required).
416 - Use of true local host name (i.e.: via gethostname()) when available, as
417 default argument to SMTP HELO/EHLO.
418 - Test case 804 for HELO fallback.
420 Daniel Stenberg (20 Feb 2010)
421 - Fixed the SMTP compliance by making sure RCPT TO addresses are specified
422 properly in angle brackets. Recipients provided with CURLOPT_MAIL_RCPT now
423 get angle bracket wrapping automatically by libcurl unless the recipient
424 starts with an angle bracket as then the app is assumed to deal with that
427 - I made the SMTP code expect a 250 response back from the server after the
428 full DATA has been sent, and I modified the test SMTP server to also send
429 that response. As usual, the DONE operation that is made after a completed
430 transfer is still not doable in a non-blocking way so this waiting for 250
431 is unfortunately made blockingly.
433 Yang Tse (14 Feb 2010)
434 - Overhauled test suite getpart() function. Fixing potential out of bounds
435 stack and memory overwrites triggered with huge test case definitions.
437 Daniel Stenberg (13 Feb 2010)
438 - Martin Hager reported and fixed a problem with a missing quote in libcurl.m4
440 (http://curl.haxx.se/bug/view.cgi?id=2951319)
442 - Tom Donovan fixed the CURL_FORMAT_* defines when building with cmake.
444 (http://curl.haxx.se/bug/view.cgi?id=2951269)
446 Daniel Stenberg (12 Feb 2010)
447 - Jack Zhang reported a problem with SMTP: we wrongly used multiple addresses
448 in the same RCPT TO line, when they should be sent in separate single
449 commands. I updated test case 802 to verify this.
451 - I also fixed a bad use of my_setopt_str() of CURLOPT_MAIL_RCPT in the curl
452 tool which made it try to output it as string for the --libcurl feature
453 which could lead to crashes.
455 Yang Tse (11 Feb 2010)
456 - Steven M. Schweda fixed VMS builder bad behavior when used in a batch job,
457 removed obsolete batch_compile.com and defines.com and updated VMS readme.
459 Version 7.20.0 (9 February 2010)
461 Daniel Stenberg (9 Feb 2010)
462 - When downloading compressed content over HTTP and the app asked libcurl to
463 automatically uncompress it with the CURLOPT_ENCODING option, libcurl could
464 wrongly provide the callback with more data than the maximum documented
465 amount. An application could thus get tricked into badness if the maximum
466 limit was trusted to be enforced by libcurl itself (as it is documented).
468 This is further detailed and explained in the libcurl security advisory
471 http://curl.haxx.se/docs/adv_20100209.html
473 Daniel Fandrich (3 Feb 2010)
474 - Changed the Watcom makefiles to make them easier to keep in sync with
475 Makefile.inc since that can't be included directly.
477 Yang Tse (2 Feb 2010)
478 - Symbol CURL_FORMAT_OFF_T now obsoleted, will be removed in a future release,
479 symbol will not be available when building with CURL_NO_OLDIES defined. Use
480 of CURL_FORMAT_CURL_OFF_T is preferred since 7.19.0
482 Daniel Stenberg (1 Feb 2010)
483 - Using the multi_socket API, it turns out at times it seemed to "forget"
484 connections (which caused a hang). It turned out to be an existing (7.19.7)
485 bug in libcurl (that's been around for a long time) and it happened like
488 The app calls curl_multi_add_handle() to add a new easy handle, libcurl will
489 then set it to timeout in 1 millisecond so libcurl will tell the app about
492 The app's timeout fires off that there's a timeout, the app calls libcurl as
493 we so often document it:
496 res = curl_multi_socket_action(... TIMEOUT ...);
497 } while(CURLM_CALL_MULTI_PERFORM == res);
499 And this is the problem number one:
501 When curl_multi_socket_action() is called with no specific handle, but only
502 a timeout-action, it will *only* perform actions within libcurl that are
503 marked to run at this time. In this case, the request would go from INIT to
504 CONNECT and return CURLM_CALL_MULTI_PERFORM. When the app then calls libcurl
505 again, there's no timer set for this handle so it remains in the CONNECT
506 state. The CONNECT state is a transitional state in libcurl so it reports no
507 sockets there, and thus libcurl never tells the app anything more about that
508 easy handle/connection.
510 libcurl _does_ set a 1ms timeout for the handle at the end of
511 multi_runsingle() if it returns CURLM_CALL_MULTI_PERFORM, but since the loop
512 is instant the new job is not ready to run at that point (and there's no
513 code that makes libcurl call the app to update the timout for this new
514 timeout). It will simply rely on that some other timeout will trigger later
515 on or that something else will update the timeout callback. This makes the
516 bug fairly hard to repeat.
518 The fix made to adress this issue:
520 We introduce a loop in lib/multi.c around all calls to multi_runsingle() and
521 simply check for CURLM_CALL_MULTI_PERFORM internally. This has the added
522 benefit that this goes in line with my long-term wishes to get rid of the
523 CURLM_CALL_MULTI_PERFORM all together from the public API.
525 The downside of this fix, is that the counter we return in 'running_handles'
526 in several of our public functions then gets a slightly new and possibly
527 confusing behavior during times:
529 If an app adds a handle that fails to connect (very quickly) it may just
530 as well never appear as a 'running_handle' with this fix. Previously it
531 would first bump the counter only to get it decreased again at next call.
532 Even I have used that change in handle counter to signal "end of a
533 transfer". The only *good* way to find the end of a individual transfer
534 is calling curl_multi_info_read() to see if it returns one.
536 Of course, if the app previously did the looping before it checked the
537 counter, it really shouldn't be any new effect.
539 Yang Tse (26 Jan 2010)
540 - Constantine Sapuntzakis' and Joshua Kwan's work done in the last four months
541 relative to the asynchronous DNS lookups, along with with some integration
542 adjustments I have done are finally committed to CVS.
544 Currently these enhancements will benefit builds done using c-ares on any
545 platform as well as Windows builds using the default threaded resolver.
547 This release does not make generally available POSIX threaded DNS lookups
548 yet. There is no configure option to enable this feature yet. It is possible
549 to experimantally try this feature running configure with compiler flags that
550 make simultaneous definition of preprocessor symbols USE_THREADS_POSIX and
551 HAVE_PTHREAD_H, as well as whatever reentrancy compiler flags and linker ones
552 are required to link and properly use pthread_* functions on each platform.
554 Daniel Stenberg (26 Jan 2010)
555 - Mike Crowe made libcurl return CURLE_COULDNT_RESOLVE_PROXY when it is the
556 proxy that cannot be resolved when using c-ares. This matches the behaviour
557 when not using c-ares.
559 Björn Stenberg (23 Jan 2010)
560 - Added a new flag: -J/--remote-header-name. This option tells the
561 -O/--remote-name option to use the server-specified Content-Disposition
562 filename instead of extracting a filename from the URL.
564 Daniel Stenberg (21 Jan 2010)
565 - Chris Conroy brought support for RTSP transfers, and with it comes 8(!) new
566 libcurl options for controlling what to get and how to receive posssibly
567 interleaved RTP data.
569 Daniel Stenberg (20 Jan 2010)
570 - As was pointed out on the http-state mailing list, the order of cookies in a
571 HTTP Cookie: header _needs_ to be sorted on the path length in the cases
572 where two cookies using the same name are set more than once using
573 (overlapping) paths. Realizing this, identically named cookies must be
574 sorted correctly. But detecting only identically named cookies and take care
575 of them individually is harder than just to blindly and unconditionally sort
576 all cookies based on their path lengths. All major browsers also already do
577 this, so this makes our behavior one step closer to them in the cookie area.
579 Test case 8 was the only one that broke due to this change and I updated it
582 Daniel Stenberg (19 Jan 2010)
583 - David McCreedy brought a fix and a new test case (129) to make libcurl work
584 again when downloading files over FTP using ASCII and it turns out that the
585 final size of the file is not the same as the initial size the server
586 reported. This is very common since servers don't take the newline
587 conversions into account.
589 Kamil Dudka (14 Jan 2010)
590 - Suppressed side effect of OpenSSL configure checks, which prevented NSS from
591 being properly detected under certain circumstances. It had been caused by
592 strange behavior of pkg-config when handling PKG_CONFIG_LIBDIR. pkg-config
593 distinguishes among empty and non-existent environment variable in that case.
595 Daniel Stenberg (12 Jan 2010)
596 - Gil Weber reported a peculiar flaw with the multi interface when doing SFTP
597 transfers: curl_multi_fdset() would return -1 and not set and file
598 descriptors several times during a transfer of a single file. It turned out
599 to be due to two different flaws now fixed. Gil's excellent recipe helped me
602 Daniel Stenberg (11 Jan 2010)
603 - Made sure that the progress callback is repeatedly called at a regular
604 interval even during very slow connects.
606 - The tests/runtests.pl script now checks to see if the test case that runs is
607 present in the tests/data/Makefile.am and outputs a notice message on the
608 screen if not. Each test file has to be included in that Makefile.am to get
609 included in release archives and forgetting to add files there is a common
610 mistake. This is an attempt to make it harder to forget.
612 Daniel Stenberg (9 Jan 2010)
613 - Johan van Selst found and fixed a OpenSSL session ref count leak:
615 ossl_connect_step3() increments an SSL session handle reference counter on
616 each call. When sessions are re-used this reference counter may be
617 incremented many times, but it will be decremented only once when done (by
618 Curl_ossl_session_free()); and the internal OpenSSL data will not be freed
619 if this reference count remains positive. When a session is re-used the
620 reference counter should be corrected by explicitly calling
621 SSL_SESSION_free() after each consecutive SSL_get1_session() to avoid
622 introducing a memory leak.
624 (http://curl.haxx.se/bug/view.cgi?id=2926284)
626 Daniel Stenberg (7 Jan 2010)
627 - Make sure the progress callback is called repeatedly even during very slow
628 name resolves when c-ares is used for resolving.
630 Claes Jakobsson (6 Jan 2010)
631 - Julien Chaffraix fixed so that the fragment part in an URL is not sent
632 to the server anymore.
634 Kamil Dudka (3 Jan 2010)
635 - Julien Chaffraix eliminated a duplicated initialization in singlesocket().
637 Daniel Stenberg (2 Jan 2010)
638 - Make curl support --ssl and --ssl-reqd instead of the previous FTP-specific
639 versions --ftp-ssl and --ftp-ssl-reqd as these options are now used to
640 control SSL/TLS for IMAP, POP3 and SMTP as well in addition to FTP. The old
641 option names are still working but the new ones are the ones listed and
644 Daniel Stenberg (1 Jan 2010)
645 - Ingmar Runge enhanced libcurl's FTP engine to support the PRET command. This
646 command is a special "hack" used by the drftpd server, but even though it is
647 a custom extension I've deemed it fine to add to libcurl since this server
648 seems to survive and people keep using it and want libcurl to support
649 it. The new libcurl option is named CURLOPT_FTP_USE_PRET, and it is also
650 usable from the curl tool with --ftp-pret. Using this option on a server
651 that doesn't support this command will make libcurl fail.
653 I added test cases 1107 and 1108 to verify the functionality.
655 The PRET command is documented at
656 http://www.drftpd.org/index.php/Distributed_PASV
658 Yang Tse (30 Dec 2009)
659 - Steven M. Schweda improved VMS build system, and Craig A. Berry helped
660 with the patch and testing.
662 Daniel Stenberg (26 Dec 2009)
663 - Renato Botelho and Peter Pentchev brought a patch that makes the libcurl
664 headers work correctly even on FreeBSD systems before v8.
666 (http://curl.haxx.se/bug/view.cgi?id=2916915)
668 Daniel Stenberg (17 Dec 2009)
669 - David Byron fixed Curl_ossl_cleanup to actually call ENGINE_cleanup when
672 - Follow-up fix for the proxy fix I did for Jon Nelson's bug. It turned out I
673 was a bit too quick and broke test case 1101 with that change. The order of
674 some of the setups is sensitive. I now changed it slightly again to make
675 sure we do them in this order:
677 1 - parse URL and figure out what protocol is used in the URL
678 2 - prepend protocol:// to URL if missing
679 3 - parse name+password off URL, which needs to know what protocol is used
680 (since only some allows for name+password in the URL)
681 4 - figure out if a proxy should be used set by an option
682 5 - if no proxy option, check proxy environment variables
683 6 - run the protocol-specific setup function, which needs to have the proxy
686 Daniel Stenberg (15 Dec 2009)
687 - Jon Nelson found a regression that turned out to be a flaw in how libcurl
688 detects and uses proxies based on the environment variables. If the proxy
689 was given as an explicit option it worked, but due to the setup order
690 mistake proxies would not be used fine for a few protocols when picked up
691 from '[protocol]_proxy'. Obviously this broke after 7.19.4. I now also added
692 test case 1106 that verifies this functionality.
694 (http://curl.haxx.se/bug/view.cgi?id=2913886)
696 Daniel Stenberg (12 Dec 2009)
697 - IMAP, POP3 and SMTP support and their TLS versions (including IMAPS, POP3S
698 and SMTPS) are now supported. The current state may not yet be solid, but
699 the foundation is in place and the test suite has some initial support for
700 these protocols. Work will now persue to make them nice libcurl citizens
703 The work with supporting these new protocols was sponsored by
704 networking4all.com - thanks!
706 Daniel Stenberg (10 Dec 2009)
707 - Siegfried Gyuricsko found out that the curl manual said --retry would retry
708 on FTP errors in the transient 5xx range. Transient FTP errors are in the
709 4xx range. The code itself only tried on 5xx errors that occured _at login_.
710 Now the retry code retries on all FTP transfer failures that ended with a
713 (http://curl.haxx.se/bug/view.cgi?id=2911279)
715 - Constantine Sapuntzakis figured out a case which would lead to libcurl
716 accessing alredy freed memory and thus crash when using HTTPS (with
717 OpenSSL), multi interface and the CURLOPT_DEBUGFUNCTION and a certain order
718 of cleaning things up. I fixed it.
720 (http://curl.haxx.se/bug/view.cgi?id=2905220)
722 Daniel Stenberg (7 Dec 2009)
723 - Martin Storsjo made libcurl use the Expect: 100-continue header for posts
724 with unknown size. Previously it was only used for posts with a known size
725 larger than 1024 bytes.
727 Daniel Stenberg (1 Dec 2009)
728 - If the Expect: 100-continue header has been set by the application through
729 curl_easy_setopt with CURLOPT_HTTPHEADER, the library should set
730 data->state.expect100header accordingly - the current code (in 7.19.7 at
731 least) doesn't handle this properly. Martin Storsjo provided the fix!
733 Yang Tse (28 Nov 2009)
734 - Added Diffie-Hellman parameters to several test harness certificate files in
735 PEM format. Required by several stunnel versions used by our test harness.
737 Daniel Stenberg (28 Nov 2009)
738 - Markus Koetter provided a polished and updated version of Chad Monroe's TFTP
739 rework patch that now integrates TFTP properly into libcurl so that it can
740 be used non-blocking with the multi interface and more. BLKSIZE also works.
742 The --tftp-blksize option was added to allow setting the TFTP BLKSIZE from
745 Daniel Stenberg (26 Nov 2009)
746 - Extended and fixed the change I did on Dec 11 for the the progress
747 meter/callback during FTP command/response sequences. It turned out it was
748 really lame before and now the progress meter SHOULD get called at least
751 Daniel Stenberg (23 Nov 2009)
752 - Bjorn Augustsson reported a bug which made curl not report any problems even
753 though it failed to write a very small download to disk (done in a single
754 fwrite call). It turned out to be because fwrite() returned success, but
755 there was insufficient error-checking for the fclose() call which tricked
756 curl to believe things were fine.
758 Yang Tse (23 Nov 2009)
759 - David Byron modified Makefile.dist vc8 and vc9 targets in order to allow
760 finer granularity control when generating src and lib makefiles.
762 Yang Tse (22 Nov 2009)
763 - I modified configure to force removal of the curlbuild.h file included in
764 distribution tarballs for use by non-configure systems. As intended, this
765 would get overwriten when doing in-tree builds. But VPATH builds would end
766 having two curlbuild.h files, one in the source tree and another in the
767 build tree. With the modification I introduced 5 Nov 2009 this could become
768 an issue when running libcurl's test suite.
770 Daniel Stenberg (20 Nov 2009)
771 - Constantine Sapuntzakis identified a write after close, as the sockets were
772 closed by libcurl before the SSL lib were shutdown and they may write to its
773 socket. Detected to at least happen with OpenSSL builds.
775 - Jad Chamcham pointed out a bug with connection re-use. If a connection had
776 CURLOPT_HTTPPROXYTUNNEL enabled over a proxy, a subsequent request using the
777 same proxy with the tunnel option disabled would still wrongly re-use that
778 previous connection and the outcome would only be badness.
780 Yang Tse (18 Nov 2009)
781 - I modified the memory tracking system to make it intolerant with zero sized
782 malloc(), calloc() and realloc() function calls.
784 Daniel Stenberg (17 Nov 2009)
785 - Constantine Sapuntzakis provided another fix for the DNS cache that could
786 end up with entries that wouldn't time-out:
788 1. Set up a first web server that redirects (307) to a http://server:port
790 2. Have curl connect to the first web server using curl multi
792 After the curl_easy_cleanup call, there will be curl dns entries hanging
793 around with in_use != 0.
795 (http://curl.haxx.se/bug/view.cgi?id=2891591)
797 - Marc Kleine-Budde fixed: curl saved the LDFLAGS set during configure into
798 its pkg-config file. So -Wl stuff ended up in the .pc file, which is really
799 bad, and breaks if there are multiple -Wl in our LDFLAGS (which are in
800 PTXdist). bug #2893592 (http://curl.haxx.se/bug/view.cgi?id=2893592)
802 Kamil Dudka (15 Nov 2009)
803 - David Byron improved the configure script to use pkg-config to find OpenSSL
804 (and in particular the list of required libraries) even if a path is given
805 as argument to --with-ssl
807 Yang Tse (15 Nov 2009)
808 - I removed enable-thread / disable-thread configure option. These were only
809 placebo options. The library is always built as thread safe as possible on
812 Claes Jakobsson (14 Nov 2009)
813 - curl-config now accepts '--configure' to see what arguments was
814 passed to the configure script when building curl.
816 Daniel Stenberg (14 Nov 2009)
817 - Claes Jakobsson restored the configure functionality to detect NSS when
818 --with-nss is set but not "yes".
820 I think we can still improve that to check for pkg-config in that path etc,
821 but at least this patch brings back the same functionality we had before.
823 - Camille Moncelier added support for the file type SSL_FILETYPE_ENGINE for
824 the client certificate. It also disable the key name test as some engines
825 can select a private key/cert automatically (When there is only one key
826 and/or certificate on the hardware device used by the engine)
828 Yang Tse (14 Nov 2009)
829 - Constantine Sapuntzakis provided the fix that ensures that an SSL connection
830 won't be reused unless protection level for peer and host verification match.
832 I refactored how preprocessor symbol _THREAD_SAFE definition is done.
834 Kamil Dudka (12 Nov 2009)
835 - Kevin Baughman provided a fix preventing libcurl-NSS from crash on doubly
836 closed NSPR descriptor. The issue was hard to find, reported several times
837 before and always closed unresolved. More info at the RH bug:
838 https://bugzilla.redhat.com/534176
840 - libcurl-NSS now tries to reconnect with TLS disabled in case it detects
841 a broken TLS server. However it does not happen if SSL version is selected
842 manually. The approach was originally taken from PSM. Kaspar Brand helped me
843 to complete the patch. Original bug reports:
844 https://bugzilla.redhat.com/525496
845 https://bugzilla.redhat.com/527771
847 Yang Tse (12 Nov 2009)
848 - I modified configure script to make the getaddrinfo function check also
849 verify if the function is thread safe.
851 Yang Tse (11 Nov 2009)
852 - Marco Maggi reported that compilation failed when configured --with-gssapi
853 and GNU GSS installed due to a missing mutual exclusion of header files in
854 the Kerberos 5 code path. He also verified that my patch worked for him.
856 Daniel Stenberg (11 Nov 2009)
857 - Constantine Sapuntzakis posted bug #2891595
858 (http://curl.haxx.se/bug/view.cgi?id=2891595) which identified how an entry
859 in the DNS cache would linger too long if the request that added it was in
860 use that long. He also provided the patch that now makes libcurl capable of
861 still doing a request while the DNS hash entry may get timed out.
863 - Christian Schmitz noticed that the progress meter/callback was not properly
864 used during the FTP connection phase (after the actual TCP connect), while
865 it of course should be. I also made the speed check get called correctly so
866 that really slow servers will trigger that properly too.
868 Kamil Dudka (5 Nov 2009)
869 - Dropped misleading timeouts in libcurl-NSS and made sure the SSL socket works
870 in non-blocking mode.
872 Yang Tse (5 Nov 2009)
873 - I removed leading 'curl' path on the 'curlbuild.h' include statement in
874 curl.h, adjusting auto-makefiles include path, to enhance portability to
875 OS's without an orthogonal directory tree structure such as OS/400.
877 Daniel Stenberg (4 Nov 2009)
878 - I fixed several problems with the transfer progress meter. It showed the
879 wrong percentage for small files, most notable for <1000 bytes and could
880 easily end up showing more than 100% at the end. It also didn't show any
881 percentage, transfer size or estimated transfer times when transferring
884 Version 7.19.7 (4 November 2009)
886 Daniel Stenberg (2 Nov 2009)
887 - As reported independent by both Stan van de Burgt and Didier Brisebourg,
888 CURLINFO_SIZE_DOWNLOAD (the -w variable size_download) didn't work when
889 getting data from ldap!
891 Daniel Stenberg (31 Oct 2009)
892 - Gabriel Kuri reported a problem with CURLINFO_CONTENT_LENGTH_DOWNLOAD if the
893 download was 0 bytes, as libcurl would then return the size as unknown (-1)
894 and not 0. I wrote a fix and test case 566 to verify it.
896 Daniel Stenberg (30 Oct 2009)
897 - Liza Alenchery mentioned a problem with re-used SCP connection when a bad
898 auth is used, as it caused a crash. I failed to repeat the issue, but still
899 made a change that now forces the TCP connection used for a freed SCP
900 session to get closed and not be re-used.
902 - "Tom" posted a bug report that mentioned how libcurl did wrong when doing a
903 POST using a read callback, with Digest authentication and
904 "Transfer-Encoding: chunked" enforced. I would then cause the first request
905 to be wrongly sent and then basically hang until the server closed the
906 connection. I fixed the problem and added test case 565 to verify it.
908 Daniel Stenberg (25 Oct 2009)
909 - Dima Barsky made the curl cookie parser accept cookies even with blank or
910 unparsable expiry dates and then treat them as session cookies - previously
911 libcurl would reject cookies with a date format it couldn't parse. Research
912 shows that the major browser treat such cookies as session cookies. I
913 modified test 8 and 31 to verify this.
915 Daniel Stenberg (21 Oct 2009)
916 - Attempt to use pkg-config for finding out libssh2 installation details
919 - A patch in bug report #2883177 (http://curl.haxx.se/bug/view.cgi?id=2883177)
920 by Johan van Selst introduced the --crlfile option to curl, which makes curl
921 tell libcurl about a file with CRL (certificate revocation list) data to
924 Daniel Stenberg (18 Oct 2009)
925 - Ray Dassen provided a patch in Debian's bug tracker (bug number #551461)
926 that now makes curl_getdate(3) actually handles RFC 822 formatted dates that
927 use the "single letter military timezones".
928 http://www.rfc-ref.org/RFC-TEXTS/822/chapter5.html has the details.
930 - Fixed memory leak in the SCP/SFTP code as it never freed the knownhosts
933 - John Dennis filed bug report #2873666
934 (http://curl.haxx.se/bug/view.cgi?id=2873666) which identified a problem
935 which made libcurl loop infinitely when given incorrect credentials when
936 using HTTP GSS negotiate authentication. He also provided a small and simple
939 - Kevin Baughman found a double close() problem with libcurl-NSS, as when
940 libcurl called NSS to close the SSL "session" it also closed the actual
943 Yang Tse (17 Oct 2009)
944 - Bug report #2866724 indicated
945 (http://curl.haxx.se/bug/view.cgi?id=2866724) that curl on Windows failed
946 when writing files whose file names originally contained characters which
947 are not valid for file names on Windows. Dan Fandrich provided an initial
948 patch and another revised one to fix this issue.
950 Daniel Stenberg (1 Oct 2009)
951 - Tom Mueller correctly reported in bug report #2870221
952 (http://curl.haxx.se/bug/view.cgi?id=2870221) that libcurl returned an
953 incorrect return code from the internal trynextip() function which caused
954 him grief. This is a regression that was introduced in 7.19.1 and I find it
955 strange it hasn't hit us harder, but I won't persue into figuring out
958 - Constantine Sapuntzakis: The current implementation will always set
959 SO_SNDBUF to CURL_WRITE_SIZE even if the SO_SNDBUF starts out larger. The
960 patch doesn't do a setsockopt if SO_SNDBUF is already greater than
961 CURL_WRITE_SIZE. This should help folks who have set up their computer with
964 Daniel Stenberg (27 Sep 2009)
965 - I introduced a maximum limit for received HTTP headers. It is controlled by
966 the define CURL_MAX_HTTP_HEADER which is even exposed in the public header
967 file to allow for users to fairly easy rebuild libcurl with a modified
968 limit. The rationale for a fixed limit is that libcurl is realloc()ing a
969 buffer to be able to put a full header into it, so that it can call the
970 header callback with the entire header, but that also risk getting it into
971 trouble if a server by mistake or willingly sends a header that is more or
972 less without an end. The limit is set to 100K.
974 Daniel Stenberg (26 Sep 2009)
975 - John P. McCaskey posted a bug report that showed how libcurl did wrong when
976 saving received cookies with no given path, if the path in the request had a
977 query part. That is means a question mark (?) and characters on the right
978 side of that. I wrote test case 1105 and fixed this problem.
980 Kamil Dudka (26 Sep 2009)
981 - Implemented a protocol independent way to specify blocking direction, used by
982 transfer.c for blocking. It is currently used only by SCP and SFTP protocols.
983 This enhancement resolves an issue with 100% CPU usage during SFTP upload,
986 Daniel Stenberg (25 Sep 2009)
987 - Chris Mumford filed bug report #2861587
988 (http://curl.haxx.se/bug/view.cgi?id=2861587) identifying that libcurl used
989 the OpenSSL function X509_load_crl_file() wrongly and failed if it would
990 load a CRL file with more than one certificate within. This is now fixed.
992 Daniel Stenberg (16 Sep 2009)
993 - Sven Anders reported that we introduced a cert verfication flaw for OpenSSL-
994 powered libcurl in 7.19.6. If there was a X509v3 Subject Alternative Name
995 field in the certficate it had to match and so even if non-DNS and non-IP
996 entry was present it caused the verification to fail.
998 Daniel Fandrich (15 Sep 2009)
999 - Moved the libssh2 checks after the SSL library checks. This helps when
1000 statically linking since libssh2 needs the SSL library link flags to be
1001 set up already to satisfy its dependencies. This wouldn't be necessary if
1002 the libssh2 configure check was changed to use pkg-config since the
1003 --static flag would add the dependencies automatically.
1005 Yang Tse (14 Sep 2009)
1006 - Revert Joshua Kwan's patch committed 11 Sep 2009.
1008 Some systems poll function sets POLLHUP in revents without setting
1009 POLLIN, and sets POLLERR without setting POLLIN and POLLOUT. In some
1010 libcurl code execution paths this could trigger busy wait loops with
1011 high CPU usage until a timeout condition aborted the loop.
1013 The reverted patch addressed the above issue for a very specific case,
1014 when awaiting c-ares to resolve. A libcurl-wide fix for Curl_poll now
1015 superceeds this one.
1017 Guenter Knauf (11 Sep 2009)
1018 - Joshua Kwan provided a patch to pass POLLERR / POLLHUP back to c-ares.
1019 This fixes a loop problem with high CPU usage.
1021 Daniel Stenberg (10 Sep 2009)
1022 - Claes Jakobsson fixed a problem with cookie expiry dates at exctly the epoch
1023 start second "Thu Jan 1 00:00:00 GMT 1970" as the date parser then returns 0
1024 which internally then is treated as a session cookie. That particular date
1025 is now made to get the value of 1.
1027 Daniel Stenberg (2 Sep 2009)
1028 - Daniel Johnson found a flaw in the code converting sftp-errors to libcurl
1031 Daniel Stenberg (1 Sep 2009)
1032 - Peter Sylvester made a debug feature for Curl_resolv() that now will force
1033 libcurl to resolve 'localhost' whatever name you use in the URL *if* you set
1034 the --interface option to (exactly) "LocalHost". This will enable us to
1035 write tests for custom hosts names but still use a local host server.
1037 - configure now tries to use pkg-config for a number of sub-dependencies even
1038 when cross-compiling. The key to success is then you properly setup
1039 PKG_CONFIG_PATH before invoking configure.
1041 I also improved how NSS is detected by trying nss-config if pkg-config isn't
1042 present, and as a last resort just use the lib name and force the user to
1043 setup the LIBS/LDFLAGS/CFLAGS etc properly. The previous last resort would
1044 add a range of various libs that would almost never be quite correct.
1046 Daniel Stenberg (31 Aug 2009)
1047 - When using the multi interface with FTP and you asked for NOBODY, you did no
1048 QUOTE commands and the request used the same path as the connection had
1049 already changed to, it would decide that no commands would be necessary for
1050 the "DO" action and that was not handled properly but libcurl would instead
1053 Kamil Dudka (28 Aug 2009)
1054 - Improved error message for not matching certificate subject name in
1055 libcurl-NSS. Originally reported at:
1056 https://bugzilla.redhat.com/show_bug.cgi?id=516056#c9
1058 Patrick Monnerat (24 Aug 2009)
1059 - Introduced a SYST-based test to properly set-up name format when dealing
1060 with the OS/400 FTP server.
1062 - Fixed an ftp_readresp() bug preventing detection of failing control socket
1063 and causing FTP client to loop forever.
1065 Daniel Stenberg (24 Aug 2009)
1066 - Marc de Bruin pointed out that configure --with-gnutls=PATH didn't work
1067 properly and provided a fix. http://curl.haxx.se/bug/view.cgi?id=2843008
1069 - Eric Wong introduced support for the new option -T. (dot) that makes curl
1070 read stdin in a non-blocking fashion. This also brings back -T- (minus) to
1071 the previous blocking behavior since it could break stuff for people at
1074 Michal Marek (21 Aug 2009)
1075 - With CURLOPT_PROXY_TRANSFER_MODE, avoid sending invalid URLs like
1076 ftp://example.com;type=i if the user specified ftp://example.com without the
1079 Daniel Stenberg (21 Aug 2009)
1080 - Andre Guibert de Bruet pointed out a missing return code check for a
1081 strdup() that could lead to segfault if it returned NULL. I extended his
1082 suggest patch to now have Curl_retry_request() return a regular return code
1083 and better check that.
1085 - Lots of good work by Krister Johansen, mostly related to pipelining:
1087 Fix SIGSEGV on free'd easy_conn when pipe unexpectedly breaks
1088 Fix data corruption issue with re-connected transfers
1089 Fix use after free if we're completed but easy_conn not NULL
1091 Kamil Dudka (13 Aug 2009)
1092 - Changed NSS code to not ignore the value of ssl.verifyhost and produce more
1093 verbose error messages. Originally reported at:
1094 https://bugzilla.redhat.com/show_bug.cgi?id=516056
1096 Daniel Stenberg (12 Aug 2009)
1097 - Karl Moerder fixed the Makefile.vc* makefiles to include the new file
1098 nonblock.c so that they work fine again
1100 - I expanded test 517 with a bunch of more dates that originate from the
1101 Chrome browser test suite. It turns out most of them get parsed the same
1104 Version 7.19.6 (12 August 2009)
1106 Daniel Stenberg (12 Aug 2009)
1107 - Carsten Lange reported a bug and provided a patch for TFTP upload and the
1108 sending of the TSIZE option. I don't like fixing bugs just hours before
1109 a release, but since it was broken and the patch fixes this for him I decided
1110 to get it in anyway.
1112 Daniel Stenberg (11 Aug 2009)
1113 - Peter Sylvester made the HTTPS test server use specific certificates for
1114 each test, so that the test suite can now be used to actually test the
1115 verification of cert names etc. This made an error show up in the OpenSSL-
1116 specific code where it would attempt to match the CN field even if a
1117 subjectAltName exists that doesn't match. This is now fixed and verified
1120 - Benbuck Nason posted the bug report #2835196
1121 (http://curl.haxx.se/bug/view.cgi?id=2835196), fixing a few compiler
1122 warnings when mixing ints and bools.
1124 Daniel Fandrich (10 Aug 2009)
1125 - Fixed a memory leak in the FTP code and an off-by-one heap buffer overflow.
1127 Daniel Fandrich (9 Aug 2009)
1128 - Fixed some memory leaks in the command-line tool that caused most of the
1129 torture tests to fail.
1131 Daniel Stenberg (2 Aug 2009)
1132 - Curt Bogmine reported a problem with SNI enabled on a particular server. We
1133 should introduce an option to disable SNI, but as we're in feature freeze
1134 now I've addressed the obvious bug here (pointed out by Peter Sylvester): we
1135 shouldn't try to enable SNI when SSLv2 or SSLv3 is explicitly selected.
1136 Code for OpenSSL and GnuTLS was fixed. NSS doesn't seem to have a particular
1137 option for SNI, or are we simply not using it?
1139 Daniel Stenberg (1 Aug 2009)
1140 - Scott Cantor posted the bug report #2829955
1141 (http://curl.haxx.se/bug/view.cgi?id=2829955) mentioning the recent SSL cert
1142 verification flaw found and exploited by Moxie Marlinspike. The presentation
1143 he did at Black Hat is available here:
1144 https://www.blackhat.com/html/bh-usa-09/bh-usa-09-archives.html#Marlinspike
1146 Apparently at least one CA allowed a subjectAltName or CN that contain a
1147 zero byte, and thus clients that assumed they would never have zero bytes
1148 were exploited to OK a certificate that didn't actually match the site. Like
1149 if the name in the cert was "example.com\0theatualsite.com", libcurl would
1150 happily verify that cert for example.com.
1152 libcurl now better uses the length of the extracted name, not using the zero
1153 termination for getting the string length.
1155 This fixing only made and needed in OpenSSL interfacing code.
1157 - Tanguy Fautre pointed out that OpenSSL's function RAND_screen() (present
1158 only in some OpenSSL installs - like on Windows) isn't thread-safe and we
1159 agreed that moving it to the global_init() function is a decent way to deal
1160 with this situation.
1162 - Alexander Beedie provided the patch for a noproxy problem: If I have set
1163 CURLOPT_NOPROXY to "*", or to a host that should not use a proxy, I actually
1164 could still end up using a proxy if a proxy environment variable was set.
1166 Daniel Stenberg (27 Jul 2009)
1167 - All the quote options (CURLOPT_QUOTE, CURLOPT_POSTQUOTE and
1168 CURLOPT_PREQUOTE) now accept a preceeding asterisk before the command to
1169 send when using FTP, as a sign that libcurl shall simply ignore the response
1170 from the server instead of treating it as an error. Not treating a 400+ FTP
1171 response code as an error means that failed commands will not abort the
1172 chain of commands, nor will they cause the connection to get disconnected.
1174 Daniel Stenberg (26 Jul 2009)
1175 - Johan van Selst posted bug report #2825989
1176 (http://curl.haxx.se/bug/view.cgi?id=2825989) pointing out that
1177 OpenSSL-powered libcurl didn't support the SHA-2 digest algorithm, and
1178 provided the solution too: to use OpenSSL_add_all_algorithms() in addition
1179 to the older SSLeay_* alternative. OpenSSL_add_all_algorithms was added in
1182 Daniel Stenberg (23 Jul 2009)
1183 - Added CURLOPT_SSH_KNOWNHOSTS, CURLOPT_SSH_KEYFUNCTION, CURLOPT_SSH_KEYDATA.
1184 They introduce known_host support for SSH keys to libcurl. See docs for
1185 details. Note that this feature depends on a new enough libssh2 version, to
1186 be supported in libssh2 1.2 and later (or current git repo at this time).
1188 Michal Marek (22 Jul 2009)
1189 - David Binderman found a memory and fd leak in lib/gtls.c:load_file()
1190 (https://bugzilla.novell.com/523919). When looking at the code, I found that
1191 also the ptr pointer can leak.
1193 Kamil Dudka (20 Jul 2009)
1194 - Claes Jakobsson improved the support for client certificates handling in
1195 NSS-powered libcurl. Now the client certificates can be selected
1196 automatically by a NSS built-in hook. Additionally pre-login to all PKCS11
1197 slots is no more performed. It used to cause problems with HW tokens.
1199 - Fixed reference counting for NSS client certificates. Now the PEM reader
1200 module should be always properly unloaded on Curl_nss_cleanup(). If the
1201 unload fails though, libcurl will try to reuse the already loaded instance.
1203 Daniel Fandrich (15 Jul 2009)
1204 - Added nonblock.c to the non-automake makefiles (note that the dependencies
1205 in the Watcom makefiles aren't quite correct).
1207 Michal Marek (15 Jul 2009)
1208 - Changed the description of CURLINFO_OS_ERRNO to make it clear that the
1209 errno is not reset on success.
1211 Guenter Knauf (14 Jul 2009)
1212 - renamed generated config.h to curl_config.h to avoid any future clashes
1213 with config.h from other projects.
1215 Daniel Stenberg (9 Jul 2009)
1216 - Eric Wong introduced curlx_nonblock() that the curl tool now (re-)uses for
1217 setting a file descriptor non-blocking. Used by the functionality Eric
1218 himself brough on June 15th.
1220 Daniel Stenberg (8 Jul 2009)
1221 - Constantine Sapuntzakis posted bug report #2813123
1222 (http://curl.haxx.se/bug/view.cgi?id=2813123) and an a patch that fixes the
1225 Url A is accessed using auth. Url A redirects to Url B (on a different
1226 server0. Url B reuses a persistent connection. Url B has auth, even though
1227 it's on a different server.
1229 Note: if Url B does not reuse a persistent connection, auth is not sent.
1233 data->state.first_host is not initialized becuase Curl_http_connect is not
1234 called when a connection is reused.
1238 move initialization of data->state.first_host to Curl_http. No code before
1239 Curl_http uses data->state.first_host anyway.
1241 Guenter Knauf (4 Jul 2009)
1242 - Markus Koetter provided a patch to avoid getnameinfo() usage which broke a
1243 couple of both IPv4 and IPv6 autobuilds.
1245 Daniel Stenberg (29 Jun 2009)
1246 - Markus Koetter made CURLOPT_FTPPORT (and curl's -P/--ftpport) support a port
1247 range if given colon-separated after the host name/address part. Like
1248 "192.168.0.1:2000-10000"
1250 - Modified the separators used for CURLOPT_CERTINFO in multi-part outputs. I
1251 don't know how they got wrong in the first place, but using this output
1252 format makes it possible to quite easily separate the string into an array
1255 Daniel Fandrich (16 June 2009)
1256 - Added a few more compiler warning options for gcc.
1258 Daniel Stenberg (16 Jun 2009)
1259 - Reuven Wachtfogel made curl -o - properly produce a binary output on windows
1260 (no newline translations). Use -B/--use-ascii if you rather get the ascii
1263 Michal Marek (16 Jun 2009)
1264 - When doing non-anonymous ftp via http proxies and the password is not
1265 provided in the url, add it there (squid needs this).
1267 Daniel Stenberg (15 Jun 2009)
1268 - Eric Wong's patch:
1270 This allows curl(1) to be used as a client-side tunnel for arbitrary stream
1271 protocols by abusing chunked transfer encoding in both the HTTP request and
1272 HTTP response. This requires server support for sending a response while a
1273 request is still being read, of course.
1275 If attempting to read from stdin returns EAGAIN, then we pause our sender.
1276 This leaves curl to attempt to read from the socket while reading from stdin
1277 (and thus sending) is paused.
1279 This change was needed to allow successfully tunneling the git protocol over
1280 HTTP (--no-buffer is needed, as well).
1282 Patrick Monnerat (15 Jun 2009)
1283 - Replaced use of standard C library rand()/srand() by our own pseudo-random
1286 Yang Tse (11 Jun 2009)
1287 - I adapted testcurl script to allow building test harness programs when
1288 cross-compiling for a *-*-mingw* host.
1290 Daniel Stenberg (10 Jun 2009)
1291 - Fabian Keil ran clang on the (lib)curl code, found a bunch of warnings and
1292 contributed a range of patches to fix them.
1294 Yang Tse (10 Jun 2009)
1295 - I introduced configure script option --enable-curldebug which now allows
1296 the decoupled enabling or disabling of the curl debug memory tracking
1297 feature from the --enable-debug option which no longer controls this.
1299 curl --version will list 'Debug' feature for debug enabled builds, and
1300 will list 'TrackMemory' feature for curl debug memory tracking capable
1301 builds. These features are independent and can be controlled when running
1302 the configure script. When --enable-debug is given both features will be
1303 enabled, unless some restriction prevents memory tracking from being used.
1305 Internally, definition of preprocessor symbol DEBUGBUILD restricts code
1306 which is only compiled for debug enabled builds. And symbol CURLDEBUG is
1307 used to differentiate code which is _only_ used for memory tracking.
1309 Yang Tse (9 Jun 2009)
1310 - Daniel Steinberg pointed out that Curl_FormInit() in formdata.c was not
1311 initializing the fread callback pointer and this triggered a compiler
1312 warning, also provided a friendly suggestion on how to fix it.
1314 Daniel Stenberg (8 Jun 2009)
1315 - Claes Jakobsson provided a patch for libcurl-NSS that fixed a bad refcount
1316 issue with client certs that caused issues like segfaults.
1317 http://curl.haxx.se/mail/lib-2009-05/0316.html
1319 - Triggered by bug report #2798852 and the patch in there, I fixed configure
1320 to detect gnutls build options with pkg-config only and not libgnutls-config
1321 anymore since GnuTLS has stopped distributing that tool. If an explicit path
1322 is given to configure, we will instead guess on how to link and use that
1323 lib. I did not use the patch from the bug report.
1325 Yang Tse (8 Jun 2009)
1326 - Igor Novoseltsev adjusted Makefile.vxworks to get sources and headers
1327 included from Makefile.inc, and provided docs\INSTALL VxWorks section.
1329 - I removed buildconf.bat from release and daily snapshot archives. This
1330 file is only for CVS tree checkout builds.
1332 Daniel Stenberg (8 Jun 2009)
1333 - Eric Wong fixed --no-buffer to actually switch off output buffering. Been
1336 Bill Hoffman (6 Jun 2009)
1337 - Added some cmake docs and fixed socklen_t in the build.
1339 Yang Tse (5 Jun 2009)
1340 - John E. Malmberg provided VMS specific patch: "This fixes an existing bug
1341 in urlglob.c where it was not converting the Curl Unix exit code to a VMS
1342 DCL compatible exit code. This fix required the enhancement described next.
1343 This also adds an enhancement to main.c so that when curl is run under a
1344 Unix shell like Bash on VMS, it will return the standard Unix exit codes
1345 and messages." And another patch for docs/examples.
1347 I introduced os-specific.c and os-specific.h for use in curl tool code
1348 and adjusted John E. Malmberg's patch placement to use these new files
1349 as an effort to prevent main.c from growing ad infinitum. Code already
1350 existing in main.c which is OS specific should be moved into these files.
1352 Daniel Stenberg (4 June 2009)
1353 - Setting the Content-Length: header from your app when you do a POST or PUT
1354 is almost always a VERY BAD IDEA. Yet there are still apps out there doing
1355 this, and now recently it triggered a bug/side-effect in libcurl as when
1356 libcurl sends a POST or PUT with NTLM, it sends an empty post first when it
1357 knows it will just get a 401/407 back. If the app then replaced the
1358 Content-Length header, it caused the server to wait for input that libcurl
1359 wouldn't send. Aaron Oneal reported this problem in bug report #2799008
1360 (http://curl.haxx.se/bug/view.cgi?id=2799008) and helped us verify the fix.
1362 Yang Tse (4 Jun 2009)
1363 - Igor Novoseltsev provided patches and information, that after some
1364 adjustments to better fit curl's way of doing things, have resulted
1365 in the posibility of building libcurl for VxWorks.
1367 Daniel Fandrich (2 June 2009)
1368 - Checked in a Google Android make file. To use it, you must first
1369 create a config.h file by running configure in the Android environment,
1370 which doesn't seem to be easy to do. If no easy way can be found, a
1371 static config-android.h may need to be created and checked in to the
1372 libcurl source tree.
1374 Daniel Stenberg (1 June 2009)
1375 - Claes Jakobsson fixed the configure script to better find and use NSS
1378 Yang Tse (1 Jun 2009)
1379 - John E. Malmberg provided a VMS specific clean-up for curl.h, and pointed
1380 out that the configure script was failing to detect the timeval struct on
1381 VMS when building with _XOPEN_SOURCE_EXTENDED undefined due to definition
1382 taking place in socket.h instead of time.h. I have adjusted configure
1383 script to also include this header when checking struct timeval.
1385 Daniel Stenberg (27 May 2009)
1386 - Frank McGeough provided a small OpenSSL #include fix to make libcurl compile
1387 fine with Nokia 5th edition 1.0 SDK for Symbian.
1389 - Andre Guibert de Bruet found a call to a OpenSSL function that didn't check
1390 for a failure properly.
1392 - Mike Crowe pointed out that setting CURLOPT_USERPWD to NULL used to clear
1393 the auth credentials back in 7.19.0 and earlier while now you have to set ""
1394 to get the same effect. His patch brings back the ability to use NULL.
1396 - Claes Jakobsson fixed libcurl-NSS to build fine even without the
1397 PK11_CreateGenericObject() function.
1399 Daniel Stenberg (25 May 2009)
1400 - bug report #2796358 (http://curl.haxx.se/bug/view.cgi?id=2796358) pointed
1401 out that the cookie parser would leak memory when it parses cookies that are
1402 received with domain, path etc set multiple times in the same header. While
1403 such a cookie is questionable, they occur in the wild and libcurl no longer
1404 leaks memory for them. I added such a header to test case 8.
1406 Daniel Fandrich (22 May 2009)
1407 - Removed some obsolete digest code that caused a valgrind error in test 551.
1409 Daniel Fandrich (20 May 2009)
1410 - Added "non-existing host" test keywords to make it easy to skip those
1411 tests on machines that have broken DNS configurations (such as
1412 those configured to use OpenDNS).
1414 Daniel Stenberg (19 May 2009)
1415 - Kamil Dudka brought the patch from the Redhat bug entry
1416 https://bugzilla.redhat.com/show_bug.cgi?id=427966 which was libcurl closing
1417 a bad file descriptor when closing down the FTP data connection. Caolan
1418 McNamara seems to be the original author of it.
1420 Version 7.19.5 (18 May 2009)
1422 Daniel Stenberg (17 May 2009)
1423 - James Bursa posted a patch to the mailing list that fixed a problem with
1424 no_proxy which made it not skip the proxy if the URL entered contained a
1425 user name. I added test case 1101 to verify.
1427 Daniel Stenberg (11 May 2009)
1428 - Balint Szilakszi reported a memory leak when libcurl did gzip decompression
1429 of streams that had some parts (legitimately) missing. We now provide and use
1430 a proper cleanup function for the content encoding submodule.
1431 http://curl.haxx.se/mail/lib-2009-05/0092.html
1433 - Kamil Dudka provided a fix for libcurl-NSS reported by Michael Cronenworth
1434 at https://bugzilla.redhat.com/show_bug.cgi?id=453612#c12
1436 If an incorrect password is given while loading a private key, libcurl ends
1437 up in an infinite loop consuming memory. The bug is critical.
1439 - I fixed the problem with doing NTLM, POST and then following a 302 redirect,
1440 as reported by Ebenezer Ikonne (on curl-users) and Laurent Rabret (on
1441 curl-library). The transfer was mistakenly marked to get more data to send
1442 but since it didn't actually have that, it just hung there...
1444 Daniel Stenberg (10 May 2009)
1445 - Andre Guibert de Bruet correctly pointed out an over-alloc with one wasted
1446 byte in the digest code.
1448 Yang Tse (9 May 2009)
1449 - Removed DOS and TPF package's subdirectory Makefile.am, it was only used
1450 to include some files in the distribution tarball serving no other purpose.
1451 Files from the DOS and TPF subdirectories are now included in the EXTRA_DIST
1452 of the Makefile in the parent subdirectory.
1454 Yang Tse (8 May 2009)
1455 - Changed host name literal in several tests to one under the haxx.se domain.
1457 - Renamed vc6 workspace and project files to avoid filename clash when used
1458 for conversion to later VS versions.
1460 Daniel Stenberg (8 May 2009)
1461 - Constantine Sapuntzakis fixed bug report #2784055
1462 (http://curl.haxx.se/bug/view.cgi?id=2784055) identifying a problem to
1463 connect to SOCKS proxies when using the multi interface. It turned out to
1464 almost not work at all previously. We need to wait for the TCP connect to
1465 be properly verified before doing the SOCKS magic.
1467 There's still a flaw in the FTP code for this.
1469 Daniel Stenberg (7 May 2009)
1470 - Made the SO_SNDBUF setting for the data connection socket for ftp uploads as
1471 well. See change 28 Apr 2009.
1473 Yang Tse (7 May 2009)
1474 - Fixed an issue affecting FTP transfers, introduced with the transfer.c
1475 patch committed May 4.
1477 Daniel Stenberg (7 May 2009)
1478 - Man page *roff problems fixed thanks to input from Colin Watson. Problems
1479 reported in the Debian package.
1481 - Vijay G filed bug report #2723236
1482 (http://curl.haxx.se/bug/view.cgi?id=2723236) identifying a problem with
1483 libcurl's TFTP code and its lack of dealing with the OACK packet.
1485 Yang Tse (5 May 2009)
1486 - Fixed the --ftp-port address of test #251 to the CLIENTIP address, and
1487 reverted the change affecting test suite harness committed 4 May.
1489 Daniel Stenberg (5 May 2009)
1490 - Inspired by Michael Smith's session id fix for OpenSSL, I did the
1491 corresponding fix in the GnuTLS code: make sure to store the new session id
1492 in case the previous re-used one is rejected.
1494 Daniel Stenberg (4 May 2009)
1495 - Michael Smith posted bug report #2786255
1496 (http://curl.haxx.se/bug/view.cgi?id=2786255) with a patch, identifying how
1497 libcurl did not deal with SSL session ids properly if the server rejected a
1498 re-use of one. Starting now, it will forget the rejected one and remember
1499 the new. This change was for OpenSSL only, it is likely that other SSL lib
1500 code needs similar fixes.
1502 Yang Tse (4 May 2009)
1503 - Applied David McCreedy's "transfer.c fixes for CURL_DO_LINEEND_CONV and
1504 non-ASCII platform HTTP requests" patch addressing two HTTP PUT problems:
1505 1) On non-ASCII platforms not all of the protocol portions of the PUT are
1506 being translated to ASCII. 2) On all platforms the line endings of part of
1507 the protocol portions are mangled from CRLF to CRCRLF if data->set.crlf or
1508 data->set.prefer_ascii are set (depending on CURL_DO_LINEEND_CONV).
1510 - Applied David McCreedy's patch to fix test suite harness to allow test FTP
1511 server and client on different machines, providing FTP client address when
1512 running the FTP test server.
1514 Daniel Fandrich (3 May 2009)
1515 - Added and disabled test case 563 which shows KNOWN_BUGS #59. The bug
1516 report failed to mention that a proxy must be used to reproduce it.
1518 Yang Tse (2 May 2009)
1519 - Use a build-time configured curl_socklen_t data type instead of socklen_t.
1521 Yang Tse (1 May 2009)
1522 - Applied David McCreedy's patches "TPF-platform specific changes to various
1523 files" and "http.c fix to Curl_proxyCONNECT for non-ASCII platforms", the
1524 former with minor edits.
1526 Daniel Stenberg (30 Apr 2009)
1527 - I was going to fix issue #59 in KNOWN_BUGS
1529 If the CURLOPT_PORT option is used on an FTP URL like
1530 "ftp://example.com/file;type=A" the ";type=A" is stripped off.
1532 I added test case 562 to verify, only to find out that I couldn't repeat
1533 this bug so I hereby consider it not a bug anymore!
1535 Daniel Stenberg (29 Apr 2009)
1536 - Based on bug report #2723219 (http://curl.haxx.se/bug/view.cgi?id=2723219)
1537 I've now made TFTP "connections" not being kept for re-use within libcurl.
1538 TFTP is UDP-based so the benefit was really low (if even existing) to begin
1539 with so instead of tracking down to fix this problem we instead removed the
1540 re-use. I also enabled test case 1099 that I wrote a few days ago to verify
1541 that this change fixes the reported problem.
1543 Daniel Stenberg (28 Apr 2009)
1544 - Constantine Sapuntzakis filed bug report #2783090
1545 (http://curl.haxx.se/bug/view.cgi?id=2783090) pointing out that on windows
1546 we need to grow the SO_SNDBUF buffer somewhat to get really good upload
1547 speeds. http://support.microsoft.com/kb/823764 has the details. Friends
1548 confirmed that simply adding 32 to CURL_MAX_WRITE_SIZE is enough.
1550 - Bug report #2709004 (http://curl.haxx.se/bug/view.cgi?id=2709004) by Tim
1551 Chen pointed out how curl couldn't upload with resume when reading from a
1554 This ended up with the introduction of a new return code for the
1555 CURLOPT_SEEKFUNCTION callback that basically says that the seek failed but
1556 that libcurl may try to resolve the situation anyway. In our case this means
1557 libcurl will attempt to instead read that much data from the stream instead
1558 of seeking and that way curl can now upload with resume when data is read
1561 Daniel Stenberg (26 Apr 2009)
1562 - Bug report #2779733 (http://curl.haxx.se/bug/view.cgi?id=2779733) by Sven
1563 Wegener pointed out that CURLINFO_APPCONNECT_TIME didn't work with the multi
1564 interface and provided a patch that fixed the problem!
1566 Daniel Stenberg (24 Apr 2009)
1567 - Kamil Dudka fixed another NSS-related leak when client certs were used.
1569 - Bug report #2779245 (http://curl.haxx.se/bug/view.cgi?id=2779245) by Rainer
1570 Koenig pointed out that the man page didn't tell that the *_proxy
1571 environment variables can be specified lower case or UPPER CASE and the
1572 lower case takes precedence,
1574 Daniel Fandrich (21 Apr 2009)
1575 - Added new libcurl source files to Amiga, RiscOS and VC6 build files.
1577 Yang Tse (21 Apr 2009)
1578 - Moved potential inclusion of system's malloc.h and memory.h header files to
1579 setup_once.h. Inclusion of each header file is based on the definition of
1580 NEED_MALLOC_H and NEED_MEMORY_H respectively.
1582 Renamed libcurl's memory.h to curl_memory.h
1584 Daniel Stenberg (20 Apr 2009)
1585 - Leanic Lefever reported a crash and did some detailed research on why and
1586 how it occurs (http://curl.haxx.se/mail/lib-2009-04/0289.html). The
1587 conclusion was that if an error is detected and Curl_done() is called for
1588 the connection, ftp_done() could at times return another error code that
1589 then would take precedence and that new code confused existing logic that
1590 works for the first error code (CURLE_SEND_ERROR) only.
1592 - Gisle Vanem noticed that --libtool would produce bogus strings at times for
1593 OBJECTPOINT options. Now we've introduced a new function - my_setopt_str -
1594 within the app for setting plain string options to avoid the risk of this
1597 Daniel Stenberg (17 Apr 2009)
1598 - Pramod Sharma reported and tracked down a bug when doing FTP over a HTTP
1599 proxy. libcurl would then wrongly close the connection after each
1600 request. In his case it had the weird side-effect that it killed NTLM auth
1601 for the proxy causing an inifinite loop!
1603 I added test case 1098 to verify this fix. The test case does however not
1604 properly verify that the transfers are done persistently - as I couldn't
1605 think of a clever way to achieve it right now - but you need to read the
1606 stderr output after a test run to see that it truly did the right thing.
1608 Daniel Stenberg (13 Apr 2009)
1609 - bug report #2727981 (http://curl.haxx.se/bug/view.cgi?id=2727981) by Martin
1610 Storsjö pointed out how setting CURLOPT_NOBODY to 0 could be downright
1611 confusing as it set the method to either GET or HEAD. The example he showed
1614 curl_easy_setopt(curl, CURLOPT_PUT, 1);
1615 curl_easy_setopt(curl, CURLOPT_NOBODY, 0);
1617 The new way doesn't alter the method until the request is about to start. If
1618 CURLOPT_NOBODY is then 1 the HTTP request will be HEAD. If CURLOPT_NOBODY is
1619 0 and the request happens to have been set to HEAD, it will then instead be
1620 set to GET. I believe this will be less surprising to users, and hopefully
1621 not hit any existing users badly.
1623 - Toshio Kuratomi reported a memory leak problem with libcurl+NSS that turned
1624 out to be leaking cacerts. Kamil Dudka helped me complete the fix. The issue
1625 is found in Redhat's bug tracker:
1626 https://bugzilla.redhat.com/show_bug.cgi?id=453612
1628 There are still memory leaks present, but they seem to have other reasons.
1630 Daniel Fandrich (11 Apr 2009)
1631 - Added new libcurl source files to Symbian OS build files.
1632 - Improved Symbian support for SSL.
1634 Yang Tse (10 Apr 2009)
1635 - Daniel Johnson improved the MacOSX-Framework shell script to now perform all
1636 the steps required to build a Mac OS X four way fat ppc/i386/ppc64/x86_64
1637 libcurl.framework. Four way fat framework requires OS X 10.5 SDK or later.
1639 Yang Tse (8 Apr 2009)
1640 - Removed Sun compilers preprocessor block from curlbuild.h.dist, this also
1641 removes it from the curlbuild.h file originally distributed by the cURL
1642 project as this file is intended for systems not capable of running the
1643 configure script. For those who have been building curl out of the source
1644 code curl distribution tarball provided by curl.haxx.se the change implies
1645 nothing. Previous change in this area committed 2 Apr becomes irrelevant.
1647 Daniel Stenberg (6 Apr 2009)
1648 - I clarified in the docs that CURLOPT_SEEKFUNCTION should return 0 on success
1649 and 1 on fatal errors. Previously it only mentioned non-zero on fatal
1650 errors. This is a slight change in meaning, but it follows what we've done
1651 elsewhere before and it opens up for LOTS of more useful return codes
1652 whenever we can think of them...
1654 Yang Tse (2 Apr 2009)
1655 - Fix curl_off_t definition for builds done using Sun compilers and a
1656 non-configured libcurl. In this case curl_off_t data type was gated
1657 to the off_t data type which depends on the _FILE_OFFSET_BITS. This
1658 configuration is exactly the unwanted configuration for our curl_off_t
1659 data type which must not depend on such setting. This breaks ABI for
1660 libcurl libraries built with Sun compilers which were built without
1661 having run the configure script with _FILE_OFFSET_BITS different than
1662 64 and using the ILP32 data model.
1664 Daniel Stenberg (1 Apr 2009)
1665 - Andre Guibert de Bruet fixed a NULL pointer use in an infof() call if a
1666 strdup() call failed.
1668 Daniel Fandrich (31 Mar 2009)
1669 - Properly return an error code in curl_easy_recv (reported by Jim Freeman).
1671 Daniel Stenberg (18 Mar 2009)
1672 - Kamil Dudka brought a patch that enables 6 additional crypto algorithms when
1673 NSS is used. These ciphers were added in NSS 3.4 and require to be enabled
1676 Daniel Stenberg (13 Mar 2009)
1677 - Use libssh2_version() to present the libssh2 version in case the libssh2
1678 library is found to support it.
1680 Yang Tse (12 Mar 2009)
1681 - Added missing Curl_read() return code checking in TELNET transfers.
1683 - Pierre Brico found and fixed TELNET transfers not being aborted upon
1684 a write callback failure.
1686 Daniel Stenberg (11 Mar 2009)
1687 - Kamil Dudka made the curl tool properly call curl_global_init() before any
1688 other libcurl function.
1690 Yang Tse (11 Mar 2009)
1691 - Added missing TELNET timeout support for Windows builds. This issue was
1692 reported by Pierre Brico.
1694 Daniel Stenberg (9 Mar 2009)
1695 - Frank Hempel found out a bug and provided the fix:
1697 curl_easy_duphandle did not necessarily duplicate the CURLOPT_COOKIEFILE
1698 option. It only enabled the cookie engine in the destination handle if
1699 data->cookies is not NULL (where data is the source handle). In case of a
1700 newly initialized handle which just had the cookie support enabled by a
1701 curl_easy_setopt(handle, CURL_COOKIEFILE, "")-call, handle->cookies was
1702 still NULL because the setopt-call only appends the value to
1703 data->change.cookielist, hence duplicating this handle would not have the
1704 cookie engine switched on.
1706 We also concluded that the slist-functionality would be suitable for being
1707 put in its own module rather than simply hanging out in lib/sendf.c so I
1708 created lib/slist.[ch] for them.
1710 - Andreas Farber made the 'buildconf' script check for the presence of m4
1711 scripts to make it detect a bad checkout earlier. People with older
1712 checkouts who don't do cvs update with the -d option won't get the new dirs
1713 and then will get funny outputs that can be a bit hard to understand and
1716 Daniel Stenberg (8 Mar 2009)
1717 - Andre Guibert de Bruet found and fixed a code segment in ssluse.c where the
1718 allocation of the memory BIO was not being properly checked.
1720 - Andre Guibert de Bruet fixed the gnutls-using code: There are a few places
1721 in the gnutls code where we were checking for negative values for errors,
1722 when the man pages state that GNUTLS_E_SUCCESS is returned on success and
1723 other values indicate error conditions.
1725 - Bill Egert pointed out (http://curl.haxx.se/bug/view.cgi?id=2671602) that
1726 curl didn't use sprintf() in a way that is documented to work in POSIX but
1727 since we use our own printf() code (from libcurl) that shouldn't be a
1728 problem. Nonetheless I modified the code to not rely on such particular
1729 features and to not cause further raised eyebrowse with no good reason.
1731 Daniel Fandrich (5 Mar 2009)
1732 - Expanded the security section of the libcurl-tutorial man page to cover
1733 more issues for authors to consider when writing robust libcurl-using
1736 Yang Tse (5 Mar 2009)
1737 - Fixed NTLM authentication memory leak on SSPI enabled Windows builds. This
1738 issue was noticed by Chris Deidun.
1740 Daniel Fandrich (4 Mar 2009)
1741 - Fixed a problem with m4 quoting in the OpenSSL configure check reported
1744 Daniel Stenberg (3 Mar 2009)
1745 - David James brought a patch that make libcurl close (all) dead connections
1746 whenever you attempt to open a new connection.
1748 1. After cleaning up a dead connection, "continue" instead of
1749 returning FALSE. This ensures that we clean up all dead connections,
1750 rather than just cleaning up the first dead connection.
1751 2. Move up the cleanup for dead connections so that it occurs for
1752 all connections, rather than just the connections which have the same
1753 preferences as our current new connection.
1755 Version 7.19.4 (3 March 2009)
1757 Daniel Stenberg (3 Mar 2009)
1758 - David Kierznowski notified us about a security flaw
1759 (http://curl.haxx.se/docs/adv_20090303.html also known as CVE-2009-0037) in
1760 which previous libcurl versions (by design) can be tricked to access an
1761 arbitrary local/different file instead of a remote one when
1762 CURLOPT_FOLLOWLOCATION is enabled. This flaw is now fixed in this release
1763 together this the addition of two new setopt options for controlling this
1766 o CURLOPT_REDIR_PROTOCOLS controls what protocols libcurl is allowed to
1767 follow to when CURLOPT_FOLLOWLOCATION is enabled. By default, this option
1768 excludes the FILE and SCP protocols and thus you nee to explicitly allow
1769 them in your app if you really want that behavior.
1771 o CURLOPT_PROTOCOLS controls what protocol(s) libcurl is allowed to fetch
1772 using the primary URL option. This is useful if you want to allow a user or
1773 other outsiders control what URL to pass to libcurl and yet not allow all
1774 protocols libcurl may have been built to support.
1776 Daniel Stenberg (27 Feb 2009)
1777 - Senthil Raja Velu reported a problem when CURLOPT_INTERFACE and
1778 CURLOPT_LOCALPORT were used together (the local port bind failed), and
1779 Markus Koetter provided the fix!
1781 Daniel Stenberg (25 Feb 2009)
1782 - As Daniel Fandrich figured out, we must do the GnuTLS initing in the
1783 curl_global_init() function to properly maintain the performing functions
1784 thread-safe. We've previously (28 April 2007) moved the init to a later time
1785 just to avoid it to fail very early when libgcrypt dislikes the situation,
1786 but that move was bad and the fix should rather be in libgcrypt or
1789 Daniel Stenberg (24 Feb 2009)
1790 - Brian J. Murrell found out that Negotiate proxy authentication didn't work.
1791 It happened because the code used the struct for server-based auth all the
1792 time for both proxy and server auth which of course was wrong.
1794 Daniel Stenberg (23 Feb 2009)
1795 - After a bug reported by James Cheng I've made curl_easy_getinfo() for
1796 CURLINFO_CONTENT_LENGTH_DOWNLOAD and CURLINFO_CONTENT_LENGTH_UPLOAD return
1797 -1 if the sizes aren't know. Previously these returned 0, make it impossible
1798 to detect the difference between actually zero and unknown.
1800 Yang Tse (23 Feb 2009)
1801 - Daniel Johnson provided a shell script that will perform all the steps needed
1802 to build a Mac OS X fat ppc/i386 or ppc64/x86_64 libcurl.framework
1804 Daniel Stenberg (23 Feb 2009)
1805 - I renamed everything in the windows builds files that used the name 'curllib'
1806 to the proper 'libcurl' as clearly this caused confusion.
1808 Yang Tse (20 Feb 2009)
1809 - Do not halt compilation when using VS2008 to build a Windows 2000 target.
1811 Daniel Stenberg (20 Feb 2009)
1812 - Linus Nielsen Feltzing reported and helped me repeat and fix a problem with
1813 FTP with the multi interface: when a transfer fails, like when aborted by a
1814 write callback, the control connection was wrongly closed and thus not
1817 This change is also an attempt to cleanup the code somewhat in this area, as
1818 now the FTP code attempts to keep (better) track on pending responses
1819 necessary to get read in ftp_done().
1821 Daniel Stenberg (19 Feb 2009)
1822 - Patrik Thunstrom reported a problem and helped me repeat it. It turned out
1823 libcurl did a superfluous 1000ms wait when doing SFTP downloads!
1825 We read data with libssh2 while doing the "DO" operation for SFTP and then
1826 when we were about to start getting data for the actual file part, the
1827 "TRANSFER" part, we waited for socket action (in 1000ms) before doing a
1828 libssh2-read. But in this case libssh2 had already read and buffered the
1829 data so we ended up always just waiting 1000ms before we get working on the
1832 Patrick Monnerat (18 Feb 2009)
1833 - FTP downloads (i.e.: RETR) ending with code 550 now return error
1834 CURLE_REMOTE_FILE_NOT_FOUND instead of CURLE_FTP_COULDNT_RETR_FILE.
1836 Daniel Stenberg (17 Feb 2009)
1837 - Kamil Dudka made NSS-powered builds compile and run again!
1839 - A second follow-up change by Andre Guibert de Bruet to fix a related memory
1840 leak like that fixed on the 14th. When zlib returns failure, we need to
1841 cleanup properly before returning error.
1843 - CURLOPT_FTP_CREATE_MISSING_DIRS can now be set to 2 in addition to 1 for
1844 plain FTP connections, and it will then allow MKD to fail once and retry the
1845 CWD afterwards. This is especially useful if you're doing many simultanoes
1846 connections against the same server and they all have this option enabled,
1847 as then CWD may first fail but then another connection does MKD before this
1848 connection and thus MKD fails but trying CWD works! The numbers can
1849 (should?) now be set with the convenience enums now called
1850 CURLFTP_CREATE_DIR and CURLFTP_CREATE_DIR_RETRY.
1852 Tests has proven that if you're making an application that uploads a set of
1853 files to an ftp server, you will get a noticable gain in speed if you're
1854 using multiple connections and this option will be then be very useful.
1856 Daniel Stenberg (14 Feb 2009)
1857 - Andre Guibert de Bruet found and fixed a memory leak in the content encoding
1858 code, which could happen on libz errors.
1860 Daniel Fandrich (12 Feb 2009)
1861 - Added support for Digest and NTLM authentication using GnuTLS.
1863 Daniel Stenberg (11 Feb 2009)
1864 - CURLINFO_CONDITION_UNMET was added to allow an application to get to know if
1865 the condition in the previous request was unmet. This is typically a time
1866 condition set with CURLOPT_TIMECONDITION and was previously not possible to
1867 reliably figure out. From bug report #2565128
1868 (http://curl.haxx.se/bug/view.cgi?id=2565128) filed by Jocelyn Jaubert.
1870 Daniel Fandrich (4 Feb 2009)
1871 - Don't add the standard /usr/lib or /usr/include paths to LDFLAGS and CPPFLAGS
1872 (respectively) when --with-ssl=/usr is used (patch based on FreeBSD).
1874 - Added an explicit buffer limit check in msdosify() (patch based on FreeBSD).
1875 This couldn't ever overflow in curl, but might if the code were used
1876 elsewhere or under different conditions.
1878 Daniel Stenberg (3 Feb 2009)
1879 - Hidemoto Nakada provided a small fix that makes it possible to get the
1880 CURLINFO_CONTENT_LENGTH_DOWNLOAD size from file:// "transfers" with
1881 CURLOPT_NOBODY set true.
1883 Daniel Stenberg (2 Feb 2009)
1884 - Patrick Scott found a rather large memory leak when using the multi
1885 interface and setting CURLMOPT_MAXCONNECTS to something less than the number
1886 of handles you add to the multi handle. All the connections that didn't fit
1887 in the cache would not be properly disconnected nor freed!
1889 - Craig A West brought us: libcurl now defaults to do CONNECT with HTTP
1890 version 1.1 instead of 1.0 like before. This change also introduces the new
1891 proxy type for libcurl called 'CURLPROXY_HTTP_1_0' that then allows apps to
1892 switch (back) to CONNECT 1.0 requests. The curl tool also got a --proxy1.0
1893 option that works exactly like --proxy but sets CURLPROXY_HTTP_1_0.
1895 I updated all test cases cases that use CONNECT and I tried to do some using
1896 --proxy1.0 and some updated to do CONNECT 1.1 to get both versions run.
1898 Daniel Stenberg (31 Jan 2009)
1899 - When building with c-ares 1.6.1 (not yet released) or later and IPv6 support
1900 enabled, we can now take advantage of its brand new AF_UNSPEC support in
1901 ares_gethostbyname(). This makes test case 241 finally run fine for me with
1902 this setup since it now parses the "::1 ip6-localhost" line fine in my
1905 Daniel Stenberg (30 Jan 2009)
1906 - Scott Cantor filed bug report #2550061
1907 (http://curl.haxx.se/bug/view.cgi?id=2550061) mentioning that I failed to
1908 properly make sure that the VC9 makefiles got included in the latest
1909 release. I've now fixed the release script and verified it so next release
1910 will hopefully include them properly!
1912 Daniel Fandrich (30 Jan 2009)
1913 - Fixed --disable-proxy for FTP and SOCKS. Thanks to Daniel Egger for
1916 Yang Tse (29 Jan 2009)
1917 - Introduced curl_sspi.c and curl_sspi.h for the implementation of functions
1918 Curl_sspi_global_init() and Curl_sspi_global_cleanup() which previously were
1919 named Curl_ntlm_global_init() and Curl_ntlm_global_cleanup() in http_ntlm.c
1920 Also adjusted socks_sspi.c to remove the link-time dependency on the Windows
1921 SSPI library using it now in the same way as it was done in http_ntlm.c.
1923 Daniel Stenberg (28 Jan 2009)
1924 - Markus Moeller introduced two new options to libcurl:
1925 CURLOPT_SOCKS5_GSSAPI_SERVICE and CURLOPT_SOCKS5_GSSAPI_NEC to allow libcurl
1926 to do GSS-style authentication with SOCKS5 proxies. The curl tool got the
1927 options called --socks5-gssapi-service and --socks5-gssapi-nec to enable
1930 Daniel Stenberg (26 Jan 2009)
1931 - Chad Monroe provided the new CURLOPT_TFTP_BLKSIZE option that allows an app
1932 to set desired block size to use for TFTP transfers instead of the default
1935 - The "-no_ticket" option was introduced in Openssl0.9.8j. It's a flag to
1936 disable "rfc4507bis session ticket support". rfc4507bis was later turned
1937 into the proper RFC5077 it seems: http://tools.ietf.org/html/rfc5077
1939 The enabled extension concerns the session management. I wonder how often
1940 libcurl stops a connection and then resumes a TLS session. also, sending the
1941 session data is some overhead. .I suggest that you just use your proposed
1942 patch (which explicitly disables TICKET).
1944 If someone writes an application with libcurl and openssl who wants to
1945 enable the feature, one can do this in the SSL callback.
1947 Sharad Gupta brought this to my attention. Peter Sylvester helped me decide
1948 on the proper action.
1950 - Alexey Borzov filed bug report #2535504
1951 (http://curl.haxx.se/bug/view.cgi?id=2535504) pointing out that realms with
1952 quoted quotation marks in HTTP Digest headers didn't work. I've now added
1953 test case 1095 that verifies my fix.
1955 - Craig A West brought CURLOPT_NOPROXY and the corresponding --noproxy option.
1956 They basically offer the same thing the NO_PROXY environment variable only
1957 offered previously: list a set of host names that shall not use the proxy
1958 even if one is specified.
1960 Daniel Fandrich (20 Jan 2009)
1961 - Call setlocale() for libtest tests to test the effects of locale-induced
1962 libc changes on libcurl.
1964 - Fixed a couple more locale-dependent toupper conversions, mainly for
1965 clarity. This does fix one problem that causes ;type=i FTP URLs
1966 to fail in the Turkish locale when CURLOPT_PROXY_TRANSFER_MODE is
1967 used (test case 561)
1969 - Added tests 561 and 1091 through 1094 to test various combinations
1970 of ;type= and ;mode= URLs that could potentially fail in the Turkish
1973 Daniel Stenberg (20 Jan 2009)
1974 - Lisa Xu pointed out that the ssh.obj file was missing from the
1975 lib/Makefile.vc6 file (and thus from the vc8 and vc9 ones too).
1977 Version 7.19.3 (19 January 2009)
1979 Daniel Stenberg (16 Jan 2009)
1980 - Andrew de los Reyes fixed curlbuild.h for "generic" gcc builds on PPC, both
1983 Daniel Stenberg (15 Jan 2009)
1984 - Tim Ansell fixed a compiler warning in lib/cookie.c
1986 Daniel Stenberg (14 Jan 2009)
1987 - Grant Erickson fixed timeouts for TFTP such that specifying a
1988 connect-timeout, a max-time or both options work correctly and as expected
1989 by passing the correct boolean value to Curl_timeleft via the
1990 'duringconnect' parameter.
1992 With this small change, curl TFTP now behaves as expected (and likely as
1993 originally-designed):
1995 1) For non-existent or unreachable dotted IP addresses:
1997 a) With no options, follows the default curl 300s timeout...
1998 b) With --connect-timeout only, follows that value...
1999 c) With --max-time only, follows that value...
2000 d) With both --connect-timeout and --max-time, follows the smaller value...
2002 and times out with a "curl: (7) Couldn't connect to server" error.
2004 2) For transfers to/from a valid host:
2006 a) With no options, follows default curl 300s timeout for the
2007 first XRQ/DATA/ACK transaction and the default TFTP 3600s
2008 timeout for the remainder of the transfer...
2010 b) With --connect-time only, follows that value for the
2011 first XRQ/DATA/ACK transaction and the default TFTP 3600s
2012 timeout for the remainder of the transfer...
2014 c) With --max-time only, follows that value for the first
2015 XRQ/DATA/ACK transaction and for the remainder of the
2018 d) With both --connect-timeout and --max-time, follows the former
2019 for the first XRQ/DATA/ACK transaction and the latter for the
2020 remainder of the transfer...
2022 and times out with a "curl: (28) Timeout was reached" error as
2025 Daniel Stenberg (13 Jan 2009)
2026 - Michael Wallner fixed a NULL pointer deref when calling
2027 curl_easy_setup(curl, CURLOPT_COOKIELIST, "SESS") on a CURL handle with no
2030 - Stefan Teleman brought a patch to fix the default curlbuild.h file for the
2033 Daniel Stenberg (12 Jan 2009)
2034 - Based on bug report #2498665 (http://curl.haxx.se/bug/view.cgi?id=2498665)
2035 by Daniel Black, I've now added magic to the configure script that makes it
2036 use pkg-config to detect gnutls details as well if the existing method
2037 (using libgnutls-config) fails. While doing this, I cleaned up and unified
2038 the pkg-config usage when detecting openssl and nss as well.
2040 Daniel Stenberg (11 Jan 2009)
2041 - Karl Moerder brought the patch that creates vc9 Makefiles, and I made
2042 'maketgz' now use the actual makefile targets to do the VC8 and VC9
2045 Daniel Stenberg (10 Jan 2009)
2046 - Emil Romanus fixed:
2048 When using the multi interface over HTTP and the server returns a Location
2049 header, the running easy handle will get stuck in the CURLM_STATE_PERFORM
2050 state, leaving the external event loop stuck waiting for data from the
2051 ingoing socket (when using the curl_multi_socket_action stuff). While this
2052 bug was pretty hard to find, it seems to require only a one-line fix. The
2053 break statement on line 1374 in multi.c caused the function to skip the call
2056 How to reproduce this bug? Well, that's another question. evhiperfifo.c in
2057 the examples directory chokes on this bug only _sometimes_, probably
2058 depending on how fast the URLs are added. One way of testing the bug out is
2059 writing to hiper.fifo from more than one source at the same time.
2061 Daniel Fandrich (7 Jan 2009)
2062 - Unified much of the SessionHandle initialization done in Curl_open() and
2063 curl_easy_reset() by creating Curl_init_userdefined(). This had the side
2064 effect of fixing curl_easy_reset() so it now also resets
2065 CURLOPT_FTP_FILEMETHOD and CURLOPT_SSL_SESSIONID_CACHE
2067 Daniel Stenberg (7 Jan 2009)
2068 - Rob Crittenden did once again provide an NSS update:
2070 I have to jump through a few hoops now with the NSS library initialization
2071 since another part of an application may have already initialized NSS by the
2072 time Curl gets invoked. This patch is more careful to only shutdown the NSS
2073 library if Curl did the initialization.
2075 It also adds in a bit of code to set the default ciphers if the app that
2076 call NSS_Init* did not call NSS_SetDomesticPolicy() or set specific
2077 ciphers. One might argue that this lets other application developers get
2078 lazy and/or they aren't using the NSS API correctly, and you'd be right.
2079 But still, this will avoid terribly difficult-to-trace crashes and is
2082 Daniel Stenberg (1 Jan 2009)
2083 - 'reconf' is removed since we rather have users use 'buildconf'
2085 Daniel Stenberg (31 Dec 2008)
2086 - Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
2087 out that 'reconf' didn't properly point out the m4 subdirectory when running
2090 Daniel Stenberg (29 Dec 2008)
2091 - Phil Lisiecki filed bug report #2413067
2092 (http://curl.haxx.se/bug/view.cgi?id=2413067) that identified a problem that
2093 would cause libcurl to mark a DNS cache entry "in use" eternally if the
2094 subsequence TCP connect failed. It would thus never get pruned and refreshed
2095 as it should've been.
2097 Phil provided his own patch to this problem that while it seemed to work
2098 wasn't complete and thus I wrote my own fix to the problem.
2100 Daniel Stenberg (28 Dec 2008)
2101 - Peter Korsgaard fixed building libcurl with "configure --with-ssl
2104 - Anthony Bryan fixed more language and spelling flaws in man pages.
2106 Daniel Stenberg (22 Dec 2008)
2107 - Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
2108 on file indexes beyond 2 or 4GB.
2110 - Anthony Bryan provided a set of patches that cleaned up manual language,
2111 corrected spellings and more.
2113 Daniel Stenberg (20 Dec 2008)
2114 - Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
2115 pipelining, as libcurl could then easily get confused and A) work on the
2116 handle that was not "first in queue" on a pipeline, or even B) tell the app
2117 to REMOVE a socket while it was in use by a second handle in a pipeline. Both
2118 errors caused hanging or stalling applications.
2120 Daniel Stenberg (19 Dec 2008)
2121 - curl_multi_timeout() could return a timeout value of 0 even though nothing
2122 was actually ready to get done, as the internal time resolution is higher
2123 than the returned millisecond timer. Therefore it could cause applications
2124 running on fast processors to do short bursts of busy-loops.
2125 curl_multi_timeout() will now only return 0 if the timeout is actually
2128 - Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
2129 now has an improved ability to do right when the multi interface (both
2130 "regular" and multi_socket) is used for SCP and SFTP transfers. This should
2131 result in (much) less busy-loop situations and thus less CPU usage with no
2134 Daniel Stenberg (17 Dec 2008)
2135 - SCP and SFTP with the multi interface had the same flaw: the 'DONE'
2136 operation didn't complete properly if the EAGAIN equivalent was returned but
2137 libcurl would simply continue with a half-completed close operation
2138 performed. This ruined persistent connection re-use and cause some
2139 SSH-protocol errors in general. The correction is unfortunately adding a
2140 blocking function - doing it entirely non-blocking should be considered for
2143 Gisle Vanem (16 Dec 2008)
2144 - Added the possibility to use the Watt-32 tcp/ip stack under Windows.
2145 The change simply involved adding a USE_WATT32 section in the
2146 config-win32.h files (under ./lib and ./src). This section disables
2147 the use of any Winsock headers.
2149 Daniel Stenberg (16 Dec 2008)
2150 - libssh2_sftp_last_error() was wrongly used at some places in libcurl which
2151 made libcurl sometimes not properly abort problematic SFTP transfers.
2153 Daniel Stenberg (12 Dec 2008)
2154 - More work with Igor Novoseltsev to first fix the remaining stuff for
2155 removing easy handles from multi handles when the easy handle is/was within
2156 a HTTP pipeline. His bug report #2351653
2157 (http://curl.haxx.se/bug/view.cgi?id=2351653) was also related and was
2158 eventually fixed by a patch by Igor himself.
2160 Yang Tse (12 Dec 2008)
2161 - Patrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
2162 OS/400 compilations with IPv6 enabled.
2164 Daniel Stenberg (12 Dec 2008)
2165 - Mark Karpeles filed bug report #2416182 titled "crash in ConnectionExists
2166 when using duphandle+curl_mutli"
2167 (http://curl.haxx.se/bug/view.cgi?id=2416182) which showed that
2168 curl_easy_duphandle() wrongly also copied the pointer to the connection
2169 cache, which was plain wrong and caused a segfault if the handle would be
2170 used in a different multi handle than the handle it was duplicated from.
2172 Daniel Stenberg (11 Dec 2008)
2173 - Keshav Krity found out that libcurl failed to deal with dotted IPv6
2174 addresses if they were very long (>39 letters) due to a too strict address
2175 validity parser. It now accepts addresses up to 45 bytes long.
2177 Daniel Stenberg (11 Dec 2008)
2178 - Internet Explorer had a broken HTTP digest authentication before v7 and
2179 there are servers "out there" that relies on the client doing this broken
2180 Digest authentication. Apache even comes with an option to work with such
2183 The difference is only for URLs that contain a query-part (a '?'-letter and
2184 text to the right of it).
2186 libcurl now supports this quirk, and you enable it by setting the
2187 CURLAUTH_DIGEST_IE bit in the bitmask you pass to the CURLOPT_HTTPAUTH or
2188 CURLOPT_PROXYAUTH options. They are thus individually controlled to server
2191 (note that there's no way to activate this with the curl tool yet)
2193 Daniel Fandrich (9 Dec 2008)
2194 - Added test cases 1089 and 1090 to test --write-out after a redirect to
2195 test a report that the size didn't work, but these test cases pass.
2197 - Documented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.
2199 Daniel Stenberg (9 Dec 2008)
2200 - Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
2201 particular state for the control connection like it did before for implicit
2202 FTPS (libcurl assumed such control connections to be encrypted while some
2203 FTPS servers such as FileZilla assumes such connections to be clear
2204 mode). Use the CURLOPT_USE_SSL option to set your desired level.
2206 Daniel Stenberg (8 Dec 2008)
2207 - Fred Machado posted about a weird FTP problem on the curl-users list and when
2208 researching it, it turned out he got a 550 response back from a SIZE command
2209 and then I fell over the text in RFC3659 that says:
2211 The presence of the 550 error response to a SIZE command MUST NOT be taken
2212 by the client as an indication that the file cannot be transferred in the
2213 current MODE and TYPE.
2215 In other words: the change I did on September 30th 2008 and that has been
2216 included in the last two releases were a regression and a bad idea. We MUST
2217 NOT take a 550 response from SIZE as a hint that the file doesn't exist.
2219 - Christian Krause filed bug #2221237
2220 (http://curl.haxx.se/bug/view.cgi?id=2221237) that identified an infinite
2221 loop during GSS authentication given some specific conditions. With his
2222 patience and great feedback I managed to narrow down the problem and
2223 eventually fix it although I can't test any of this myself!
2225 Daniel Fandrich (3 Dec 2008)
2226 - Fixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
2227 support (e.g. Minix)
2229 Daniel Stenberg (3 Dec 2008)
2230 - Igor Novoseltsev filed bug #2351645
2231 (http://curl.haxx.se/bug/view.cgi?id=2351645) that identified a problem with
2232 the multi interface that occured if you removed an easy handle while in
2233 progress and the handle was used in a HTTP pipeline.
2235 - Pawel Kierski pointed out a mistake in the cookie code that could lead to a
2236 bad fclose() after a fatal error had occured.
2237 (http://curl.haxx.se/bug/view.cgi?id=2382219)
2239 Daniel Fandrich (25 Nov 2008)
2240 - If a HTTP request is Basic and num is already >=1000, the HTTP test
2241 server adds 1 to num to get the data section to return. This allows
2242 testing authentication negotiations using the Basic authentication
2245 - Added tests 1087 and 1088 to test Basic authentication on a redirect
2246 with and without --location-trusted
2248 Daniel Stenberg (24 Nov 2008)
2249 - Based on a patch by Vlad Grachov, libcurl now uses a new libssh2 0.19
2250 function when built to support SCP and SFTP that helps the library to know
2251 in which direction a particular libssh2 operation would return EAGAIN so
2252 that libcurl knows what socket conditions to wait for before trying the
2253 function call again. Previously (and still when using libssh2 0.18 or
2254 earlier), libcurl will busy-loop in this situation when the easy interface
2257 Daniel Fandrich (20 Nov 2008)
2258 - Automatically detect OpenBSD's CA cert bundle.
2260 Daniel Stenberg (19 Nov 2008)
2261 - I removed the default use of "Pragma: no-cache" from libcurl when a proxy is
2262 used. It has been used since forever but it was never a good idea to use
2263 unless explicitly asked for.
2265 - Josef Wolf's extension that allows a $TESTDIR/gdbinit$testnum file that when
2266 you use runtests.pl -g, will be sourced by gdb to allow additional fancy or
2267 whatever you see fit
2269 - Christian Krause reported and fixed a memory leak that would occur with HTTP
2270 GSS/kerberos authentication (http://curl.haxx.se/bug/view.cgi?id=2284386)
2272 - Andreas Wurf and Markus Koetter helped me analyze a problem that Andreas got
2273 when uploading files to a single FTP server using multiple easy handle
2274 handles with the multi interface. Occasionally a handle would stall in
2277 The problem turned out to be a side-effect of the ConnectionExists()
2278 function's eagerness to re-use a handle for HTTP pipelining so it would
2279 select it even if already being in use, due to an inadequate check for its
2280 chances of being used for pipelnining.
2282 Daniel Fandrich (17 Nov 2008)
2283 - Added more compiler warning options for gcc 4.3
2285 Yang Tse (17 Nov 2008)
2286 - Fix a remaining problem in the inet_pton() runtime configure check. And
2287 fix internal Curl_inet_pton() failures to reject certain malformed literals.
2289 - Make configure script check if ioctl with the SIOCGIFADDR command can be
2290 used, and define HAVE_IOCTL_SIOCGIFADDR if appropriate.
2292 Daniel Stenberg (16 Nov 2008)
2293 - Christian Krause fixed a build failure when building with gss support
2294 enabled and FTP disabled.
2296 - Added check for NULL returns from strdup() in src/main.c and lib/formdata.c
2297 - reported by Jim Meyering also prevent buffer overflow on MSDOS when you do
2298 for example -O on a url with a file name part longer than PATH_MAX letters
2300 - lib/nss.c fixes based on the report by Jim Meyering: I went over and added
2301 checks for return codes for all calls to malloc and strdup that were
2302 missing. I also changed a few malloc(13) to use arrays on the stack and a
2303 few malloc(PATH_MAX) to instead use aprintf() to lower memory use.
2305 - I fixed a memory leak in Curl_nss_connect() when CURLOPT_ISSUERCERT is
2308 Daniel Fandrich (14 Nov 2008)
2309 - Added .xml as one of the few common file extensions known by the multipart
2312 - Added some #ifdefs around header files and change the EAGAIN test to
2313 fix compilation on Cell (reported by Jeff Curley).
2315 Yang Tse (14 Nov 2008)
2316 - Fixed several configure script issues affecting checks for inet_ntoa_r(),
2317 inet_ntop(), inet_pton(), getifaddrs(), fcntl() and getaddrinfo().
2319 Yang Tse (13 Nov 2008)
2320 - Refactored configure script detection of functions used to set sockets into
2321 non-blocking mode, and decouple function detection from function capability.