5 \___|\___/|_| \_\_____|
9 Kamil Dudka (19 Mar 2010)
10 - Improved Curl_read() to not ignore the error returned from Curl_ssl_recv().
12 Daniel Stenberg (15 Mar 2010)
13 - Constantine Sapuntzakis brought a patch:
15 The problem mentioned on Dec 10 2009
16 (http://curl.haxx.se/bug/view.cgi?id=2905220) was only partially fixed.
17 Partially because an easy handle can be associated with many connections in
18 the cache (e.g. if there is a redirect during the lifetime of the easy
19 handle). The previous patch only cleaned up the first one. The new fix now
20 removes the easy handle from all connections, not just the first one.
22 Daniel Stenberg (6 Mar 2010)
23 - Ben Greear brought a patch that fixed the rate limiting logic for TFTP when
24 the easy interface was used.
26 Daniel Stenberg (5 Mar 2010)
27 - Daniel Johnson provided fixes for building curl with the clang compiler.
30 - Constantine Sapuntzakis detected and fixed a double free in builds done
31 with threaded resolver enabled (Windows default configuration) that would
32 get triggered when a curl handle is closed while doing DNS resolution.
34 Daniel Stenberg (2 Mar 2010)
35 - [Daniel Johnson] I've been trying to build libcurl with clang on Darwin and
36 ran into some issues with the GSSAPI tests in configure.ac. The tests first
37 try to determine the include dirs and libs and set CPPFLAGS and LIBS
38 accordingly. It then checks for the headers and finally sets LIBS a second
39 time, causing the libs to be included twice. The first setting of LIBS seems
40 redundant and should be left out, since the first part is otherwise just
41 about finding headers.
43 My second issue is that 'krb5-config --libs gssapi' on Darwin is less than
44 useless and returns junk that, while it happens to work with gcc, causes
45 clang to choke. For example, --libs returns $CFLAGS along with the libs,
46 which is really retarded. Simply setting 'LIBS="$LIBS -lgssapi_krb5
47 -lresolv"' on Darwin is sufficient.
49 - Based on patch provided by Jacob Moshenko, the transfer logic now properly
50 makes sure that when using sub-second timeouts, there's no final bad 1000ms
51 wait. Previously, a sub-second timeout would often make the elapsed time end
52 up the time rounded up to the nearest second (e.g. 1s for 200ms timeout)
54 - Andrei Benea filed bug report #2956698 and pointed out that the
55 CURLOPT_CERTINFO feature leaked memory due to a missing OpenSSL function
56 call. He provided the patch to fix it too.
58 http://curl.haxx.se/bug/view.cgi?id=2956698
60 - Markus Duft pointed out in bug #2961796 that even though Interix has a
61 poll() function it doesn't quite work the way we want it so we must disable
62 it, and he also provided a patch for it.
64 http://curl.haxx.se/bug/view.cgi?id=2961796
66 - Made the pingpong timeout code properly deal with the response timeout AND
67 the global timeout if set. Also, as was reported in the bug report #2956437
68 by Ryan Chan, the time stamp to use as basis for the per command timeout was
69 not set properly in the DONE phase for FTP (and not for SMTP) so I fixed
70 that just now. This was a regression compared to 7.19.7 due to the
71 conversion of FTP code over to the generic pingpong concepts.
73 http://curl.haxx.se/bug/view.cgi?id=2956437
75 Daniel Stenberg (1 Mar 2010)
76 - Ben Greear provided an update for TFTP that fixes upload.
78 - Wesley Miaw reported bug #2958179 which identified a case of looping during
79 OpenSSL based SSL handshaking even though the multi interface was used and
80 there was no good reason for it.
82 http://curl.haxx.se/bug/view.cgi?id=2958179
84 Daniel Stenberg (26 Feb 2010)
85 - Pat Ray in bug #2958474 pointed out an off-by-one case when receiving a
86 chunked-encoding trailer.
88 http://curl.haxx.se/bug/view.cgi?id=2958474
90 Daniel Fandrich (25 Feb 2010)
91 - Fixed a couple of out of memory leaks and a segfault in the SMTP & IMAP code.
93 Yang Tse (25 Feb 2010)
94 - I fixed bug report #2958074 indicating
95 (http://curl.haxx.se/bug/view.cgi?id=2958074) that curl on Windows with
96 option --trace-time did not use local time when timestamping trace lines.
97 This could also happen on other systems depending on time souurce.
99 Patrick Monnerat (22 Feb 2010)
100 - Proper handling of STARTTLS on SMTP, taking CURLUSESSL_TRY into account.
101 - SMTP falls back to RFC821 HELO when EHLO fails (and SSL is not required).
102 - Use of true local host name (i.e.: via gethostname()) when available, as
103 default argument to SMTP HELO/EHLO.
104 - Test case 804 for HELO fallback.
106 Daniel Stenberg (20 Feb 2010)
107 - Fixed the SMTP compliance by making sure RCPT TO addresses are specified
108 properly in angle brackets. Recipients provided with CURLOPT_MAIL_RCPT now
109 get angle bracket wrapping automatically by libcurl unless the recipient
110 starts with an angle bracket as then the app is assumed to deal with that
113 - I made the SMTP code expect a 250 response back from the server after the
114 full DATA has been sent, and I modified the test SMTP server to also send
115 that response. As usual, the DONE operation that is made after a completed
116 transfer is still not doable in a non-blocking way so this waiting for 250
117 is unfortunately made blockingly.
119 Yang Tse (14 Feb 2010)
120 - Overhauled test suite getpart() function. Fixing potential out of bounds
121 stack and memory overwrites triggered with huge test case definitions.
123 Daniel Stenberg (13 Feb 2010)
124 - Martin Hager reported and fixed a problem with a missing quote in libcurl.m4
126 (http://curl.haxx.se/bug/view.cgi?id=2951319)
128 - Tom Donovan fixed the CURL_FORMAT_* defines when building with cmake.
130 (http://curl.haxx.se/bug/view.cgi?id=2951269)
132 Daniel Stenberg (12 Feb 2010)
133 - Jack Zhang reported a problem with SMTP: we wrongly used multiple addresses
134 in the same RCPT TO line, when they should be sent in separate single
135 commands. I updated test case 802 to verify this.
137 - I also fixed a bad use of my_setopt_str() of CURLOPT_MAIL_RCPT in the curl
138 tool which made it try to output it as string for the --libcurl feature
139 which could lead to crashes.
141 Yang Tse (11 Feb 2010)
142 - Steven M. Schweda fixed VMS builder bad behavior when used in a batch job,
143 removed obsolete batch_compile.com and defines.com and updated VMS readme.
145 Version 7.20.0 (9 February 2010)
147 Daniel Stenberg (9 Feb 2010)
148 - When downloading compressed content over HTTP and the app asked libcurl to
149 automatically uncompress it with the CURLOPT_ENCODING option, libcurl could
150 wrongly provide the callback with more data than the maximum documented
151 amount. An application could thus get tricked into badness if the maximum
152 limit was trusted to be enforced by libcurl itself (as it is documented).
154 This is further detailed and explained in the libcurl security advisory
157 http://curl.haxx.se/docs/adv_20100209.html
159 Daniel Fandrich (3 Feb 2010)
160 - Changed the Watcom makefiles to make them easier to keep in sync with
161 Makefile.inc since that can't be included directly.
163 Yang Tse (2 Feb 2010)
164 - Symbol CURL_FORMAT_OFF_T now obsoleted, will be removed in a future release,
165 symbol will not be available when building with CURL_NO_OLDIES defined. Use
166 of CURL_FORMAT_CURL_OFF_T is preferred since 7.19.0
168 Daniel Stenberg (1 Feb 2010)
169 - Using the multi_socket API, it turns out at times it seemed to "forget"
170 connections (which caused a hang). It turned out to be an existing (7.19.7)
171 bug in libcurl (that's been around for a long time) and it happened like
174 The app calls curl_multi_add_handle() to add a new easy handle, libcurl will
175 then set it to timeout in 1 millisecond so libcurl will tell the app about
178 The app's timeout fires off that there's a timeout, the app calls libcurl as
179 we so often document it:
182 res = curl_multi_socket_action(... TIMEOUT ...);
183 } while(CURLM_CALL_MULTI_PERFORM == res);
185 And this is the problem number one:
187 When curl_multi_socket_action() is called with no specific handle, but only
188 a timeout-action, it will *only* perform actions within libcurl that are
189 marked to run at this time. In this case, the request would go from INIT to
190 CONNECT and return CURLM_CALL_MULTI_PERFORM. When the app then calls libcurl
191 again, there's no timer set for this handle so it remains in the CONNECT
192 state. The CONNECT state is a transitional state in libcurl so it reports no
193 sockets there, and thus libcurl never tells the app anything more about that
194 easy handle/connection.
196 libcurl _does_ set a 1ms timeout for the handle at the end of
197 multi_runsingle() if it returns CURLM_CALL_MULTI_PERFORM, but since the loop
198 is instant the new job is not ready to run at that point (and there's no
199 code that makes libcurl call the app to update the timout for this new
200 timeout). It will simply rely on that some other timeout will trigger later
201 on or that something else will update the timeout callback. This makes the
202 bug fairly hard to repeat.
204 The fix made to adress this issue:
206 We introduce a loop in lib/multi.c around all calls to multi_runsingle() and
207 simply check for CURLM_CALL_MULTI_PERFORM internally. This has the added
208 benefit that this goes in line with my long-term wishes to get rid of the
209 CURLM_CALL_MULTI_PERFORM all together from the public API.
211 The downside of this fix, is that the counter we return in 'running_handles'
212 in several of our public functions then gets a slightly new and possibly
213 confusing behavior during times:
215 If an app adds a handle that fails to connect (very quickly) it may just
216 as well never appear as a 'running_handle' with this fix. Previously it
217 would first bump the counter only to get it decreased again at next call.
218 Even I have used that change in handle counter to signal "end of a
219 transfer". The only *good* way to find the end of a individual transfer
220 is calling curl_multi_info_read() to see if it returns one.
222 Of course, if the app previously did the looping before it checked the
223 counter, it really shouldn't be any new effect.
225 Yang Tse (26 Jan 2010)
226 - Constantine Sapuntzakis' and Joshua Kwan's work done in the last four months
227 relative to the asynchronous DNS lookups, along with with some integration
228 adjustments I have done are finally committed to CVS.
230 Currently these enhancements will benefit builds done using c-ares on any
231 platform as well as Windows builds using the default threaded resolver.
233 This release does not make generally available POSIX threaded DNS lookups
234 yet. There is no configure option to enable this feature yet. It is possible
235 to experimantally try this feature running configure with compiler flags that
236 make simultaneous definition of preprocessor symbols USE_THREADS_POSIX and
237 HAVE_PTHREAD_H, as well as whatever reentrancy compiler flags and linker ones
238 are required to link and properly use pthread_* functions on each platform.
240 Daniel Stenberg (26 Jan 2010)
241 - Mike Crowe made libcurl return CURLE_COULDNT_RESOLVE_PROXY when it is the
242 proxy that cannot be resolved when using c-ares. This matches the behaviour
243 when not using c-ares.
245 Björn Stenberg (23 Jan 2010)
246 - Added a new flag: -J/--remote-header-name. This option tells the
247 -O/--remote-name option to use the server-specified Content-Disposition
248 filename instead of extracting a filename from the URL.
250 Daniel Stenberg (21 Jan 2010)
251 - Chris Conroy brought support for RTSP transfers, and with it comes 8(!) new
252 libcurl options for controlling what to get and how to receive posssibly
253 interleaved RTP data.
255 Daniel Stenberg (20 Jan 2010)
256 - As was pointed out on the http-state mailing list, the order of cookies in a
257 HTTP Cookie: header _needs_ to be sorted on the path length in the cases
258 where two cookies using the same name are set more than once using
259 (overlapping) paths. Realizing this, identically named cookies must be
260 sorted correctly. But detecting only identically named cookies and take care
261 of them individually is harder than just to blindly and unconditionally sort
262 all cookies based on their path lengths. All major browsers also already do
263 this, so this makes our behavior one step closer to them in the cookie area.
265 Test case 8 was the only one that broke due to this change and I updated it
268 Daniel Stenberg (19 Jan 2010)
269 - David McCreedy brought a fix and a new test case (129) to make libcurl work
270 again when downloading files over FTP using ASCII and it turns out that the
271 final size of the file is not the same as the initial size the server
272 reported. This is very common since servers don't take the newline
273 conversions into account.
275 Kamil Dudka (14 Jan 2010)
276 - Suppressed side effect of OpenSSL configure checks, which prevented NSS from
277 being properly detected under certain circumstances. It had been caused by
278 strange behavior of pkg-config when handling PKG_CONFIG_LIBDIR. pkg-config
279 distinguishes among empty and non-existent environment variable in that case.
281 Daniel Stenberg (12 Jan 2010)
282 - Gil Weber reported a peculiar flaw with the multi interface when doing SFTP
283 transfers: curl_multi_fdset() would return -1 and not set and file
284 descriptors several times during a transfer of a single file. It turned out
285 to be due to two different flaws now fixed. Gil's excellent recipe helped me
288 Daniel Stenberg (11 Jan 2010)
289 - Made sure that the progress callback is repeatedly called at a regular
290 interval even during very slow connects.
292 - The tests/runtests.pl script now checks to see if the test case that runs is
293 present in the tests/data/Makefile.am and outputs a notice message on the
294 screen if not. Each test file has to be included in that Makefile.am to get
295 included in release archives and forgetting to add files there is a common
296 mistake. This is an attempt to make it harder to forget.
298 Daniel Stenberg (9 Jan 2010)
299 - Johan van Selst found and fixed a OpenSSL session ref count leak:
301 ossl_connect_step3() increments an SSL session handle reference counter on
302 each call. When sessions are re-used this reference counter may be
303 incremented many times, but it will be decremented only once when done (by
304 Curl_ossl_session_free()); and the internal OpenSSL data will not be freed
305 if this reference count remains positive. When a session is re-used the
306 reference counter should be corrected by explicitly calling
307 SSL_SESSION_free() after each consecutive SSL_get1_session() to avoid
308 introducing a memory leak.
310 (http://curl.haxx.se/bug/view.cgi?id=2926284)
312 Daniel Stenberg (7 Jan 2010)
313 - Make sure the progress callback is called repeatedly even during very slow
314 name resolves when c-ares is used for resolving.
316 Claes Jakobsson (6 Jan 2010)
317 - Julien Chaffraix fixed so that the fragment part in an URL is not sent
318 to the server anymore.
320 Kamil Dudka (3 Jan 2010)
321 - Julien Chaffraix eliminated a duplicated initialization in singlesocket().
323 Daniel Stenberg (2 Jan 2010)
324 - Make curl support --ssl and --ssl-reqd instead of the previous FTP-specific
325 versions --ftp-ssl and --ftp-ssl-reqd as these options are now used to
326 control SSL/TLS for IMAP, POP3 and SMTP as well in addition to FTP. The old
327 option names are still working but the new ones are the ones listed and
330 Daniel Stenberg (1 Jan 2010)
331 - Ingmar Runge enhanced libcurl's FTP engine to support the PRET command. This
332 command is a special "hack" used by the drftpd server, but even though it is
333 a custom extension I've deemed it fine to add to libcurl since this server
334 seems to survive and people keep using it and want libcurl to support
335 it. The new libcurl option is named CURLOPT_FTP_USE_PRET, and it is also
336 usable from the curl tool with --ftp-pret. Using this option on a server
337 that doesn't support this command will make libcurl fail.
339 I added test cases 1107 and 1108 to verify the functionality.
341 The PRET command is documented at
342 http://www.drftpd.org/index.php/Distributed_PASV
344 Yang Tse (30 Dec 2009)
345 - Steven M. Schweda improved VMS build system, and Craig A. Berry helped
346 with the patch and testing.
348 Daniel Stenberg (26 Dec 2009)
349 - Renato Botelho and Peter Pentchev brought a patch that makes the libcurl
350 headers work correctly even on FreeBSD systems before v8.
352 (http://curl.haxx.se/bug/view.cgi?id=2916915)
354 Daniel Stenberg (17 Dec 2009)
355 - David Byron fixed Curl_ossl_cleanup to actually call ENGINE_cleanup when
358 - Follow-up fix for the proxy fix I did for Jon Nelson's bug. It turned out I
359 was a bit too quick and broke test case 1101 with that change. The order of
360 some of the setups is sensitive. I now changed it slightly again to make
361 sure we do them in this order:
363 1 - parse URL and figure out what protocol is used in the URL
364 2 - prepend protocol:// to URL if missing
365 3 - parse name+password off URL, which needs to know what protocol is used
366 (since only some allows for name+password in the URL)
367 4 - figure out if a proxy should be used set by an option
368 5 - if no proxy option, check proxy environment variables
369 6 - run the protocol-specific setup function, which needs to have the proxy
372 Daniel Stenberg (15 Dec 2009)
373 - Jon Nelson found a regression that turned out to be a flaw in how libcurl
374 detects and uses proxies based on the environment variables. If the proxy
375 was given as an explicit option it worked, but due to the setup order
376 mistake proxies would not be used fine for a few protocols when picked up
377 from '[protocol]_proxy'. Obviously this broke after 7.19.4. I now also added
378 test case 1106 that verifies this functionality.
380 (http://curl.haxx.se/bug/view.cgi?id=2913886)
382 Daniel Stenberg (12 Dec 2009)
383 - IMAP, POP3 and SMTP support and their TLS versions (including IMAPS, POP3S
384 and SMTPS) are now supported. The current state may not yet be solid, but
385 the foundation is in place and the test suite has some initial support for
386 these protocols. Work will now persue to make them nice libcurl citizens
389 The work with supporting these new protocols was sponsored by
390 networking4all.com - thanks!
392 Daniel Stenberg (10 Dec 2009)
393 - Siegfried Gyuricsko found out that the curl manual said --retry would retry
394 on FTP errors in the transient 5xx range. Transient FTP errors are in the
395 4xx range. The code itself only tried on 5xx errors that occured _at login_.
396 Now the retry code retries on all FTP transfer failures that ended with a
399 (http://curl.haxx.se/bug/view.cgi?id=2911279)
401 - Constantine Sapuntzakis figured out a case which would lead to libcurl
402 accessing alredy freed memory and thus crash when using HTTPS (with
403 OpenSSL), multi interface and the CURLOPT_DEBUGFUNCTION and a certain order
404 of cleaning things up. I fixed it.
406 (http://curl.haxx.se/bug/view.cgi?id=2905220)
408 Daniel Stenberg (7 Dec 2009)
409 - Martin Storsjo made libcurl use the Expect: 100-continue header for posts
410 with unknown size. Previously it was only used for posts with a known size
411 larger than 1024 bytes.
413 Daniel Stenberg (1 Dec 2009)
414 - If the Expect: 100-continue header has been set by the application through
415 curl_easy_setopt with CURLOPT_HTTPHEADER, the library should set
416 data->state.expect100header accordingly - the current code (in 7.19.7 at
417 least) doesn't handle this properly. Martin Storsjo provided the fix!
419 Yang Tse (28 Nov 2009)
420 - Added Diffie-Hellman parameters to several test harness certificate files in
421 PEM format. Required by several stunnel versions used by our test harness.
423 Daniel Stenberg (28 Nov 2009)
424 - Markus Koetter provided a polished and updated version of Chad Monroe's TFTP
425 rework patch that now integrates TFTP properly into libcurl so that it can
426 be used non-blocking with the multi interface and more. BLKSIZE also works.
428 The --tftp-blksize option was added to allow setting the TFTP BLKSIZE from
431 Daniel Stenberg (26 Nov 2009)
432 - Extended and fixed the change I did on Dec 11 for the the progress
433 meter/callback during FTP command/response sequences. It turned out it was
434 really lame before and now the progress meter SHOULD get called at least
437 Daniel Stenberg (23 Nov 2009)
438 - Bjorn Augustsson reported a bug which made curl not report any problems even
439 though it failed to write a very small download to disk (done in a single
440 fwrite call). It turned out to be because fwrite() returned success, but
441 there was insufficient error-checking for the fclose() call which tricked
442 curl to believe things were fine.
444 Yang Tse (23 Nov 2009)
445 - David Byron modified Makefile.dist vc8 and vc9 targets in order to allow
446 finer granularity control when generating src and lib makefiles.
448 Yang Tse (22 Nov 2009)
449 - I modified configure to force removal of the curlbuild.h file included in
450 distribution tarballs for use by non-configure systems. As intended, this
451 would get overwriten when doing in-tree builds. But VPATH builds would end
452 having two curlbuild.h files, one in the source tree and another in the
453 build tree. With the modification I introduced 5 Nov 2009 this could become
454 an issue when running libcurl's test suite.
456 Daniel Stenberg (20 Nov 2009)
457 - Constantine Sapuntzakis identified a write after close, as the sockets were
458 closed by libcurl before the SSL lib were shutdown and they may write to its
459 socket. Detected to at least happen with OpenSSL builds.
461 - Jad Chamcham pointed out a bug with connection re-use. If a connection had
462 CURLOPT_HTTPPROXYTUNNEL enabled over a proxy, a subsequent request using the
463 same proxy with the tunnel option disabled would still wrongly re-use that
464 previous connection and the outcome would only be badness.
466 Yang Tse (18 Nov 2009)
467 - I modified the memory tracking system to make it intolerant with zero sized
468 malloc(), calloc() and realloc() function calls.
470 Daniel Stenberg (17 Nov 2009)
471 - Constantine Sapuntzakis provided another fix for the DNS cache that could
472 end up with entries that wouldn't time-out:
474 1. Set up a first web server that redirects (307) to a http://server:port
476 2. Have curl connect to the first web server using curl multi
478 After the curl_easy_cleanup call, there will be curl dns entries hanging
479 around with in_use != 0.
481 (http://curl.haxx.se/bug/view.cgi?id=2891591)
483 - Marc Kleine-Budde fixed: curl saved the LDFLAGS set during configure into
484 its pkg-config file. So -Wl stuff ended up in the .pc file, which is really
485 bad, and breaks if there are multiple -Wl in our LDFLAGS (which are in
486 PTXdist). bug #2893592 (http://curl.haxx.se/bug/view.cgi?id=2893592)
488 Kamil Dudka (15 Nov 2009)
489 - David Byron improved the configure script to use pkg-config to find OpenSSL
490 (and in particular the list of required libraries) even if a path is given
491 as argument to --with-ssl
493 Yang Tse (15 Nov 2009)
494 - I removed enable-thread / disable-thread configure option. These were only
495 placebo options. The library is always built as thread safe as possible on
498 Claes Jakobsson (14 Nov 2009)
499 - curl-config now accepts '--configure' to see what arguments was
500 passed to the configure script when building curl.
502 Daniel Stenberg (14 Nov 2009)
503 - Claes Jakobsson restored the configure functionality to detect NSS when
504 --with-nss is set but not "yes".
506 I think we can still improve that to check for pkg-config in that path etc,
507 but at least this patch brings back the same functionality we had before.
509 - Camille Moncelier added support for the file type SSL_FILETYPE_ENGINE for
510 the client certificate. It also disable the key name test as some engines
511 can select a private key/cert automatically (When there is only one key
512 and/or certificate on the hardware device used by the engine)
514 Yang Tse (14 Nov 2009)
515 - Constantine Sapuntzakis provided the fix that ensures that an SSL connection
516 won't be reused unless protection level for peer and host verification match.
518 I refactored how preprocessor symbol _THREAD_SAFE definition is done.
520 Kamil Dudka (12 Nov 2009)
521 - Kevin Baughman provided a fix preventing libcurl-NSS from crash on doubly
522 closed NSPR descriptor. The issue was hard to find, reported several times
523 before and always closed unresolved. More info at the RH bug:
524 https://bugzilla.redhat.com/534176
526 - libcurl-NSS now tries to reconnect with TLS disabled in case it detects
527 a broken TLS server. However it does not happen if SSL version is selected
528 manually. The approach was originally taken from PSM. Kaspar Brand helped me
529 to complete the patch. Original bug reports:
530 https://bugzilla.redhat.com/525496
531 https://bugzilla.redhat.com/527771
533 Yang Tse (12 Nov 2009)
534 - I modified configure script to make the getaddrinfo function check also
535 verify if the function is thread safe.
537 Yang Tse (11 Nov 2009)
538 - Marco Maggi reported that compilation failed when configured --with-gssapi
539 and GNU GSS installed due to a missing mutual exclusion of header files in
540 the Kerberos 5 code path. He also verified that my patch worked for him.
542 Daniel Stenberg (11 Nov 2009)
543 - Constantine Sapuntzakis posted bug #2891595
544 (http://curl.haxx.se/bug/view.cgi?id=2891595) which identified how an entry
545 in the DNS cache would linger too long if the request that added it was in
546 use that long. He also provided the patch that now makes libcurl capable of
547 still doing a request while the DNS hash entry may get timed out.
549 - Christian Schmitz noticed that the progress meter/callback was not properly
550 used during the FTP connection phase (after the actual TCP connect), while
551 it of course should be. I also made the speed check get called correctly so
552 that really slow servers will trigger that properly too.
554 Kamil Dudka (5 Nov 2009)
555 - Dropped misleading timeouts in libcurl-NSS and made sure the SSL socket works
556 in non-blocking mode.
558 Yang Tse (5 Nov 2009)
559 - I removed leading 'curl' path on the 'curlbuild.h' include statement in
560 curl.h, adjusting auto-makefiles include path, to enhance portability to
561 OS's without an orthogonal directory tree structure such as OS/400.
563 Daniel Stenberg (4 Nov 2009)
564 - I fixed several problems with the transfer progress meter. It showed the
565 wrong percentage for small files, most notable for <1000 bytes and could
566 easily end up showing more than 100% at the end. It also didn't show any
567 percentage, transfer size or estimated transfer times when transferring
570 Version 7.19.7 (4 November 2009)
572 Daniel Stenberg (2 Nov 2009)
573 - As reported independent by both Stan van de Burgt and Didier Brisebourg,
574 CURLINFO_SIZE_DOWNLOAD (the -w variable size_download) didn't work when
575 getting data from ldap!
577 Daniel Stenberg (31 Oct 2009)
578 - Gabriel Kuri reported a problem with CURLINFO_CONTENT_LENGTH_DOWNLOAD if the
579 download was 0 bytes, as libcurl would then return the size as unknown (-1)
580 and not 0. I wrote a fix and test case 566 to verify it.
582 Daniel Stenberg (30 Oct 2009)
583 - Liza Alenchery mentioned a problem with re-used SCP connection when a bad
584 auth is used, as it caused a crash. I failed to repeat the issue, but still
585 made a change that now forces the TCP connection used for a freed SCP
586 session to get closed and not be re-used.
588 - "Tom" posted a bug report that mentioned how libcurl did wrong when doing a
589 POST using a read callback, with Digest authentication and
590 "Transfer-Encoding: chunked" enforced. I would then cause the first request
591 to be wrongly sent and then basically hang until the server closed the
592 connection. I fixed the problem and added test case 565 to verify it.
594 Daniel Stenberg (25 Oct 2009)
595 - Dima Barsky made the curl cookie parser accept cookies even with blank or
596 unparsable expiry dates and then treat them as session cookies - previously
597 libcurl would reject cookies with a date format it couldn't parse. Research
598 shows that the major browser treat such cookies as session cookies. I
599 modified test 8 and 31 to verify this.
601 Daniel Stenberg (21 Oct 2009)
602 - Attempt to use pkg-config for finding out libssh2 installation details
605 - A patch in bug report #2883177 (http://curl.haxx.se/bug/view.cgi?id=2883177)
606 by Johan van Selst introduced the --crlfile option to curl, which makes curl
607 tell libcurl about a file with CRL (certificate revocation list) data to
610 Daniel Stenberg (18 Oct 2009)
611 - Ray Dassen provided a patch in Debian's bug tracker (bug number #551461)
612 that now makes curl_getdate(3) actually handles RFC 822 formatted dates that
613 use the "single letter military timezones".
614 http://www.rfc-ref.org/RFC-TEXTS/822/chapter5.html has the details.
616 - Fixed memory leak in the SCP/SFTP code as it never freed the knownhosts
619 - John Dennis filed bug report #2873666
620 (http://curl.haxx.se/bug/view.cgi?id=2873666) which identified a problem
621 which made libcurl loop infinitely when given incorrect credentials when
622 using HTTP GSS negotiate authentication. He also provided a small and simple
625 - Kevin Baughman found a double close() problem with libcurl-NSS, as when
626 libcurl called NSS to close the SSL "session" it also closed the actual
629 Yang Tse (17 Oct 2009)
630 - Bug report #2866724 indicated
631 (http://curl.haxx.se/bug/view.cgi?id=2866724) that curl on Windows failed
632 when writing files whose file names originally contained characters which
633 are not valid for file names on Windows. Dan Fandrich provided an initial
634 patch and another revised one to fix this issue.
636 Daniel Stenberg (1 Oct 2009)
637 - Tom Mueller correctly reported in bug report #2870221
638 (http://curl.haxx.se/bug/view.cgi?id=2870221) that libcurl returned an
639 incorrect return code from the internal trynextip() function which caused
640 him grief. This is a regression that was introduced in 7.19.1 and I find it
641 strange it hasn't hit us harder, but I won't persue into figuring out
644 - Constantine Sapuntzakis: The current implementation will always set
645 SO_SNDBUF to CURL_WRITE_SIZE even if the SO_SNDBUF starts out larger. The
646 patch doesn't do a setsockopt if SO_SNDBUF is already greater than
647 CURL_WRITE_SIZE. This should help folks who have set up their computer with
650 Daniel Stenberg (27 Sep 2009)
651 - I introduced a maximum limit for received HTTP headers. It is controlled by
652 the define CURL_MAX_HTTP_HEADER which is even exposed in the public header
653 file to allow for users to fairly easy rebuild libcurl with a modified
654 limit. The rationale for a fixed limit is that libcurl is realloc()ing a
655 buffer to be able to put a full header into it, so that it can call the
656 header callback with the entire header, but that also risk getting it into
657 trouble if a server by mistake or willingly sends a header that is more or
658 less without an end. The limit is set to 100K.
660 Daniel Stenberg (26 Sep 2009)
661 - John P. McCaskey posted a bug report that showed how libcurl did wrong when
662 saving received cookies with no given path, if the path in the request had a
663 query part. That is means a question mark (?) and characters on the right
664 side of that. I wrote test case 1105 and fixed this problem.
666 Kamil Dudka (26 Sep 2009)
667 - Implemented a protocol independent way to specify blocking direction, used by
668 transfer.c for blocking. It is currently used only by SCP and SFTP protocols.
669 This enhancement resolves an issue with 100% CPU usage during SFTP upload,
672 Daniel Stenberg (25 Sep 2009)
673 - Chris Mumford filed bug report #2861587
674 (http://curl.haxx.se/bug/view.cgi?id=2861587) identifying that libcurl used
675 the OpenSSL function X509_load_crl_file() wrongly and failed if it would
676 load a CRL file with more than one certificate within. This is now fixed.
678 Daniel Stenberg (16 Sep 2009)
679 - Sven Anders reported that we introduced a cert verfication flaw for OpenSSL-
680 powered libcurl in 7.19.6. If there was a X509v3 Subject Alternative Name
681 field in the certficate it had to match and so even if non-DNS and non-IP
682 entry was present it caused the verification to fail.
684 Daniel Fandrich (15 Sep 2009)
685 - Moved the libssh2 checks after the SSL library checks. This helps when
686 statically linking since libssh2 needs the SSL library link flags to be
687 set up already to satisfy its dependencies. This wouldn't be necessary if
688 the libssh2 configure check was changed to use pkg-config since the
689 --static flag would add the dependencies automatically.
691 Yang Tse (14 Sep 2009)
692 - Revert Joshua Kwan's patch committed 11 Sep 2009.
694 Some systems poll function sets POLLHUP in revents without setting
695 POLLIN, and sets POLLERR without setting POLLIN and POLLOUT. In some
696 libcurl code execution paths this could trigger busy wait loops with
697 high CPU usage until a timeout condition aborted the loop.
699 The reverted patch addressed the above issue for a very specific case,
700 when awaiting c-ares to resolve. A libcurl-wide fix for Curl_poll now
703 Guenter Knauf (11 Sep 2009)
704 - Joshua Kwan provided a patch to pass POLLERR / POLLHUP back to c-ares.
705 This fixes a loop problem with high CPU usage.
707 Daniel Stenberg (10 Sep 2009)
708 - Claes Jakobsson fixed a problem with cookie expiry dates at exctly the epoch
709 start second "Thu Jan 1 00:00:00 GMT 1970" as the date parser then returns 0
710 which internally then is treated as a session cookie. That particular date
711 is now made to get the value of 1.
713 Daniel Stenberg (2 Sep 2009)
714 - Daniel Johnson found a flaw in the code converting sftp-errors to libcurl
717 Daniel Stenberg (1 Sep 2009)
718 - Peter Sylvester made a debug feature for Curl_resolv() that now will force
719 libcurl to resolve 'localhost' whatever name you use in the URL *if* you set
720 the --interface option to (exactly) "LocalHost". This will enable us to
721 write tests for custom hosts names but still use a local host server.
723 - configure now tries to use pkg-config for a number of sub-dependencies even
724 when cross-compiling. The key to success is then you properly setup
725 PKG_CONFIG_PATH before invoking configure.
727 I also improved how NSS is detected by trying nss-config if pkg-config isn't
728 present, and as a last resort just use the lib name and force the user to
729 setup the LIBS/LDFLAGS/CFLAGS etc properly. The previous last resort would
730 add a range of various libs that would almost never be quite correct.
732 Daniel Stenberg (31 Aug 2009)
733 - When using the multi interface with FTP and you asked for NOBODY, you did no
734 QUOTE commands and the request used the same path as the connection had
735 already changed to, it would decide that no commands would be necessary for
736 the "DO" action and that was not handled properly but libcurl would instead
739 Kamil Dudka (28 Aug 2009)
740 - Improved error message for not matching certificate subject name in
741 libcurl-NSS. Originally reported at:
742 https://bugzilla.redhat.com/show_bug.cgi?id=516056#c9
744 Patrick Monnerat (24 Aug 2009)
745 - Introduced a SYST-based test to properly set-up name format when dealing
746 with the OS/400 FTP server.
748 - Fixed an ftp_readresp() bug preventing detection of failing control socket
749 and causing FTP client to loop forever.
751 Daniel Stenberg (24 Aug 2009)
752 - Marc de Bruin pointed out that configure --with-gnutls=PATH didn't work
753 properly and provided a fix. http://curl.haxx.se/bug/view.cgi?id=2843008
755 - Eric Wong introduced support for the new option -T. (dot) that makes curl
756 read stdin in a non-blocking fashion. This also brings back -T- (minus) to
757 the previous blocking behavior since it could break stuff for people at
760 Michal Marek (21 Aug 2009)
761 - With CURLOPT_PROXY_TRANSFER_MODE, avoid sending invalid URLs like
762 ftp://example.com;type=i if the user specified ftp://example.com without the
765 Daniel Stenberg (21 Aug 2009)
766 - Andre Guibert de Bruet pointed out a missing return code check for a
767 strdup() that could lead to segfault if it returned NULL. I extended his
768 suggest patch to now have Curl_retry_request() return a regular return code
769 and better check that.
771 - Lots of good work by Krister Johansen, mostly related to pipelining:
773 Fix SIGSEGV on free'd easy_conn when pipe unexpectedly breaks
774 Fix data corruption issue with re-connected transfers
775 Fix use after free if we're completed but easy_conn not NULL
777 Kamil Dudka (13 Aug 2009)
778 - Changed NSS code to not ignore the value of ssl.verifyhost and produce more
779 verbose error messages. Originally reported at:
780 https://bugzilla.redhat.com/show_bug.cgi?id=516056
782 Daniel Stenberg (12 Aug 2009)
783 - Karl Moerder fixed the Makefile.vc* makefiles to include the new file
784 nonblock.c so that they work fine again
786 - I expanded test 517 with a bunch of more dates that originate from the
787 Chrome browser test suite. It turns out most of them get parsed the same
790 Version 7.19.6 (12 August 2009)
792 Daniel Stenberg (12 Aug 2009)
793 - Carsten Lange reported a bug and provided a patch for TFTP upload and the
794 sending of the TSIZE option. I don't like fixing bugs just hours before
795 a release, but since it was broken and the patch fixes this for him I decided
798 Daniel Stenberg (11 Aug 2009)
799 - Peter Sylvester made the HTTPS test server use specific certificates for
800 each test, so that the test suite can now be used to actually test the
801 verification of cert names etc. This made an error show up in the OpenSSL-
802 specific code where it would attempt to match the CN field even if a
803 subjectAltName exists that doesn't match. This is now fixed and verified
806 - Benbuck Nason posted the bug report #2835196
807 (http://curl.haxx.se/bug/view.cgi?id=2835196), fixing a few compiler
808 warnings when mixing ints and bools.
810 Daniel Fandrich (10 Aug 2009)
811 - Fixed a memory leak in the FTP code and an off-by-one heap buffer overflow.
813 Daniel Fandrich (9 Aug 2009)
814 - Fixed some memory leaks in the command-line tool that caused most of the
815 torture tests to fail.
817 Daniel Stenberg (2 Aug 2009)
818 - Curt Bogmine reported a problem with SNI enabled on a particular server. We
819 should introduce an option to disable SNI, but as we're in feature freeze
820 now I've addressed the obvious bug here (pointed out by Peter Sylvester): we
821 shouldn't try to enable SNI when SSLv2 or SSLv3 is explicitly selected.
822 Code for OpenSSL and GnuTLS was fixed. NSS doesn't seem to have a particular
823 option for SNI, or are we simply not using it?
825 Daniel Stenberg (1 Aug 2009)
826 - Scott Cantor posted the bug report #2829955
827 (http://curl.haxx.se/bug/view.cgi?id=2829955) mentioning the recent SSL cert
828 verification flaw found and exploited by Moxie Marlinspike. The presentation
829 he did at Black Hat is available here:
830 https://www.blackhat.com/html/bh-usa-09/bh-usa-09-archives.html#Marlinspike
832 Apparently at least one CA allowed a subjectAltName or CN that contain a
833 zero byte, and thus clients that assumed they would never have zero bytes
834 were exploited to OK a certificate that didn't actually match the site. Like
835 if the name in the cert was "example.com\0theatualsite.com", libcurl would
836 happily verify that cert for example.com.
838 libcurl now better uses the length of the extracted name, not using the zero
839 termination for getting the string length.
841 This fixing only made and needed in OpenSSL interfacing code.
843 - Tanguy Fautre pointed out that OpenSSL's function RAND_screen() (present
844 only in some OpenSSL installs - like on Windows) isn't thread-safe and we
845 agreed that moving it to the global_init() function is a decent way to deal
848 - Alexander Beedie provided the patch for a noproxy problem: If I have set
849 CURLOPT_NOPROXY to "*", or to a host that should not use a proxy, I actually
850 could still end up using a proxy if a proxy environment variable was set.
852 Daniel Stenberg (27 Jul 2009)
853 - All the quote options (CURLOPT_QUOTE, CURLOPT_POSTQUOTE and
854 CURLOPT_PREQUOTE) now accept a preceeding asterisk before the command to
855 send when using FTP, as a sign that libcurl shall simply ignore the response
856 from the server instead of treating it as an error. Not treating a 400+ FTP
857 response code as an error means that failed commands will not abort the
858 chain of commands, nor will they cause the connection to get disconnected.
860 Daniel Stenberg (26 Jul 2009)
861 - Johan van Selst posted bug report #2825989
862 (http://curl.haxx.se/bug/view.cgi?id=2825989) pointing out that
863 OpenSSL-powered libcurl didn't support the SHA-2 digest algorithm, and
864 provided the solution too: to use OpenSSL_add_all_algorithms() in addition
865 to the older SSLeay_* alternative. OpenSSL_add_all_algorithms was added in
868 Daniel Stenberg (23 Jul 2009)
869 - Added CURLOPT_SSH_KNOWNHOSTS, CURLOPT_SSH_KEYFUNCTION, CURLOPT_SSH_KEYDATA.
870 They introduce known_host support for SSH keys to libcurl. See docs for
871 details. Note that this feature depends on a new enough libssh2 version, to
872 be supported in libssh2 1.2 and later (or current git repo at this time).
874 Michal Marek (22 Jul 2009)
875 - David Binderman found a memory and fd leak in lib/gtls.c:load_file()
876 (https://bugzilla.novell.com/523919). When looking at the code, I found that
877 also the ptr pointer can leak.
879 Kamil Dudka (20 Jul 2009)
880 - Claes Jakobsson improved the support for client certificates handling in
881 NSS-powered libcurl. Now the client certificates can be selected
882 automatically by a NSS built-in hook. Additionally pre-login to all PKCS11
883 slots is no more performed. It used to cause problems with HW tokens.
885 - Fixed reference counting for NSS client certificates. Now the PEM reader
886 module should be always properly unloaded on Curl_nss_cleanup(). If the
887 unload fails though, libcurl will try to reuse the already loaded instance.
889 Daniel Fandrich (15 Jul 2009)
890 - Added nonblock.c to the non-automake makefiles (note that the dependencies
891 in the Watcom makefiles aren't quite correct).
893 Michal Marek (15 Jul 2009)
894 - Changed the description of CURLINFO_OS_ERRNO to make it clear that the
895 errno is not reset on success.
897 Guenter Knauf (14 Jul 2009)
898 - renamed generated config.h to curl_config.h to avoid any future clashes
899 with config.h from other projects.
901 Daniel Stenberg (9 Jul 2009)
902 - Eric Wong introduced curlx_nonblock() that the curl tool now (re-)uses for
903 setting a file descriptor non-blocking. Used by the functionality Eric
904 himself brough on June 15th.
906 Daniel Stenberg (8 Jul 2009)
907 - Constantine Sapuntzakis posted bug report #2813123
908 (http://curl.haxx.se/bug/view.cgi?id=2813123) and an a patch that fixes the
911 Url A is accessed using auth. Url A redirects to Url B (on a different
912 server0. Url B reuses a persistent connection. Url B has auth, even though
913 it's on a different server.
915 Note: if Url B does not reuse a persistent connection, auth is not sent.
919 data->state.first_host is not initialized becuase Curl_http_connect is not
920 called when a connection is reused.
924 move initialization of data->state.first_host to Curl_http. No code before
925 Curl_http uses data->state.first_host anyway.
927 Guenter Knauf (4 Jul 2009)
928 - Markus Koetter provided a patch to avoid getnameinfo() usage which broke a
929 couple of both IPv4 and IPv6 autobuilds.
931 Daniel Stenberg (29 Jun 2009)
932 - Markus Koetter made CURLOPT_FTPPORT (and curl's -P/--ftpport) support a port
933 range if given colon-separated after the host name/address part. Like
934 "192.168.0.1:2000-10000"
936 - Modified the separators used for CURLOPT_CERTINFO in multi-part outputs. I
937 don't know how they got wrong in the first place, but using this output
938 format makes it possible to quite easily separate the string into an array
941 Daniel Fandrich (16 June 2009)
942 - Added a few more compiler warning options for gcc.
944 Daniel Stenberg (16 Jun 2009)
945 - Reuven Wachtfogel made curl -o - properly produce a binary output on windows
946 (no newline translations). Use -B/--use-ascii if you rather get the ascii
949 Michal Marek (16 Jun 2009)
950 - When doing non-anonymous ftp via http proxies and the password is not
951 provided in the url, add it there (squid needs this).
953 Daniel Stenberg (15 Jun 2009)
956 This allows curl(1) to be used as a client-side tunnel for arbitrary stream
957 protocols by abusing chunked transfer encoding in both the HTTP request and
958 HTTP response. This requires server support for sending a response while a
959 request is still being read, of course.
961 If attempting to read from stdin returns EAGAIN, then we pause our sender.
962 This leaves curl to attempt to read from the socket while reading from stdin
963 (and thus sending) is paused.
965 This change was needed to allow successfully tunneling the git protocol over
966 HTTP (--no-buffer is needed, as well).
968 Patrick Monnerat (15 Jun 2009)
969 - Replaced use of standard C library rand()/srand() by our own pseudo-random
972 Yang Tse (11 Jun 2009)
973 - I adapted testcurl script to allow building test harness programs when
974 cross-compiling for a *-*-mingw* host.
976 Daniel Stenberg (10 Jun 2009)
977 - Fabian Keil ran clang on the (lib)curl code, found a bunch of warnings and
978 contributed a range of patches to fix them.
980 Yang Tse (10 Jun 2009)
981 - I introduced configure script option --enable-curldebug which now allows
982 the decoupled enabling or disabling of the curl debug memory tracking
983 feature from the --enable-debug option which no longer controls this.
985 curl --version will list 'Debug' feature for debug enabled builds, and
986 will list 'TrackMemory' feature for curl debug memory tracking capable
987 builds. These features are independent and can be controlled when running
988 the configure script. When --enable-debug is given both features will be
989 enabled, unless some restriction prevents memory tracking from being used.
991 Internally, definition of preprocessor symbol DEBUGBUILD restricts code
992 which is only compiled for debug enabled builds. And symbol CURLDEBUG is
993 used to differentiate code which is _only_ used for memory tracking.
995 Yang Tse (9 Jun 2009)
996 - Daniel Steinberg pointed out that Curl_FormInit() in formdata.c was not
997 initializing the fread callback pointer and this triggered a compiler
998 warning, also provided a friendly suggestion on how to fix it.
1000 Daniel Stenberg (8 Jun 2009)
1001 - Claes Jakobsson provided a patch for libcurl-NSS that fixed a bad refcount
1002 issue with client certs that caused issues like segfaults.
1003 http://curl.haxx.se/mail/lib-2009-05/0316.html
1005 - Triggered by bug report #2798852 and the patch in there, I fixed configure
1006 to detect gnutls build options with pkg-config only and not libgnutls-config
1007 anymore since GnuTLS has stopped distributing that tool. If an explicit path
1008 is given to configure, we will instead guess on how to link and use that
1009 lib. I did not use the patch from the bug report.
1011 Yang Tse (8 Jun 2009)
1012 - Igor Novoseltsev adjusted Makefile.vxworks to get sources and headers
1013 included from Makefile.inc, and provided docs\INSTALL VxWorks section.
1015 - I removed buildconf.bat from release and daily snapshot archives. This
1016 file is only for CVS tree checkout builds.
1018 Daniel Stenberg (8 Jun 2009)
1019 - Eric Wong fixed --no-buffer to actually switch off output buffering. Been
1022 Bill Hoffman (6 Jun 2009)
1023 - Added some cmake docs and fixed socklen_t in the build.
1025 Yang Tse (5 Jun 2009)
1026 - John E. Malmberg provided VMS specific patch: "This fixes an existing bug
1027 in urlglob.c where it was not converting the Curl Unix exit code to a VMS
1028 DCL compatible exit code. This fix required the enhancement described next.
1029 This also adds an enhancement to main.c so that when curl is run under a
1030 Unix shell like Bash on VMS, it will return the standard Unix exit codes
1031 and messages." And another patch for docs/examples.
1033 I introduced os-specific.c and os-specific.h for use in curl tool code
1034 and adjusted John E. Malmberg's patch placement to use these new files
1035 as an effort to prevent main.c from growing ad infinitum. Code already
1036 existing in main.c which is OS specific should be moved into these files.
1038 Daniel Stenberg (4 June 2009)
1039 - Setting the Content-Length: header from your app when you do a POST or PUT
1040 is almost always a VERY BAD IDEA. Yet there are still apps out there doing
1041 this, and now recently it triggered a bug/side-effect in libcurl as when
1042 libcurl sends a POST or PUT with NTLM, it sends an empty post first when it
1043 knows it will just get a 401/407 back. If the app then replaced the
1044 Content-Length header, it caused the server to wait for input that libcurl
1045 wouldn't send. Aaron Oneal reported this problem in bug report #2799008
1046 (http://curl.haxx.se/bug/view.cgi?id=2799008) and helped us verify the fix.
1048 Yang Tse (4 Jun 2009)
1049 - Igor Novoseltsev provided patches and information, that after some
1050 adjustments to better fit curl's way of doing things, have resulted
1051 in the posibility of building libcurl for VxWorks.
1053 Daniel Fandrich (2 June 2009)
1054 - Checked in a Google Android make file. To use it, you must first
1055 create a config.h file by running configure in the Android environment,
1056 which doesn't seem to be easy to do. If no easy way can be found, a
1057 static config-android.h may need to be created and checked in to the
1058 libcurl source tree.
1060 Daniel Stenberg (1 June 2009)
1061 - Claes Jakobsson fixed the configure script to better find and use NSS
1064 Yang Tse (1 Jun 2009)
1065 - John E. Malmberg provided a VMS specific clean-up for curl.h, and pointed
1066 out that the configure script was failing to detect the timeval struct on
1067 VMS when building with _XOPEN_SOURCE_EXTENDED undefined due to definition
1068 taking place in socket.h instead of time.h. I have adjusted configure
1069 script to also include this header when checking struct timeval.
1071 Daniel Stenberg (27 May 2009)
1072 - Frank McGeough provided a small OpenSSL #include fix to make libcurl compile
1073 fine with Nokia 5th edition 1.0 SDK for Symbian.
1075 - Andre Guibert de Bruet found a call to a OpenSSL function that didn't check
1076 for a failure properly.
1078 - Mike Crowe pointed out that setting CURLOPT_USERPWD to NULL used to clear
1079 the auth credentials back in 7.19.0 and earlier while now you have to set ""
1080 to get the same effect. His patch brings back the ability to use NULL.
1082 - Claes Jakobsson fixed libcurl-NSS to build fine even without the
1083 PK11_CreateGenericObject() function.
1085 Daniel Stenberg (25 May 2009)
1086 - bug report #2796358 (http://curl.haxx.se/bug/view.cgi?id=2796358) pointed
1087 out that the cookie parser would leak memory when it parses cookies that are
1088 received with domain, path etc set multiple times in the same header. While
1089 such a cookie is questionable, they occur in the wild and libcurl no longer
1090 leaks memory for them. I added such a header to test case 8.
1092 Daniel Fandrich (22 May 2009)
1093 - Removed some obsolete digest code that caused a valgrind error in test 551.
1095 Daniel Fandrich (20 May 2009)
1096 - Added "non-existing host" test keywords to make it easy to skip those
1097 tests on machines that have broken DNS configurations (such as
1098 those configured to use OpenDNS).
1100 Daniel Stenberg (19 May 2009)
1101 - Kamil Dudka brought the patch from the Redhat bug entry
1102 https://bugzilla.redhat.com/show_bug.cgi?id=427966 which was libcurl closing
1103 a bad file descriptor when closing down the FTP data connection. Caolan
1104 McNamara seems to be the original author of it.
1106 Version 7.19.5 (18 May 2009)
1108 Daniel Stenberg (17 May 2009)
1109 - James Bursa posted a patch to the mailing list that fixed a problem with
1110 no_proxy which made it not skip the proxy if the URL entered contained a
1111 user name. I added test case 1101 to verify.
1113 Daniel Stenberg (11 May 2009)
1114 - Balint Szilakszi reported a memory leak when libcurl did gzip decompression
1115 of streams that had some parts (legitimately) missing. We now provide and use
1116 a proper cleanup function for the content encoding submodule.
1117 http://curl.haxx.se/mail/lib-2009-05/0092.html
1119 - Kamil Dudka provided a fix for libcurl-NSS reported by Michael Cronenworth
1120 at https://bugzilla.redhat.com/show_bug.cgi?id=453612#c12
1122 If an incorrect password is given while loading a private key, libcurl ends
1123 up in an infinite loop consuming memory. The bug is critical.
1125 - I fixed the problem with doing NTLM, POST and then following a 302 redirect,
1126 as reported by Ebenezer Ikonne (on curl-users) and Laurent Rabret (on
1127 curl-library). The transfer was mistakenly marked to get more data to send
1128 but since it didn't actually have that, it just hung there...
1130 Daniel Stenberg (10 May 2009)
1131 - Andre Guibert de Bruet correctly pointed out an over-alloc with one wasted
1132 byte in the digest code.
1134 Yang Tse (9 May 2009)
1135 - Removed DOS and TPF package's subdirectory Makefile.am, it was only used
1136 to include some files in the distribution tarball serving no other purpose.
1137 Files from the DOS and TPF subdirectories are now included in the EXTRA_DIST
1138 of the Makefile in the parent subdirectory.
1140 Yang Tse (8 May 2009)
1141 - Changed host name literal in several tests to one under the haxx.se domain.
1143 - Renamed vc6 workspace and project files to avoid filename clash when used
1144 for conversion to later VS versions.
1146 Daniel Stenberg (8 May 2009)
1147 - Constantine Sapuntzakis fixed bug report #2784055
1148 (http://curl.haxx.se/bug/view.cgi?id=2784055) identifying a problem to
1149 connect to SOCKS proxies when using the multi interface. It turned out to
1150 almost not work at all previously. We need to wait for the TCP connect to
1151 be properly verified before doing the SOCKS magic.
1153 There's still a flaw in the FTP code for this.
1155 Daniel Stenberg (7 May 2009)
1156 - Made the SO_SNDBUF setting for the data connection socket for ftp uploads as
1157 well. See change 28 Apr 2009.
1159 Yang Tse (7 May 2009)
1160 - Fixed an issue affecting FTP transfers, introduced with the transfer.c
1161 patch committed May 4.
1163 Daniel Stenberg (7 May 2009)
1164 - Man page *roff problems fixed thanks to input from Colin Watson. Problems
1165 reported in the Debian package.
1167 - Vijay G filed bug report #2723236
1168 (http://curl.haxx.se/bug/view.cgi?id=2723236) identifying a problem with
1169 libcurl's TFTP code and its lack of dealing with the OACK packet.
1171 Yang Tse (5 May 2009)
1172 - Fixed the --ftp-port address of test #251 to the CLIENTIP address, and
1173 reverted the change affecting test suite harness committed 4 May.
1175 Daniel Stenberg (5 May 2009)
1176 - Inspired by Michael Smith's session id fix for OpenSSL, I did the
1177 corresponding fix in the GnuTLS code: make sure to store the new session id
1178 in case the previous re-used one is rejected.
1180 Daniel Stenberg (4 May 2009)
1181 - Michael Smith posted bug report #2786255
1182 (http://curl.haxx.se/bug/view.cgi?id=2786255) with a patch, identifying how
1183 libcurl did not deal with SSL session ids properly if the server rejected a
1184 re-use of one. Starting now, it will forget the rejected one and remember
1185 the new. This change was for OpenSSL only, it is likely that other SSL lib
1186 code needs similar fixes.
1188 Yang Tse (4 May 2009)
1189 - Applied David McCreedy's "transfer.c fixes for CURL_DO_LINEEND_CONV and
1190 non-ASCII platform HTTP requests" patch addressing two HTTP PUT problems:
1191 1) On non-ASCII platforms not all of the protocol portions of the PUT are
1192 being translated to ASCII. 2) On all platforms the line endings of part of
1193 the protocol portions are mangled from CRLF to CRCRLF if data->set.crlf or
1194 data->set.prefer_ascii are set (depending on CURL_DO_LINEEND_CONV).
1196 - Applied David McCreedy's patch to fix test suite harness to allow test FTP
1197 server and client on different machines, providing FTP client address when
1198 running the FTP test server.
1200 Daniel Fandrich (3 May 2009)
1201 - Added and disabled test case 563 which shows KNOWN_BUGS #59. The bug
1202 report failed to mention that a proxy must be used to reproduce it.
1204 Yang Tse (2 May 2009)
1205 - Use a build-time configured curl_socklen_t data type instead of socklen_t.
1207 Yang Tse (1 May 2009)
1208 - Applied David McCreedy's patches "TPF-platform specific changes to various
1209 files" and "http.c fix to Curl_proxyCONNECT for non-ASCII platforms", the
1210 former with minor edits.
1212 Daniel Stenberg (30 Apr 2009)
1213 - I was going to fix issue #59 in KNOWN_BUGS
1215 If the CURLOPT_PORT option is used on an FTP URL like
1216 "ftp://example.com/file;type=A" the ";type=A" is stripped off.
1218 I added test case 562 to verify, only to find out that I couldn't repeat
1219 this bug so I hereby consider it not a bug anymore!
1221 Daniel Stenberg (29 Apr 2009)
1222 - Based on bug report #2723219 (http://curl.haxx.se/bug/view.cgi?id=2723219)
1223 I've now made TFTP "connections" not being kept for re-use within libcurl.
1224 TFTP is UDP-based so the benefit was really low (if even existing) to begin
1225 with so instead of tracking down to fix this problem we instead removed the
1226 re-use. I also enabled test case 1099 that I wrote a few days ago to verify
1227 that this change fixes the reported problem.
1229 Daniel Stenberg (28 Apr 2009)
1230 - Constantine Sapuntzakis filed bug report #2783090
1231 (http://curl.haxx.se/bug/view.cgi?id=2783090) pointing out that on windows
1232 we need to grow the SO_SNDBUF buffer somewhat to get really good upload
1233 speeds. http://support.microsoft.com/kb/823764 has the details. Friends
1234 confirmed that simply adding 32 to CURL_MAX_WRITE_SIZE is enough.
1236 - Bug report #2709004 (http://curl.haxx.se/bug/view.cgi?id=2709004) by Tim
1237 Chen pointed out how curl couldn't upload with resume when reading from a
1240 This ended up with the introduction of a new return code for the
1241 CURLOPT_SEEKFUNCTION callback that basically says that the seek failed but
1242 that libcurl may try to resolve the situation anyway. In our case this means
1243 libcurl will attempt to instead read that much data from the stream instead
1244 of seeking and that way curl can now upload with resume when data is read
1247 Daniel Stenberg (26 Apr 2009)
1248 - Bug report #2779733 (http://curl.haxx.se/bug/view.cgi?id=2779733) by Sven
1249 Wegener pointed out that CURLINFO_APPCONNECT_TIME didn't work with the multi
1250 interface and provided a patch that fixed the problem!
1252 Daniel Stenberg (24 Apr 2009)
1253 - Kamil Dudka fixed another NSS-related leak when client certs were used.
1255 - Bug report #2779245 (http://curl.haxx.se/bug/view.cgi?id=2779245) by Rainer
1256 Koenig pointed out that the man page didn't tell that the *_proxy
1257 environment variables can be specified lower case or UPPER CASE and the
1258 lower case takes precedence,
1260 Daniel Fandrich (21 Apr 2009)
1261 - Added new libcurl source files to Amiga, RiscOS and VC6 build files.
1263 Yang Tse (21 Apr 2009)
1264 - Moved potential inclusion of system's malloc.h and memory.h header files to
1265 setup_once.h. Inclusion of each header file is based on the definition of
1266 NEED_MALLOC_H and NEED_MEMORY_H respectively.
1268 Renamed libcurl's memory.h to curl_memory.h
1270 Daniel Stenberg (20 Apr 2009)
1271 - Leanic Lefever reported a crash and did some detailed research on why and
1272 how it occurs (http://curl.haxx.se/mail/lib-2009-04/0289.html). The
1273 conclusion was that if an error is detected and Curl_done() is called for
1274 the connection, ftp_done() could at times return another error code that
1275 then would take precedence and that new code confused existing logic that
1276 works for the first error code (CURLE_SEND_ERROR) only.
1278 - Gisle Vanem noticed that --libtool would produce bogus strings at times for
1279 OBJECTPOINT options. Now we've introduced a new function - my_setopt_str -
1280 within the app for setting plain string options to avoid the risk of this
1283 Daniel Stenberg (17 Apr 2009)
1284 - Pramod Sharma reported and tracked down a bug when doing FTP over a HTTP
1285 proxy. libcurl would then wrongly close the connection after each
1286 request. In his case it had the weird side-effect that it killed NTLM auth
1287 for the proxy causing an inifinite loop!
1289 I added test case 1098 to verify this fix. The test case does however not
1290 properly verify that the transfers are done persistently - as I couldn't
1291 think of a clever way to achieve it right now - but you need to read the
1292 stderr output after a test run to see that it truly did the right thing.
1294 Daniel Stenberg (13 Apr 2009)
1295 - bug report #2727981 (http://curl.haxx.se/bug/view.cgi?id=2727981) by Martin
1296 Storsjö pointed out how setting CURLOPT_NOBODY to 0 could be downright
1297 confusing as it set the method to either GET or HEAD. The example he showed
1300 curl_easy_setopt(curl, CURLOPT_PUT, 1);
1301 curl_easy_setopt(curl, CURLOPT_NOBODY, 0);
1303 The new way doesn't alter the method until the request is about to start. If
1304 CURLOPT_NOBODY is then 1 the HTTP request will be HEAD. If CURLOPT_NOBODY is
1305 0 and the request happens to have been set to HEAD, it will then instead be
1306 set to GET. I believe this will be less surprising to users, and hopefully
1307 not hit any existing users badly.
1309 - Toshio Kuratomi reported a memory leak problem with libcurl+NSS that turned
1310 out to be leaking cacerts. Kamil Dudka helped me complete the fix. The issue
1311 is found in Redhat's bug tracker:
1312 https://bugzilla.redhat.com/show_bug.cgi?id=453612
1314 There are still memory leaks present, but they seem to have other reasons.
1316 Daniel Fandrich (11 Apr 2009)
1317 - Added new libcurl source files to Symbian OS build files.
1318 - Improved Symbian support for SSL.
1320 Yang Tse (10 Apr 2009)
1321 - Daniel Johnson improved the MacOSX-Framework shell script to now perform all
1322 the steps required to build a Mac OS X four way fat ppc/i386/ppc64/x86_64
1323 libcurl.framework. Four way fat framework requires OS X 10.5 SDK or later.
1325 Yang Tse (8 Apr 2009)
1326 - Removed Sun compilers preprocessor block from curlbuild.h.dist, this also
1327 removes it from the curlbuild.h file originally distributed by the cURL
1328 project as this file is intended for systems not capable of running the
1329 configure script. For those who have been building curl out of the source
1330 code curl distribution tarball provided by curl.haxx.se the change implies
1331 nothing. Previous change in this area committed 2 Apr becomes irrelevant.
1333 Daniel Stenberg (6 Apr 2009)
1334 - I clarified in the docs that CURLOPT_SEEKFUNCTION should return 0 on success
1335 and 1 on fatal errors. Previously it only mentioned non-zero on fatal
1336 errors. This is a slight change in meaning, but it follows what we've done
1337 elsewhere before and it opens up for LOTS of more useful return codes
1338 whenever we can think of them...
1340 Yang Tse (2 Apr 2009)
1341 - Fix curl_off_t definition for builds done using Sun compilers and a
1342 non-configured libcurl. In this case curl_off_t data type was gated
1343 to the off_t data type which depends on the _FILE_OFFSET_BITS. This
1344 configuration is exactly the unwanted configuration for our curl_off_t
1345 data type which must not depend on such setting. This breaks ABI for
1346 libcurl libraries built with Sun compilers which were built without
1347 having run the configure script with _FILE_OFFSET_BITS different than
1348 64 and using the ILP32 data model.
1350 Daniel Stenberg (1 Apr 2009)
1351 - Andre Guibert de Bruet fixed a NULL pointer use in an infof() call if a
1352 strdup() call failed.
1354 Daniel Fandrich (31 Mar 2009)
1355 - Properly return an error code in curl_easy_recv (reported by Jim Freeman).
1357 Daniel Stenberg (18 Mar 2009)
1358 - Kamil Dudka brought a patch that enables 6 additional crypto algorithms when
1359 NSS is used. These ciphers were added in NSS 3.4 and require to be enabled
1362 Daniel Stenberg (13 Mar 2009)
1363 - Use libssh2_version() to present the libssh2 version in case the libssh2
1364 library is found to support it.
1366 Yang Tse (12 Mar 2009)
1367 - Added missing Curl_read() return code checking in TELNET transfers.
1369 - Pierre Brico found and fixed TELNET transfers not being aborted upon
1370 a write callback failure.
1372 Daniel Stenberg (11 Mar 2009)
1373 - Kamil Dudka made the curl tool properly call curl_global_init() before any
1374 other libcurl function.
1376 Yang Tse (11 Mar 2009)
1377 - Added missing TELNET timeout support for Windows builds. This issue was
1378 reported by Pierre Brico.
1380 Daniel Stenberg (9 Mar 2009)
1381 - Frank Hempel found out a bug and provided the fix:
1383 curl_easy_duphandle did not necessarily duplicate the CURLOPT_COOKIEFILE
1384 option. It only enabled the cookie engine in the destination handle if
1385 data->cookies is not NULL (where data is the source handle). In case of a
1386 newly initialized handle which just had the cookie support enabled by a
1387 curl_easy_setopt(handle, CURL_COOKIEFILE, "")-call, handle->cookies was
1388 still NULL because the setopt-call only appends the value to
1389 data->change.cookielist, hence duplicating this handle would not have the
1390 cookie engine switched on.
1392 We also concluded that the slist-functionality would be suitable for being
1393 put in its own module rather than simply hanging out in lib/sendf.c so I
1394 created lib/slist.[ch] for them.
1396 - Andreas Farber made the 'buildconf' script check for the presence of m4
1397 scripts to make it detect a bad checkout earlier. People with older
1398 checkouts who don't do cvs update with the -d option won't get the new dirs
1399 and then will get funny outputs that can be a bit hard to understand and
1402 Daniel Stenberg (8 Mar 2009)
1403 - Andre Guibert de Bruet found and fixed a code segment in ssluse.c where the
1404 allocation of the memory BIO was not being properly checked.
1406 - Andre Guibert de Bruet fixed the gnutls-using code: There are a few places
1407 in the gnutls code where we were checking for negative values for errors,
1408 when the man pages state that GNUTLS_E_SUCCESS is returned on success and
1409 other values indicate error conditions.
1411 - Bill Egert pointed out (http://curl.haxx.se/bug/view.cgi?id=2671602) that
1412 curl didn't use sprintf() in a way that is documented to work in POSIX but
1413 since we use our own printf() code (from libcurl) that shouldn't be a
1414 problem. Nonetheless I modified the code to not rely on such particular
1415 features and to not cause further raised eyebrowse with no good reason.
1417 Daniel Fandrich (5 Mar 2009)
1418 - Expanded the security section of the libcurl-tutorial man page to cover
1419 more issues for authors to consider when writing robust libcurl-using
1422 Yang Tse (5 Mar 2009)
1423 - Fixed NTLM authentication memory leak on SSPI enabled Windows builds. This
1424 issue was noticed by Chris Deidun.
1426 Daniel Fandrich (4 Mar 2009)
1427 - Fixed a problem with m4 quoting in the OpenSSL configure check reported
1430 Daniel Stenberg (3 Mar 2009)
1431 - David James brought a patch that make libcurl close (all) dead connections
1432 whenever you attempt to open a new connection.
1434 1. After cleaning up a dead connection, "continue" instead of
1435 returning FALSE. This ensures that we clean up all dead connections,
1436 rather than just cleaning up the first dead connection.
1437 2. Move up the cleanup for dead connections so that it occurs for
1438 all connections, rather than just the connections which have the same
1439 preferences as our current new connection.
1441 Version 7.19.4 (3 March 2009)
1443 Daniel Stenberg (3 Mar 2009)
1444 - David Kierznowski notified us about a security flaw
1445 (http://curl.haxx.se/docs/adv_20090303.html also known as CVE-2009-0037) in
1446 which previous libcurl versions (by design) can be tricked to access an
1447 arbitrary local/different file instead of a remote one when
1448 CURLOPT_FOLLOWLOCATION is enabled. This flaw is now fixed in this release
1449 together this the addition of two new setopt options for controlling this
1452 o CURLOPT_REDIR_PROTOCOLS controls what protocols libcurl is allowed to
1453 follow to when CURLOPT_FOLLOWLOCATION is enabled. By default, this option
1454 excludes the FILE and SCP protocols and thus you nee to explicitly allow
1455 them in your app if you really want that behavior.
1457 o CURLOPT_PROTOCOLS controls what protocol(s) libcurl is allowed to fetch
1458 using the primary URL option. This is useful if you want to allow a user or
1459 other outsiders control what URL to pass to libcurl and yet not allow all
1460 protocols libcurl may have been built to support.
1462 Daniel Stenberg (27 Feb 2009)
1463 - Senthil Raja Velu reported a problem when CURLOPT_INTERFACE and
1464 CURLOPT_LOCALPORT were used together (the local port bind failed), and
1465 Markus Koetter provided the fix!
1467 Daniel Stenberg (25 Feb 2009)
1468 - As Daniel Fandrich figured out, we must do the GnuTLS initing in the
1469 curl_global_init() function to properly maintain the performing functions
1470 thread-safe. We've previously (28 April 2007) moved the init to a later time
1471 just to avoid it to fail very early when libgcrypt dislikes the situation,
1472 but that move was bad and the fix should rather be in libgcrypt or
1475 Daniel Stenberg (24 Feb 2009)
1476 - Brian J. Murrell found out that Negotiate proxy authentication didn't work.
1477 It happened because the code used the struct for server-based auth all the
1478 time for both proxy and server auth which of course was wrong.
1480 Daniel Stenberg (23 Feb 2009)
1481 - After a bug reported by James Cheng I've made curl_easy_getinfo() for
1482 CURLINFO_CONTENT_LENGTH_DOWNLOAD and CURLINFO_CONTENT_LENGTH_UPLOAD return
1483 -1 if the sizes aren't know. Previously these returned 0, make it impossible
1484 to detect the difference between actually zero and unknown.
1486 Yang Tse (23 Feb 2009)
1487 - Daniel Johnson provided a shell script that will perform all the steps needed
1488 to build a Mac OS X fat ppc/i386 or ppc64/x86_64 libcurl.framework
1490 Daniel Stenberg (23 Feb 2009)
1491 - I renamed everything in the windows builds files that used the name 'curllib'
1492 to the proper 'libcurl' as clearly this caused confusion.
1494 Yang Tse (20 Feb 2009)
1495 - Do not halt compilation when using VS2008 to build a Windows 2000 target.
1497 Daniel Stenberg (20 Feb 2009)
1498 - Linus Nielsen Feltzing reported and helped me repeat and fix a problem with
1499 FTP with the multi interface: when a transfer fails, like when aborted by a
1500 write callback, the control connection was wrongly closed and thus not
1503 This change is also an attempt to cleanup the code somewhat in this area, as
1504 now the FTP code attempts to keep (better) track on pending responses
1505 necessary to get read in ftp_done().
1507 Daniel Stenberg (19 Feb 2009)
1508 - Patrik Thunstrom reported a problem and helped me repeat it. It turned out
1509 libcurl did a superfluous 1000ms wait when doing SFTP downloads!
1511 We read data with libssh2 while doing the "DO" operation for SFTP and then
1512 when we were about to start getting data for the actual file part, the
1513 "TRANSFER" part, we waited for socket action (in 1000ms) before doing a
1514 libssh2-read. But in this case libssh2 had already read and buffered the
1515 data so we ended up always just waiting 1000ms before we get working on the
1518 Patrick Monnerat (18 Feb 2009)
1519 - FTP downloads (i.e.: RETR) ending with code 550 now return error
1520 CURLE_REMOTE_FILE_NOT_FOUND instead of CURLE_FTP_COULDNT_RETR_FILE.
1522 Daniel Stenberg (17 Feb 2009)
1523 - Kamil Dudka made NSS-powered builds compile and run again!
1525 - A second follow-up change by Andre Guibert de Bruet to fix a related memory
1526 leak like that fixed on the 14th. When zlib returns failure, we need to
1527 cleanup properly before returning error.
1529 - CURLOPT_FTP_CREATE_MISSING_DIRS can now be set to 2 in addition to 1 for
1530 plain FTP connections, and it will then allow MKD to fail once and retry the
1531 CWD afterwards. This is especially useful if you're doing many simultanoes
1532 connections against the same server and they all have this option enabled,
1533 as then CWD may first fail but then another connection does MKD before this
1534 connection and thus MKD fails but trying CWD works! The numbers can
1535 (should?) now be set with the convenience enums now called
1536 CURLFTP_CREATE_DIR and CURLFTP_CREATE_DIR_RETRY.
1538 Tests has proven that if you're making an application that uploads a set of
1539 files to an ftp server, you will get a noticable gain in speed if you're
1540 using multiple connections and this option will be then be very useful.
1542 Daniel Stenberg (14 Feb 2009)
1543 - Andre Guibert de Bruet found and fixed a memory leak in the content encoding
1544 code, which could happen on libz errors.
1546 Daniel Fandrich (12 Feb 2009)
1547 - Added support for Digest and NTLM authentication using GnuTLS.
1549 Daniel Stenberg (11 Feb 2009)
1550 - CURLINFO_CONDITION_UNMET was added to allow an application to get to know if
1551 the condition in the previous request was unmet. This is typically a time
1552 condition set with CURLOPT_TIMECONDITION and was previously not possible to
1553 reliably figure out. From bug report #2565128
1554 (http://curl.haxx.se/bug/view.cgi?id=2565128) filed by Jocelyn Jaubert.
1556 Daniel Fandrich (4 Feb 2009)
1557 - Don't add the standard /usr/lib or /usr/include paths to LDFLAGS and CPPFLAGS
1558 (respectively) when --with-ssl=/usr is used (patch based on FreeBSD).
1560 - Added an explicit buffer limit check in msdosify() (patch based on FreeBSD).
1561 This couldn't ever overflow in curl, but might if the code were used
1562 elsewhere or under different conditions.
1564 Daniel Stenberg (3 Feb 2009)
1565 - Hidemoto Nakada provided a small fix that makes it possible to get the
1566 CURLINFO_CONTENT_LENGTH_DOWNLOAD size from file:// "transfers" with
1567 CURLOPT_NOBODY set true.
1569 Daniel Stenberg (2 Feb 2009)
1570 - Patrick Scott found a rather large memory leak when using the multi
1571 interface and setting CURLMOPT_MAXCONNECTS to something less than the number
1572 of handles you add to the multi handle. All the connections that didn't fit
1573 in the cache would not be properly disconnected nor freed!
1575 - Craig A West brought us: libcurl now defaults to do CONNECT with HTTP
1576 version 1.1 instead of 1.0 like before. This change also introduces the new
1577 proxy type for libcurl called 'CURLPROXY_HTTP_1_0' that then allows apps to
1578 switch (back) to CONNECT 1.0 requests. The curl tool also got a --proxy1.0
1579 option that works exactly like --proxy but sets CURLPROXY_HTTP_1_0.
1581 I updated all test cases cases that use CONNECT and I tried to do some using
1582 --proxy1.0 and some updated to do CONNECT 1.1 to get both versions run.
1584 Daniel Stenberg (31 Jan 2009)
1585 - When building with c-ares 1.6.1 (not yet released) or later and IPv6 support
1586 enabled, we can now take advantage of its brand new AF_UNSPEC support in
1587 ares_gethostbyname(). This makes test case 241 finally run fine for me with
1588 this setup since it now parses the "::1 ip6-localhost" line fine in my
1591 Daniel Stenberg (30 Jan 2009)
1592 - Scott Cantor filed bug report #2550061
1593 (http://curl.haxx.se/bug/view.cgi?id=2550061) mentioning that I failed to
1594 properly make sure that the VC9 makefiles got included in the latest
1595 release. I've now fixed the release script and verified it so next release
1596 will hopefully include them properly!
1598 Daniel Fandrich (30 Jan 2009)
1599 - Fixed --disable-proxy for FTP and SOCKS. Thanks to Daniel Egger for
1602 Yang Tse (29 Jan 2009)
1603 - Introduced curl_sspi.c and curl_sspi.h for the implementation of functions
1604 Curl_sspi_global_init() and Curl_sspi_global_cleanup() which previously were
1605 named Curl_ntlm_global_init() and Curl_ntlm_global_cleanup() in http_ntlm.c
1606 Also adjusted socks_sspi.c to remove the link-time dependency on the Windows
1607 SSPI library using it now in the same way as it was done in http_ntlm.c.
1609 Daniel Stenberg (28 Jan 2009)
1610 - Markus Moeller introduced two new options to libcurl:
1611 CURLOPT_SOCKS5_GSSAPI_SERVICE and CURLOPT_SOCKS5_GSSAPI_NEC to allow libcurl
1612 to do GSS-style authentication with SOCKS5 proxies. The curl tool got the
1613 options called --socks5-gssapi-service and --socks5-gssapi-nec to enable
1616 Daniel Stenberg (26 Jan 2009)
1617 - Chad Monroe provided the new CURLOPT_TFTP_BLKSIZE option that allows an app
1618 to set desired block size to use for TFTP transfers instead of the default
1621 - The "-no_ticket" option was introduced in Openssl0.9.8j. It's a flag to
1622 disable "rfc4507bis session ticket support". rfc4507bis was later turned
1623 into the proper RFC5077 it seems: http://tools.ietf.org/html/rfc5077
1625 The enabled extension concerns the session management. I wonder how often
1626 libcurl stops a connection and then resumes a TLS session. also, sending the
1627 session data is some overhead. .I suggest that you just use your proposed
1628 patch (which explicitly disables TICKET).
1630 If someone writes an application with libcurl and openssl who wants to
1631 enable the feature, one can do this in the SSL callback.
1633 Sharad Gupta brought this to my attention. Peter Sylvester helped me decide
1634 on the proper action.
1636 - Alexey Borzov filed bug report #2535504
1637 (http://curl.haxx.se/bug/view.cgi?id=2535504) pointing out that realms with
1638 quoted quotation marks in HTTP Digest headers didn't work. I've now added
1639 test case 1095 that verifies my fix.
1641 - Craig A West brought CURLOPT_NOPROXY and the corresponding --noproxy option.
1642 They basically offer the same thing the NO_PROXY environment variable only
1643 offered previously: list a set of host names that shall not use the proxy
1644 even if one is specified.
1646 Daniel Fandrich (20 Jan 2009)
1647 - Call setlocale() for libtest tests to test the effects of locale-induced
1648 libc changes on libcurl.
1650 - Fixed a couple more locale-dependent toupper conversions, mainly for
1651 clarity. This does fix one problem that causes ;type=i FTP URLs
1652 to fail in the Turkish locale when CURLOPT_PROXY_TRANSFER_MODE is
1653 used (test case 561)
1655 - Added tests 561 and 1091 through 1094 to test various combinations
1656 of ;type= and ;mode= URLs that could potentially fail in the Turkish
1659 Daniel Stenberg (20 Jan 2009)
1660 - Lisa Xu pointed out that the ssh.obj file was missing from the
1661 lib/Makefile.vc6 file (and thus from the vc8 and vc9 ones too).
1663 Version 7.19.3 (19 January 2009)
1665 Daniel Stenberg (16 Jan 2009)
1666 - Andrew de los Reyes fixed curlbuild.h for "generic" gcc builds on PPC, both
1669 Daniel Stenberg (15 Jan 2009)
1670 - Tim Ansell fixed a compiler warning in lib/cookie.c
1672 Daniel Stenberg (14 Jan 2009)
1673 - Grant Erickson fixed timeouts for TFTP such that specifying a
1674 connect-timeout, a max-time or both options work correctly and as expected
1675 by passing the correct boolean value to Curl_timeleft via the
1676 'duringconnect' parameter.
1678 With this small change, curl TFTP now behaves as expected (and likely as
1679 originally-designed):
1681 1) For non-existent or unreachable dotted IP addresses:
1683 a) With no options, follows the default curl 300s timeout...
1684 b) With --connect-timeout only, follows that value...
1685 c) With --max-time only, follows that value...
1686 d) With both --connect-timeout and --max-time, follows the smaller value...
1688 and times out with a "curl: (7) Couldn't connect to server" error.
1690 2) For transfers to/from a valid host:
1692 a) With no options, follows default curl 300s timeout for the
1693 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1694 timeout for the remainder of the transfer...
1696 b) With --connect-time only, follows that value for the
1697 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1698 timeout for the remainder of the transfer...
1700 c) With --max-time only, follows that value for the first
1701 XRQ/DATA/ACK transaction and for the remainder of the
1704 d) With both --connect-timeout and --max-time, follows the former
1705 for the first XRQ/DATA/ACK transaction and the latter for the
1706 remainder of the transfer...
1708 and times out with a "curl: (28) Timeout was reached" error as
1711 Daniel Stenberg (13 Jan 2009)
1712 - Michael Wallner fixed a NULL pointer deref when calling
1713 curl_easy_setup(curl, CURLOPT_COOKIELIST, "SESS") on a CURL handle with no
1716 - Stefan Teleman brought a patch to fix the default curlbuild.h file for the
1719 Daniel Stenberg (12 Jan 2009)
1720 - Based on bug report #2498665 (http://curl.haxx.se/bug/view.cgi?id=2498665)
1721 by Daniel Black, I've now added magic to the configure script that makes it
1722 use pkg-config to detect gnutls details as well if the existing method
1723 (using libgnutls-config) fails. While doing this, I cleaned up and unified
1724 the pkg-config usage when detecting openssl and nss as well.
1726 Daniel Stenberg (11 Jan 2009)
1727 - Karl Moerder brought the patch that creates vc9 Makefiles, and I made
1728 'maketgz' now use the actual makefile targets to do the VC8 and VC9
1731 Daniel Stenberg (10 Jan 2009)
1732 - Emil Romanus fixed:
1734 When using the multi interface over HTTP and the server returns a Location
1735 header, the running easy handle will get stuck in the CURLM_STATE_PERFORM
1736 state, leaving the external event loop stuck waiting for data from the
1737 ingoing socket (when using the curl_multi_socket_action stuff). While this
1738 bug was pretty hard to find, it seems to require only a one-line fix. The
1739 break statement on line 1374 in multi.c caused the function to skip the call
1742 How to reproduce this bug? Well, that's another question. evhiperfifo.c in
1743 the examples directory chokes on this bug only _sometimes_, probably
1744 depending on how fast the URLs are added. One way of testing the bug out is
1745 writing to hiper.fifo from more than one source at the same time.
1747 Daniel Fandrich (7 Jan 2009)
1748 - Unified much of the SessionHandle initialization done in Curl_open() and
1749 curl_easy_reset() by creating Curl_init_userdefined(). This had the side
1750 effect of fixing curl_easy_reset() so it now also resets
1751 CURLOPT_FTP_FILEMETHOD and CURLOPT_SSL_SESSIONID_CACHE
1753 Daniel Stenberg (7 Jan 2009)
1754 - Rob Crittenden did once again provide an NSS update:
1756 I have to jump through a few hoops now with the NSS library initialization
1757 since another part of an application may have already initialized NSS by the
1758 time Curl gets invoked. This patch is more careful to only shutdown the NSS
1759 library if Curl did the initialization.
1761 It also adds in a bit of code to set the default ciphers if the app that
1762 call NSS_Init* did not call NSS_SetDomesticPolicy() or set specific
1763 ciphers. One might argue that this lets other application developers get
1764 lazy and/or they aren't using the NSS API correctly, and you'd be right.
1765 But still, this will avoid terribly difficult-to-trace crashes and is
1768 Daniel Stenberg (1 Jan 2009)
1769 - 'reconf' is removed since we rather have users use 'buildconf'
1771 Daniel Stenberg (31 Dec 2008)
1772 - Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
1773 out that 'reconf' didn't properly point out the m4 subdirectory when running
1776 Daniel Stenberg (29 Dec 2008)
1777 - Phil Lisiecki filed bug report #2413067
1778 (http://curl.haxx.se/bug/view.cgi?id=2413067) that identified a problem that
1779 would cause libcurl to mark a DNS cache entry "in use" eternally if the
1780 subsequence TCP connect failed. It would thus never get pruned and refreshed
1781 as it should've been.
1783 Phil provided his own patch to this problem that while it seemed to work
1784 wasn't complete and thus I wrote my own fix to the problem.
1786 Daniel Stenberg (28 Dec 2008)
1787 - Peter Korsgaard fixed building libcurl with "configure --with-ssl
1790 - Anthony Bryan fixed more language and spelling flaws in man pages.
1792 Daniel Stenberg (22 Dec 2008)
1793 - Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
1794 on file indexes beyond 2 or 4GB.
1796 - Anthony Bryan provided a set of patches that cleaned up manual language,
1797 corrected spellings and more.
1799 Daniel Stenberg (20 Dec 2008)
1800 - Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
1801 pipelining, as libcurl could then easily get confused and A) work on the
1802 handle that was not "first in queue" on a pipeline, or even B) tell the app
1803 to REMOVE a socket while it was in use by a second handle in a pipeline. Both
1804 errors caused hanging or stalling applications.
1806 Daniel Stenberg (19 Dec 2008)
1807 - curl_multi_timeout() could return a timeout value of 0 even though nothing
1808 was actually ready to get done, as the internal time resolution is higher
1809 than the returned millisecond timer. Therefore it could cause applications
1810 running on fast processors to do short bursts of busy-loops.
1811 curl_multi_timeout() will now only return 0 if the timeout is actually
1814 - Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
1815 now has an improved ability to do right when the multi interface (both
1816 "regular" and multi_socket) is used for SCP and SFTP transfers. This should
1817 result in (much) less busy-loop situations and thus less CPU usage with no
1820 Daniel Stenberg (17 Dec 2008)
1821 - SCP and SFTP with the multi interface had the same flaw: the 'DONE'
1822 operation didn't complete properly if the EAGAIN equivalent was returned but
1823 libcurl would simply continue with a half-completed close operation
1824 performed. This ruined persistent connection re-use and cause some
1825 SSH-protocol errors in general. The correction is unfortunately adding a
1826 blocking function - doing it entirely non-blocking should be considered for
1829 Gisle Vanem (16 Dec 2008)
1830 - Added the possibility to use the Watt-32 tcp/ip stack under Windows.
1831 The change simply involved adding a USE_WATT32 section in the
1832 config-win32.h files (under ./lib and ./src). This section disables
1833 the use of any Winsock headers.
1835 Daniel Stenberg (16 Dec 2008)
1836 - libssh2_sftp_last_error() was wrongly used at some places in libcurl which
1837 made libcurl sometimes not properly abort problematic SFTP transfers.
1839 Daniel Stenberg (12 Dec 2008)
1840 - More work with Igor Novoseltsev to first fix the remaining stuff for
1841 removing easy handles from multi handles when the easy handle is/was within
1842 a HTTP pipeline. His bug report #2351653
1843 (http://curl.haxx.se/bug/view.cgi?id=2351653) was also related and was
1844 eventually fixed by a patch by Igor himself.
1846 Yang Tse (12 Dec 2008)
1847 - Patrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
1848 OS/400 compilations with IPv6 enabled.
1850 Daniel Stenberg (12 Dec 2008)
1851 - Mark Karpeles filed bug report #2416182 titled "crash in ConnectionExists
1852 when using duphandle+curl_mutli"
1853 (http://curl.haxx.se/bug/view.cgi?id=2416182) which showed that
1854 curl_easy_duphandle() wrongly also copied the pointer to the connection
1855 cache, which was plain wrong and caused a segfault if the handle would be
1856 used in a different multi handle than the handle it was duplicated from.
1858 Daniel Stenberg (11 Dec 2008)
1859 - Keshav Krity found out that libcurl failed to deal with dotted IPv6
1860 addresses if they were very long (>39 letters) due to a too strict address
1861 validity parser. It now accepts addresses up to 45 bytes long.
1863 Daniel Stenberg (11 Dec 2008)
1864 - Internet Explorer had a broken HTTP digest authentication before v7 and
1865 there are servers "out there" that relies on the client doing this broken
1866 Digest authentication. Apache even comes with an option to work with such
1869 The difference is only for URLs that contain a query-part (a '?'-letter and
1870 text to the right of it).
1872 libcurl now supports this quirk, and you enable it by setting the
1873 CURLAUTH_DIGEST_IE bit in the bitmask you pass to the CURLOPT_HTTPAUTH or
1874 CURLOPT_PROXYAUTH options. They are thus individually controlled to server
1877 (note that there's no way to activate this with the curl tool yet)
1879 Daniel Fandrich (9 Dec 2008)
1880 - Added test cases 1089 and 1090 to test --write-out after a redirect to
1881 test a report that the size didn't work, but these test cases pass.
1883 - Documented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.
1885 Daniel Stenberg (9 Dec 2008)
1886 - Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
1887 particular state for the control connection like it did before for implicit
1888 FTPS (libcurl assumed such control connections to be encrypted while some
1889 FTPS servers such as FileZilla assumes such connections to be clear
1890 mode). Use the CURLOPT_USE_SSL option to set your desired level.
1892 Daniel Stenberg (8 Dec 2008)
1893 - Fred Machado posted about a weird FTP problem on the curl-users list and when
1894 researching it, it turned out he got a 550 response back from a SIZE command
1895 and then I fell over the text in RFC3659 that says:
1897 The presence of the 550 error response to a SIZE command MUST NOT be taken
1898 by the client as an indication that the file cannot be transferred in the
1899 current MODE and TYPE.
1901 In other words: the change I did on September 30th 2008 and that has been
1902 included in the last two releases were a regression and a bad idea. We MUST
1903 NOT take a 550 response from SIZE as a hint that the file doesn't exist.
1905 - Christian Krause filed bug #2221237
1906 (http://curl.haxx.se/bug/view.cgi?id=2221237) that identified an infinite
1907 loop during GSS authentication given some specific conditions. With his
1908 patience and great feedback I managed to narrow down the problem and
1909 eventually fix it although I can't test any of this myself!
1911 Daniel Fandrich (3 Dec 2008)
1912 - Fixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
1913 support (e.g. Minix)
1915 Daniel Stenberg (3 Dec 2008)
1916 - Igor Novoseltsev filed bug #2351645
1917 (http://curl.haxx.se/bug/view.cgi?id=2351645) that identified a problem with
1918 the multi interface that occured if you removed an easy handle while in
1919 progress and the handle was used in a HTTP pipeline.
1921 - Pawel Kierski pointed out a mistake in the cookie code that could lead to a
1922 bad fclose() after a fatal error had occured.
1923 (http://curl.haxx.se/bug/view.cgi?id=2382219)
1925 Daniel Fandrich (25 Nov 2008)
1926 - If a HTTP request is Basic and num is already >=1000, the HTTP test
1927 server adds 1 to num to get the data section to return. This allows
1928 testing authentication negotiations using the Basic authentication
1931 - Added tests 1087 and 1088 to test Basic authentication on a redirect
1932 with and without --location-trusted
1934 Daniel Stenberg (24 Nov 2008)
1935 - Based on a patch by Vlad Grachov, libcurl now uses a new libssh2 0.19
1936 function when built to support SCP and SFTP that helps the library to know
1937 in which direction a particular libssh2 operation would return EAGAIN so
1938 that libcurl knows what socket conditions to wait for before trying the
1939 function call again. Previously (and still when using libssh2 0.18 or
1940 earlier), libcurl will busy-loop in this situation when the easy interface
1943 Daniel Fandrich (20 Nov 2008)
1944 - Automatically detect OpenBSD's CA cert bundle.
1946 Daniel Stenberg (19 Nov 2008)
1947 - I removed the default use of "Pragma: no-cache" from libcurl when a proxy is
1948 used. It has been used since forever but it was never a good idea to use
1949 unless explicitly asked for.
1951 - Josef Wolf's extension that allows a $TESTDIR/gdbinit$testnum file that when
1952 you use runtests.pl -g, will be sourced by gdb to allow additional fancy or
1953 whatever you see fit
1955 - Christian Krause reported and fixed a memory leak that would occur with HTTP
1956 GSS/kerberos authentication (http://curl.haxx.se/bug/view.cgi?id=2284386)
1958 - Andreas Wurf and Markus Koetter helped me analyze a problem that Andreas got
1959 when uploading files to a single FTP server using multiple easy handle
1960 handles with the multi interface. Occasionally a handle would stall in
1963 The problem turned out to be a side-effect of the ConnectionExists()
1964 function's eagerness to re-use a handle for HTTP pipelining so it would
1965 select it even if already being in use, due to an inadequate check for its
1966 chances of being used for pipelnining.
1968 Daniel Fandrich (17 Nov 2008)
1969 - Added more compiler warning options for gcc 4.3
1971 Yang Tse (17 Nov 2008)
1972 - Fix a remaining problem in the inet_pton() runtime configure check. And
1973 fix internal Curl_inet_pton() failures to reject certain malformed literals.
1975 - Make configure script check if ioctl with the SIOCGIFADDR command can be
1976 used, and define HAVE_IOCTL_SIOCGIFADDR if appropriate.
1978 Daniel Stenberg (16 Nov 2008)
1979 - Christian Krause fixed a build failure when building with gss support
1980 enabled and FTP disabled.
1982 - Added check for NULL returns from strdup() in src/main.c and lib/formdata.c
1983 - reported by Jim Meyering also prevent buffer overflow on MSDOS when you do
1984 for example -O on a url with a file name part longer than PATH_MAX letters
1986 - lib/nss.c fixes based on the report by Jim Meyering: I went over and added
1987 checks for return codes for all calls to malloc and strdup that were
1988 missing. I also changed a few malloc(13) to use arrays on the stack and a
1989 few malloc(PATH_MAX) to instead use aprintf() to lower memory use.
1991 - I fixed a memory leak in Curl_nss_connect() when CURLOPT_ISSUERCERT is
1994 Daniel Fandrich (14 Nov 2008)
1995 - Added .xml as one of the few common file extensions known by the multipart
1998 - Added some #ifdefs around header files and change the EAGAIN test to
1999 fix compilation on Cell (reported by Jeff Curley).
2001 Yang Tse (14 Nov 2008)
2002 - Fixed several configure script issues affecting checks for inet_ntoa_r(),
2003 inet_ntop(), inet_pton(), getifaddrs(), fcntl() and getaddrinfo().
2005 Yang Tse (13 Nov 2008)
2006 - Refactored configure script detection of functions used to set sockets into
2007 non-blocking mode, and decouple function detection from function capability.