5 \___|\___/|_| \_\_____|
10 - Steven M. Schweda fixed VMS builder bad behavior when used in a batch job,
11 removed obsolete batch_compile.com and defines.com and updated VMS readme.
13 Version 7.20.0 (9 February 2010)
15 Daniel Stenberg (9 Feb 2010)
16 - When downloading compressed content over HTTP and the app asked libcurl to
17 automatically uncompress it with the CURLOPT_ENCODING option, libcurl could
18 wrongly provide the callback with more data than the maximum documented
19 amount. An application could thus get tricked into badness if the maximum
20 limit was trusted to be enforced by libcurl itself (as it is documented).
22 This is further detailed and explained in the libcurl security advisory
25 http://curl.haxx.se/docs/adv_20100209.html
27 Daniel Fandrich (3 Feb 2010)
28 - Changed the Watcom makefiles to make them easier to keep in sync with
29 Makefile.inc since that can't be included directly.
32 - Symbol CURL_FORMAT_OFF_T now obsoleted, will be removed in a future release,
33 symbol will not be available when building with CURL_NO_OLDIES defined. Use
34 of CURL_FORMAT_CURL_OFF_T is preferred since 7.19.0
36 Daniel Stenberg (1 Feb 2010)
37 - Using the multi_socket API, it turns out at times it seemed to "forget"
38 connections (which caused a hang). It turned out to be an existing (7.19.7)
39 bug in libcurl (that's been around for a long time) and it happened like
42 The app calls curl_multi_add_handle() to add a new easy handle, libcurl will
43 then set it to timeout in 1 millisecond so libcurl will tell the app about
46 The app's timeout fires off that there's a timeout, the app calls libcurl as
47 we so often document it:
50 res = curl_multi_socket_action(... TIMEOUT ...);
51 } while(CURLM_CALL_MULTI_PERFORM == res);
53 And this is the problem number one:
55 When curl_multi_socket_action() is called with no specific handle, but only
56 a timeout-action, it will *only* perform actions within libcurl that are
57 marked to run at this time. In this case, the request would go from INIT to
58 CONNECT and return CURLM_CALL_MULTI_PERFORM. When the app then calls libcurl
59 again, there's no timer set for this handle so it remains in the CONNECT
60 state. The CONNECT state is a transitional state in libcurl so it reports no
61 sockets there, and thus libcurl never tells the app anything more about that
62 easy handle/connection.
64 libcurl _does_ set a 1ms timeout for the handle at the end of
65 multi_runsingle() if it returns CURLM_CALL_MULTI_PERFORM, but since the loop
66 is instant the new job is not ready to run at that point (and there's no
67 code that makes libcurl call the app to update the timout for this new
68 timeout). It will simply rely on that some other timeout will trigger later
69 on or that something else will update the timeout callback. This makes the
70 bug fairly hard to repeat.
72 The fix made to adress this issue:
74 We introduce a loop in lib/multi.c around all calls to multi_runsingle() and
75 simply check for CURLM_CALL_MULTI_PERFORM internally. This has the added
76 benefit that this goes in line with my long-term wishes to get rid of the
77 CURLM_CALL_MULTI_PERFORM all together from the public API.
79 The downside of this fix, is that the counter we return in 'running_handles'
80 in several of our public functions then gets a slightly new and possibly
81 confusing behavior during times:
83 If an app adds a handle that fails to connect (very quickly) it may just
84 as well never appear as a 'running_handle' with this fix. Previously it
85 would first bump the counter only to get it decreased again at next call.
86 Even I have used that change in handle counter to signal "end of a
87 transfer". The only *good* way to find the end of a individual transfer
88 is calling curl_multi_info_read() to see if it returns one.
90 Of course, if the app previously did the looping before it checked the
91 counter, it really shouldn't be any new effect.
93 Yang Tse (26 Jan 2010)
94 - Constantine Sapuntzakis' and Joshua Kwan's work done in the last four months
95 relative to the asynchronous DNS lookups, along with with some integration
96 adjustments I have done are finally committed to CVS.
98 Currently these enhancements will benefit builds done using c-ares on any
99 platform as well as Windows builds using the default threaded resolver.
101 This release does not make generally available POSIX threaded DNS lookups
102 yet. There is no configure option to enable this feature yet. It is possible
103 to experimantally try this feature running configure with compiler flags that
104 make simultaneous definition of preprocessor symbols USE_THREADS_POSIX and
105 HAVE_PTHREAD_H, as well as whatever reentrancy compiler flags and linker ones
106 are required to link and properly use pthread_* functions on each platform.
108 Daniel Stenberg (26 Jan 2010)
109 - Mike Crowe made libcurl return CURLE_COULDNT_RESOLVE_PROXY when it is the
110 proxy that cannot be resolved when using c-ares. This matches the behaviour
111 when not using c-ares.
113 Björn Stenberg (23 Jan 2010)
114 - Added a new flag: -J/--remote-header-name. This option tells the
115 -O/--remote-name option to use the server-specified Content-Disposition
116 filename instead of extracting a filename from the URL.
118 Daniel Stenberg (21 Jan 2010)
119 - Chris Conroy brought support for RTSP transfers, and with it comes 8(!) new
120 libcurl options for controlling what to get and how to receive posssibly
121 interleaved RTP data.
123 Daniel Stenberg (20 Jan 2010)
124 - As was pointed out on the http-state mailing list, the order of cookies in a
125 HTTP Cookie: header _needs_ to be sorted on the path length in the cases
126 where two cookies using the same name are set more than once using
127 (overlapping) paths. Realizing this, identically named cookies must be
128 sorted correctly. But detecting only identically named cookies and take care
129 of them individually is harder than just to blindly and unconditionally sort
130 all cookies based on their path lengths. All major browsers also already do
131 this, so this makes our behavior one step closer to them in the cookie area.
133 Test case 8 was the only one that broke due to this change and I updated it
136 Daniel Stenberg (19 Jan 2010)
137 - David McCreedy brought a fix and a new test case (129) to make libcurl work
138 again when downloading files over FTP using ASCII and it turns out that the
139 final size of the file is not the same as the initial size the server
140 reported. This is very common since servers don't take the newline
141 conversions into account.
143 Kamil Dudka (14 Jan 2010)
144 - Suppressed side effect of OpenSSL configure checks, which prevented NSS from
145 being properly detected under certain circumstances. It had been caused by
146 strange behavior of pkg-config when handling PKG_CONFIG_LIBDIR. pkg-config
147 distinguishes among empty and non-existent environment variable in that case.
149 Daniel Stenberg (12 Jan 2010)
150 - Gil Weber reported a peculiar flaw with the multi interface when doing SFTP
151 transfers: curl_multi_fdset() would return -1 and not set and file
152 descriptors several times during a transfer of a single file. It turned out
153 to be due to two different flaws now fixed. Gil's excellent recipe helped me
156 Daniel Stenberg (11 Jan 2010)
157 - Made sure that the progress callback is repeatedly called at a regular
158 interval even during very slow connects.
160 - The tests/runtests.pl script now checks to see if the test case that runs is
161 present in the tests/data/Makefile.am and outputs a notice message on the
162 screen if not. Each test file has to be included in that Makefile.am to get
163 included in release archives and forgetting to add files there is a common
164 mistake. This is an attempt to make it harder to forget.
166 Daniel Stenberg (9 Jan 2010)
167 - Johan van Selst found and fixed a OpenSSL session ref count leak:
169 ossl_connect_step3() increments an SSL session handle reference counter on
170 each call. When sessions are re-used this reference counter may be
171 incremented many times, but it will be decremented only once when done (by
172 Curl_ossl_session_free()); and the internal OpenSSL data will not be freed
173 if this reference count remains positive. When a session is re-used the
174 reference counter should be corrected by explicitly calling
175 SSL_SESSION_free() after each consecutive SSL_get1_session() to avoid
176 introducing a memory leak.
178 (http://curl.haxx.se/bug/view.cgi?id=2926284)
180 Daniel Stenberg (7 Jan 2010)
181 - Make sure the progress callback is called repeatedly even during very slow
182 name resolves when c-ares is used for resolving.
184 Claes Jakobsson (6 Jan 2010)
185 - Julien Chaffraix fixed so that the fragment part in an URL is not sent
186 to the server anymore.
188 Kamil Dudka (3 Jan 2010)
189 - Julien Chaffraix eliminated a duplicated initialization in singlesocket().
191 Daniel Stenberg (2 Jan 2010)
192 - Make curl support --ssl and --ssl-reqd instead of the previous FTP-specific
193 versions --ftp-ssl and --ftp-ssl-reqd as these options are now used to
194 control SSL/TLS for IMAP, POP3 and SMTP as well in addition to FTP. The old
195 option names are still working but the new ones are the ones listed and
198 Daniel Stenberg (1 Jan 2010)
199 - Ingmar Runge enhanced libcurl's FTP engine to support the PRET command. This
200 command is a special "hack" used by the drftpd server, but even though it is
201 a custom extension I've deemed it fine to add to libcurl since this server
202 seems to survive and people keep using it and want libcurl to support
203 it. The new libcurl option is named CURLOPT_FTP_USE_PRET, and it is also
204 usable from the curl tool with --ftp-pret. Using this option on a server
205 that doesn't support this command will make libcurl fail.
207 I added test cases 1107 and 1108 to verify the functionality.
209 The PRET command is documented at
210 http://www.drftpd.org/index.php/Distributed_PASV
212 Yang Tse (30 Dec 2009)
213 - Steven M. Schweda improved VMS build system, and Craig A. Berry helped
214 with the patch and testing.
216 Daniel Stenberg (26 Dec 2009)
217 - Renato Botelho and Peter Pentchev brought a patch that makes the libcurl
218 headers work correctly even on FreeBSD systems before v8.
220 (http://curl.haxx.se/bug/view.cgi?id=2916915)
222 Daniel Stenberg (17 Dec 2009)
223 - David Byron fixed Curl_ossl_cleanup to actually call ENGINE_cleanup when
226 - Follow-up fix for the proxy fix I did for Jon Nelson's bug. It turned out I
227 was a bit too quick and broke test case 1101 with that change. The order of
228 some of the setups is sensitive. I now changed it slightly again to make
229 sure we do them in this order:
231 1 - parse URL and figure out what protocol is used in the URL
232 2 - prepend protocol:// to URL if missing
233 3 - parse name+password off URL, which needs to know what protocol is used
234 (since only some allows for name+password in the URL)
235 4 - figure out if a proxy should be used set by an option
236 5 - if no proxy option, check proxy environment variables
237 6 - run the protocol-specific setup function, which needs to have the proxy
240 Daniel Stenberg (15 Dec 2009)
241 - Jon Nelson found a regression that turned out to be a flaw in how libcurl
242 detects and uses proxies based on the environment variables. If the proxy
243 was given as an explicit option it worked, but due to the setup order
244 mistake proxies would not be used fine for a few protocols when picked up
245 from '[protocol]_proxy'. Obviously this broke after 7.19.4. I now also added
246 test case 1106 that verifies this functionality.
248 (http://curl.haxx.se/bug/view.cgi?id=2913886)
250 Daniel Stenberg (12 Dec 2009)
251 - IMAP, POP3 and SMTP support and their TLS versions (including IMAPS, POP3S
252 and SMTPS) are now supported. The current state may not yet be solid, but
253 the foundation is in place and the test suite has some initial support for
254 these protocols. Work will now persue to make them nice libcurl citizens
257 The work with supporting these new protocols was sponsored by
258 networking4all.com - thanks!
260 Daniel Stenberg (10 Dec 2009)
261 - Siegfried Gyuricsko found out that the curl manual said --retry would retry
262 on FTP errors in the transient 5xx range. Transient FTP errors are in the
263 4xx range. The code itself only tried on 5xx errors that occured _at login_.
264 Now the retry code retries on all FTP transfer failures that ended with a
267 (http://curl.haxx.se/bug/view.cgi?id=2911279)
269 - Constantine Sapuntzakis figured out a case which would lead to libcurl
270 accessing alredy freed memory and thus crash when using HTTPS (with
271 OpenSSL), multi interface and the CURLOPT_DEBUGFUNCTION and a certain order
272 of cleaning things up. I fixed it.
274 (http://curl.haxx.se/bug/view.cgi?id=2905220)
276 Daniel Stenberg (7 Dec 2009)
277 - Martin Storsjo made libcurl use the Expect: 100-continue header for posts
278 with unknown size. Previously it was only used for posts with a known size
279 larger than 1024 bytes.
281 Daniel Stenberg (1 Dec 2009)
282 - If the Expect: 100-continue header has been set by the application through
283 curl_easy_setopt with CURLOPT_HTTPHEADER, the library should set
284 data->state.expect100header accordingly - the current code (in 7.19.7 at
285 least) doesn't handle this properly. Martin Storsjo provided the fix!
287 Yang Tse (28 Nov 2009)
288 - Added Diffie-Hellman parameters to several test harness certificate files in
289 PEM format. Required by several stunnel versions used by our test harness.
291 Daniel Stenberg (28 Nov 2009)
292 - Markus Koetter provided a polished and updated version of Chad Monroe's TFTP
293 rework patch that now integrates TFTP properly into libcurl so that it can
294 be used non-blocking with the multi interface and more. BLKSIZE also works.
296 The --tftp-blksize option was added to allow setting the TFTP BLKSIZE from
299 Daniel Stenberg (26 Nov 2009)
300 - Extended and fixed the change I did on Dec 11 for the the progress
301 meter/callback during FTP command/response sequences. It turned out it was
302 really lame before and now the progress meter SHOULD get called at least
305 Daniel Stenberg (23 Nov 2009)
306 - Bjorn Augustsson reported a bug which made curl not report any problems even
307 though it failed to write a very small download to disk (done in a single
308 fwrite call). It turned out to be because fwrite() returned success, but
309 there was insufficient error-checking for the fclose() call which tricked
310 curl to believe things were fine.
312 Yang Tse (23 Nov 2009)
313 - David Byron modified Makefile.dist vc8 and vc9 targets in order to allow
314 finer granularity control when generating src and lib makefiles.
316 Yang Tse (22 Nov 2009)
317 - I modified configure to force removal of the curlbuild.h file included in
318 distribution tarballs for use by non-configure systems. As intended, this
319 would get overwriten when doing in-tree builds. But VPATH builds would end
320 having two curlbuild.h files, one in the source tree and another in the
321 build tree. With the modification I introduced 5 Nov 2009 this could become
322 an issue when running libcurl's test suite.
324 Daniel Stenberg (20 Nov 2009)
325 - Constantine Sapuntzakis identified a write after close, as the sockets were
326 closed by libcurl before the SSL lib were shutdown and they may write to its
327 socket. Detected to at least happen with OpenSSL builds.
329 - Jad Chamcham pointed out a bug with connection re-use. If a connection had
330 CURLOPT_HTTPPROXYTUNNEL enabled over a proxy, a subsequent request using the
331 same proxy with the tunnel option disabled would still wrongly re-use that
332 previous connection and the outcome would only be badness.
334 Yang Tse (18 Nov 2009)
335 - I modified the memory tracking system to make it intolerant with zero sized
336 malloc(), calloc() and realloc() function calls.
338 Daniel Stenberg (17 Nov 2009)
339 - Constantine Sapuntzakis provided another fix for the DNS cache that could
340 end up with entries that wouldn't time-out:
342 1. Set up a first web server that redirects (307) to a http://server:port
344 2. Have curl connect to the first web server using curl multi
346 After the curl_easy_cleanup call, there will be curl dns entries hanging
347 around with in_use != 0.
349 (http://curl.haxx.se/bug/view.cgi?id=2891591)
351 - Marc Kleine-Budde fixed: curl saved the LDFLAGS set during configure into
352 its pkg-config file. So -Wl stuff ended up in the .pc file, which is really
353 bad, and breaks if there are multiple -Wl in our LDFLAGS (which are in
354 PTXdist). bug #2893592 (http://curl.haxx.se/bug/view.cgi?id=2893592)
356 Kamil Dudka (15 Nov 2009)
357 - David Byron improved the configure script to use pkg-config to find OpenSSL
358 (and in particular the list of required libraries) even if a path is given
359 as argument to --with-ssl
361 Yang Tse (15 Nov 2009)
362 - I removed enable-thread / disable-thread configure option. These were only
363 placebo options. The library is always built as thread safe as possible on
366 Claes Jakobsson (14 Nov 2009)
367 - curl-config now accepts '--configure' to see what arguments was
368 passed to the configure script when building curl.
370 Daniel Stenberg (14 Nov 2009)
371 - Claes Jakobsson restored the configure functionality to detect NSS when
372 --with-nss is set but not "yes".
374 I think we can still improve that to check for pkg-config in that path etc,
375 but at least this patch brings back the same functionality we had before.
377 - Camille Moncelier added support for the file type SSL_FILETYPE_ENGINE for
378 the client certificate. It also disable the key name test as some engines
379 can select a private key/cert automatically (When there is only one key
380 and/or certificate on the hardware device used by the engine)
382 Yang Tse (14 Nov 2009)
383 - Constantine Sapuntzakis provided the fix that ensures that an SSL connection
384 won't be reused unless protection level for peer and host verification match.
386 I refactored how preprocessor symbol _THREAD_SAFE definition is done.
388 Kamil Dudka (12 Nov 2009)
389 - Kevin Baughman provided a fix preventing libcurl-NSS from crash on doubly
390 closed NSPR descriptor. The issue was hard to find, reported several times
391 before and always closed unresolved. More info at the RH bug:
392 https://bugzilla.redhat.com/534176
394 - libcurl-NSS now tries to reconnect with TLS disabled in case it detects
395 a broken TLS server. However it does not happen if SSL version is selected
396 manually. The approach was originally taken from PSM. Kaspar Brand helped me
397 to complete the patch. Original bug reports:
398 https://bugzilla.redhat.com/525496
399 https://bugzilla.redhat.com/527771
401 Yang Tse (12 Nov 2009)
402 - I modified configure script to make the getaddrinfo function check also
403 verify if the function is thread safe.
405 Yang Tse (11 Nov 2009)
406 - Marco Maggi reported that compilation failed when configured --with-gssapi
407 and GNU GSS installed due to a missing mutual exclusion of header files in
408 the Kerberos 5 code path. He also verified that my patch worked for him.
410 Daniel Stenberg (11 Nov 2009)
411 - Constantine Sapuntzakis posted bug #2891595
412 (http://curl.haxx.se/bug/view.cgi?id=2891595) which identified how an entry
413 in the DNS cache would linger too long if the request that added it was in
414 use that long. He also provided the patch that now makes libcurl capable of
415 still doing a request while the DNS hash entry may get timed out.
417 - Christian Schmitz noticed that the progress meter/callback was not properly
418 used during the FTP connection phase (after the actual TCP connect), while
419 it of course should be. I also made the speed check get called correctly so
420 that really slow servers will trigger that properly too.
422 Kamil Dudka (5 Nov 2009)
423 - Dropped misleading timeouts in libcurl-NSS and made sure the SSL socket works
424 in non-blocking mode.
426 Yang Tse (5 Nov 2009)
427 - I removed leading 'curl' path on the 'curlbuild.h' include statement in
428 curl.h, adjusting auto-makefiles include path, to enhance portability to
429 OS's without an orthogonal directory tree structure such as OS/400.
431 Daniel Stenberg (4 Nov 2009)
432 - I fixed several problems with the transfer progress meter. It showed the
433 wrong percentage for small files, most notable for <1000 bytes and could
434 easily end up showing more than 100% at the end. It also didn't show any
435 percentage, transfer size or estimated transfer times when transferring
438 Version 7.19.7 (4 November 2009)
440 Daniel Stenberg (2 Nov 2009)
441 - As reported independent by both Stan van de Burgt and Didier Brisebourg,
442 CURLINFO_SIZE_DOWNLOAD (the -w variable size_download) didn't work when
443 getting data from ldap!
445 Daniel Stenberg (31 Oct 2009)
446 - Gabriel Kuri reported a problem with CURLINFO_CONTENT_LENGTH_DOWNLOAD if the
447 download was 0 bytes, as libcurl would then return the size as unknown (-1)
448 and not 0. I wrote a fix and test case 566 to verify it.
450 Daniel Stenberg (30 Oct 2009)
451 - Liza Alenchery mentioned a problem with re-used SCP connection when a bad
452 auth is used, as it caused a crash. I failed to repeat the issue, but still
453 made a change that now forces the TCP connection used for a freed SCP
454 session to get closed and not be re-used.
456 - "Tom" posted a bug report that mentioned how libcurl did wrong when doing a
457 POST using a read callback, with Digest authentication and
458 "Transfer-Encoding: chunked" enforced. I would then cause the first request
459 to be wrongly sent and then basically hang until the server closed the
460 connection. I fixed the problem and added test case 565 to verify it.
462 Daniel Stenberg (25 Oct 2009)
463 - Dima Barsky made the curl cookie parser accept cookies even with blank or
464 unparsable expiry dates and then treat them as session cookies - previously
465 libcurl would reject cookies with a date format it couldn't parse. Research
466 shows that the major browser treat such cookies as session cookies. I
467 modified test 8 and 31 to verify this.
469 Daniel Stenberg (21 Oct 2009)
470 - Attempt to use pkg-config for finding out libssh2 installation details
473 - A patch in bug report #2883177 (http://curl.haxx.se/bug/view.cgi?id=2883177)
474 by Johan van Selst introduced the --crlfile option to curl, which makes curl
475 tell libcurl about a file with CRL (certificate revocation list) data to
478 Daniel Stenberg (18 Oct 2009)
479 - Ray Dassen provided a patch in Debian's bug tracker (bug number #551461)
480 that now makes curl_getdate(3) actually handles RFC 822 formatted dates that
481 use the "single letter military timezones".
482 http://www.rfc-ref.org/RFC-TEXTS/822/chapter5.html has the details.
484 - Fixed memory leak in the SCP/SFTP code as it never freed the knownhosts
487 - John Dennis filed bug report #2873666
488 (http://curl.haxx.se/bug/view.cgi?id=2873666) which identified a problem
489 which made libcurl loop infinitely when given incorrect credentials when
490 using HTTP GSS negotiate authentication. He also provided a small and simple
493 - Kevin Baughman found a double close() problem with libcurl-NSS, as when
494 libcurl called NSS to close the SSL "session" it also closed the actual
497 Yang Tse (17 Oct 2009)
498 - Bug report #2866724 indicated
499 (http://curl.haxx.se/bug/view.cgi?id=2866724) that curl on Windows failed
500 when writing files whose file names originally contained characters which
501 are not valid for file names on Windows. Dan Fandrich provided an initial
502 patch and another revised one to fix this issue.
504 Daniel Stenberg (1 Oct 2009)
505 - Tom Mueller correctly reported in bug report #2870221
506 (http://curl.haxx.se/bug/view.cgi?id=2870221) that libcurl returned an
507 incorrect return code from the internal trynextip() function which caused
508 him grief. This is a regression that was introduced in 7.19.1 and I find it
509 strange it hasn't hit us harder, but I won't persue into figuring out
512 - Constantine Sapuntzakis: The current implementation will always set
513 SO_SNDBUF to CURL_WRITE_SIZE even if the SO_SNDBUF starts out larger. The
514 patch doesn't do a setsockopt if SO_SNDBUF is already greater than
515 CURL_WRITE_SIZE. This should help folks who have set up their computer with
518 Daniel Stenberg (27 Sep 2009)
519 - I introduced a maximum limit for received HTTP headers. It is controlled by
520 the define CURL_MAX_HTTP_HEADER which is even exposed in the public header
521 file to allow for users to fairly easy rebuild libcurl with a modified
522 limit. The rationale for a fixed limit is that libcurl is realloc()ing a
523 buffer to be able to put a full header into it, so that it can call the
524 header callback with the entire header, but that also risk getting it into
525 trouble if a server by mistake or willingly sends a header that is more or
526 less without an end. The limit is set to 100K.
528 Daniel Stenberg (26 Sep 2009)
529 - John P. McCaskey posted a bug report that showed how libcurl did wrong when
530 saving received cookies with no given path, if the path in the request had a
531 query part. That is means a question mark (?) and characters on the right
532 side of that. I wrote test case 1105 and fixed this problem.
534 Kamil Dudka (26 Sep 2009)
535 - Implemented a protocol independent way to specify blocking direction, used by
536 transfer.c for blocking. It is currently used only by SCP and SFTP protocols.
537 This enhancement resolves an issue with 100% CPU usage during SFTP upload,
540 Daniel Stenberg (25 Sep 2009)
541 - Chris Mumford filed bug report #2861587
542 (http://curl.haxx.se/bug/view.cgi?id=2861587) identifying that libcurl used
543 the OpenSSL function X509_load_crl_file() wrongly and failed if it would
544 load a CRL file with more than one certificate within. This is now fixed.
546 Daniel Stenberg (16 Sep 2009)
547 - Sven Anders reported that we introduced a cert verfication flaw for OpenSSL-
548 powered libcurl in 7.19.6. If there was a X509v3 Subject Alternative Name
549 field in the certficate it had to match and so even if non-DNS and non-IP
550 entry was present it caused the verification to fail.
552 Daniel Fandrich (15 Sep 2009)
553 - Moved the libssh2 checks after the SSL library checks. This helps when
554 statically linking since libssh2 needs the SSL library link flags to be
555 set up already to satisfy its dependencies. This wouldn't be necessary if
556 the libssh2 configure check was changed to use pkg-config since the
557 --static flag would add the dependencies automatically.
559 Yang Tse (14 Sep 2009)
560 - Revert Joshua Kwan's patch committed 11 Sep 2009.
562 Some systems poll function sets POLLHUP in revents without setting
563 POLLIN, and sets POLLERR without setting POLLIN and POLLOUT. In some
564 libcurl code execution paths this could trigger busy wait loops with
565 high CPU usage until a timeout condition aborted the loop.
567 The reverted patch addressed the above issue for a very specific case,
568 when awaiting c-ares to resolve. A libcurl-wide fix for Curl_poll now
571 Guenter Knauf (11 Sep 2009)
572 - Joshua Kwan provided a patch to pass POLLERR / POLLHUP back to c-ares.
573 This fixes a loop problem with high CPU usage.
575 Daniel Stenberg (10 Sep 2009)
576 - Claes Jakobsson fixed a problem with cookie expiry dates at exctly the epoch
577 start second "Thu Jan 1 00:00:00 GMT 1970" as the date parser then returns 0
578 which internally then is treated as a session cookie. That particular date
579 is now made to get the value of 1.
581 Daniel Stenberg (2 Sep 2009)
582 - Daniel Johnson found a flaw in the code converting sftp-errors to libcurl
585 Daniel Stenberg (1 Sep 2009)
586 - Peter Sylvester made a debug feature for Curl_resolv() that now will force
587 libcurl to resolve 'localhost' whatever name you use in the URL *if* you set
588 the --interface option to (exactly) "LocalHost". This will enable us to
589 write tests for custom hosts names but still use a local host server.
591 - configure now tries to use pkg-config for a number of sub-dependencies even
592 when cross-compiling. The key to success is then you properly setup
593 PKG_CONFIG_PATH before invoking configure.
595 I also improved how NSS is detected by trying nss-config if pkg-config isn't
596 present, and as a last resort just use the lib name and force the user to
597 setup the LIBS/LDFLAGS/CFLAGS etc properly. The previous last resort would
598 add a range of various libs that would almost never be quite correct.
600 Daniel Stenberg (31 Aug 2009)
601 - When using the multi interface with FTP and you asked for NOBODY, you did no
602 QUOTE commands and the request used the same path as the connection had
603 already changed to, it would decide that no commands would be necessary for
604 the "DO" action and that was not handled properly but libcurl would instead
607 Kamil Dudka (28 Aug 2009)
608 - Improved error message for not matching certificate subject name in
609 libcurl-NSS. Originally reported at:
610 https://bugzilla.redhat.com/show_bug.cgi?id=516056#c9
612 Patrick Monnerat (24 Aug 2009)
613 - Introduced a SYST-based test to properly set-up name format when dealing
614 with the OS/400 FTP server.
616 - Fixed an ftp_readresp() bug preventing detection of failing control socket
617 and causing FTP client to loop forever.
619 Daniel Stenberg (24 Aug 2009)
620 - Marc de Bruin pointed out that configure --with-gnutls=PATH didn't work
621 properly and provided a fix. http://curl.haxx.se/bug/view.cgi?id=2843008
623 - Eric Wong introduced support for the new option -T. (dot) that makes curl
624 read stdin in a non-blocking fashion. This also brings back -T- (minus) to
625 the previous blocking behavior since it could break stuff for people at
628 Michal Marek (21 Aug 2009)
629 - With CURLOPT_PROXY_TRANSFER_MODE, avoid sending invalid URLs like
630 ftp://example.com;type=i if the user specified ftp://example.com without the
633 Daniel Stenberg (21 Aug 2009)
634 - Andre Guibert de Bruet pointed out a missing return code check for a
635 strdup() that could lead to segfault if it returned NULL. I extended his
636 suggest patch to now have Curl_retry_request() return a regular return code
637 and better check that.
639 - Lots of good work by Krister Johansen, mostly related to pipelining:
641 Fix SIGSEGV on free'd easy_conn when pipe unexpectedly breaks
642 Fix data corruption issue with re-connected transfers
643 Fix use after free if we're completed but easy_conn not NULL
645 Kamil Dudka (13 Aug 2009)
646 - Changed NSS code to not ignore the value of ssl.verifyhost and produce more
647 verbose error messages. Originally reported at:
648 https://bugzilla.redhat.com/show_bug.cgi?id=516056
650 Daniel Stenberg (12 Aug 2009)
651 - Karl Moerder fixed the Makefile.vc* makefiles to include the new file
652 nonblock.c so that they work fine again
654 - I expanded test 517 with a bunch of more dates that originate from the
655 Chrome browser test suite. It turns out most of them get parsed the same
658 Version 7.19.6 (12 August 2009)
660 Daniel Stenberg (12 Aug 2009)
661 - Carsten Lange reported a bug and provided a patch for TFTP upload and the
662 sending of the TSIZE option. I don't like fixing bugs just hours before
663 a release, but since it was broken and the patch fixes this for him I decided
666 Daniel Stenberg (11 Aug 2009)
667 - Peter Sylvester made the HTTPS test server use specific certificates for
668 each test, so that the test suite can now be used to actually test the
669 verification of cert names etc. This made an error show up in the OpenSSL-
670 specific code where it would attempt to match the CN field even if a
671 subjectAltName exists that doesn't match. This is now fixed and verified
674 - Benbuck Nason posted the bug report #2835196
675 (http://curl.haxx.se/bug/view.cgi?id=2835196), fixing a few compiler
676 warnings when mixing ints and bools.
678 Daniel Fandrich (10 Aug 2009)
679 - Fixed a memory leak in the FTP code and an off-by-one heap buffer overflow.
681 Daniel Fandrich (9 Aug 2009)
682 - Fixed some memory leaks in the command-line tool that caused most of the
683 torture tests to fail.
685 Daniel Stenberg (2 Aug 2009)
686 - Curt Bogmine reported a problem with SNI enabled on a particular server. We
687 should introduce an option to disable SNI, but as we're in feature freeze
688 now I've addressed the obvious bug here (pointed out by Peter Sylvester): we
689 shouldn't try to enable SNI when SSLv2 or SSLv3 is explicitly selected.
690 Code for OpenSSL and GnuTLS was fixed. NSS doesn't seem to have a particular
691 option for SNI, or are we simply not using it?
693 Daniel Stenberg (1 Aug 2009)
694 - Scott Cantor posted the bug report #2829955
695 (http://curl.haxx.se/bug/view.cgi?id=2829955) mentioning the recent SSL cert
696 verification flaw found and exploited by Moxie Marlinspike. The presentation
697 he did at Black Hat is available here:
698 https://www.blackhat.com/html/bh-usa-09/bh-usa-09-archives.html#Marlinspike
700 Apparently at least one CA allowed a subjectAltName or CN that contain a
701 zero byte, and thus clients that assumed they would never have zero bytes
702 were exploited to OK a certificate that didn't actually match the site. Like
703 if the name in the cert was "example.com\0theatualsite.com", libcurl would
704 happily verify that cert for example.com.
706 libcurl now better uses the length of the extracted name, not using the zero
707 termination for getting the string length.
709 This fixing only made and needed in OpenSSL interfacing code.
711 - Tanguy Fautre pointed out that OpenSSL's function RAND_screen() (present
712 only in some OpenSSL installs - like on Windows) isn't thread-safe and we
713 agreed that moving it to the global_init() function is a decent way to deal
716 - Alexander Beedie provided the patch for a noproxy problem: If I have set
717 CURLOPT_NOPROXY to "*", or to a host that should not use a proxy, I actually
718 could still end up using a proxy if a proxy environment variable was set.
720 Daniel Stenberg (27 Jul 2009)
721 - All the quote options (CURLOPT_QUOTE, CURLOPT_POSTQUOTE and
722 CURLOPT_PREQUOTE) now accept a preceeding asterisk before the command to
723 send when using FTP, as a sign that libcurl shall simply ignore the response
724 from the server instead of treating it as an error. Not treating a 400+ FTP
725 response code as an error means that failed commands will not abort the
726 chain of commands, nor will they cause the connection to get disconnected.
728 Daniel Stenberg (26 Jul 2009)
729 - Johan van Selst posted bug report #2825989
730 (http://curl.haxx.se/bug/view.cgi?id=2825989) pointing out that
731 OpenSSL-powered libcurl didn't support the SHA-2 digest algorithm, and
732 provided the solution too: to use OpenSSL_add_all_algorithms() in addition
733 to the older SSLeay_* alternative. OpenSSL_add_all_algorithms was added in
736 Daniel Stenberg (23 Jul 2009)
737 - Added CURLOPT_SSH_KNOWNHOSTS, CURLOPT_SSH_KEYFUNCTION, CURLOPT_SSH_KEYDATA.
738 They introduce known_host support for SSH keys to libcurl. See docs for
739 details. Note that this feature depends on a new enough libssh2 version, to
740 be supported in libssh2 1.2 and later (or current git repo at this time).
742 Michal Marek (22 Jul 2009)
743 - David Binderman found a memory and fd leak in lib/gtls.c:load_file()
744 (https://bugzilla.novell.com/523919). When looking at the code, I found that
745 also the ptr pointer can leak.
747 Kamil Dudka (20 Jul 2009)
748 - Claes Jakobsson improved the support for client certificates handling in
749 NSS-powered libcurl. Now the client certificates can be selected
750 automatically by a NSS built-in hook. Additionally pre-login to all PKCS11
751 slots is no more performed. It used to cause problems with HW tokens.
753 - Fixed reference counting for NSS client certificates. Now the PEM reader
754 module should be always properly unloaded on Curl_nss_cleanup(). If the
755 unload fails though, libcurl will try to reuse the already loaded instance.
757 Daniel Fandrich (15 Jul 2009)
758 - Added nonblock.c to the non-automake makefiles (note that the dependencies
759 in the Watcom makefiles aren't quite correct).
761 Michal Marek (15 Jul 2009)
762 - Changed the description of CURLINFO_OS_ERRNO to make it clear that the
763 errno is not reset on success.
765 Guenter Knauf (14 Jul 2009)
766 - renamed generated config.h to curl_config.h to avoid any future clashes
767 with config.h from other projects.
769 Daniel Stenberg (9 Jul 2009)
770 - Eric Wong introduced curlx_nonblock() that the curl tool now (re-)uses for
771 setting a file descriptor non-blocking. Used by the functionality Eric
772 himself brough on June 15th.
774 Daniel Stenberg (8 Jul 2009)
775 - Constantine Sapuntzakis posted bug report #2813123
776 (http://curl.haxx.se/bug/view.cgi?id=2813123) and an a patch that fixes the
779 Url A is accessed using auth. Url A redirects to Url B (on a different
780 server0. Url B reuses a persistent connection. Url B has auth, even though
781 it's on a different server.
783 Note: if Url B does not reuse a persistent connection, auth is not sent.
787 data->state.first_host is not initialized becuase Curl_http_connect is not
788 called when a connection is reused.
792 move initialization of data->state.first_host to Curl_http. No code before
793 Curl_http uses data->state.first_host anyway.
795 Guenter Knauf (4 Jul 2009)
796 - Markus Koetter provided a patch to avoid getnameinfo() usage which broke a
797 couple of both IPv4 and IPv6 autobuilds.
799 Daniel Stenberg (29 Jun 2009)
800 - Markus Koetter made CURLOPT_FTPPORT (and curl's -P/--ftpport) support a port
801 range if given colon-separated after the host name/address part. Like
802 "192.168.0.1:2000-10000"
804 - Modified the separators used for CURLOPT_CERTINFO in multi-part outputs. I
805 don't know how they got wrong in the first place, but using this output
806 format makes it possible to quite easily separate the string into an array
809 Daniel Fandrich (16 June 2009)
810 - Added a few more compiler warning options for gcc.
812 Daniel Stenberg (16 Jun 2009)
813 - Reuven Wachtfogel made curl -o - properly produce a binary output on windows
814 (no newline translations). Use -B/--use-ascii if you rather get the ascii
817 Michal Marek (16 Jun 2009)
818 - When doing non-anonymous ftp via http proxies and the password is not
819 provided in the url, add it there (squid needs this).
821 Daniel Stenberg (15 Jun 2009)
824 This allows curl(1) to be used as a client-side tunnel for arbitrary stream
825 protocols by abusing chunked transfer encoding in both the HTTP request and
826 HTTP response. This requires server support for sending a response while a
827 request is still being read, of course.
829 If attempting to read from stdin returns EAGAIN, then we pause our sender.
830 This leaves curl to attempt to read from the socket while reading from stdin
831 (and thus sending) is paused.
833 This change was needed to allow successfully tunneling the git protocol over
834 HTTP (--no-buffer is needed, as well).
836 Patrick Monnerat (15 Jun 2009)
837 - Replaced use of standard C library rand()/srand() by our own pseudo-random
840 Yang Tse (11 Jun 2009)
841 - I adapted testcurl script to allow building test harness programs when
842 cross-compiling for a *-*-mingw* host.
844 Daniel Stenberg (10 Jun 2009)
845 - Fabian Keil ran clang on the (lib)curl code, found a bunch of warnings and
846 contributed a range of patches to fix them.
848 Yang Tse (10 Jun 2009)
849 - I introduced configure script option --enable-curldebug which now allows
850 the decoupled enabling or disabling of the curl debug memory tracking
851 feature from the --enable-debug option which no longer controls this.
853 curl --version will list 'Debug' feature for debug enabled builds, and
854 will list 'TrackMemory' feature for curl debug memory tracking capable
855 builds. These features are independent and can be controlled when running
856 the configure script. When --enable-debug is given both features will be
857 enabled, unless some restriction prevents memory tracking from being used.
859 Internally, definition of preprocessor symbol DEBUGBUILD restricts code
860 which is only compiled for debug enabled builds. And symbol CURLDEBUG is
861 used to differentiate code which is _only_ used for memory tracking.
863 Yang Tse (9 Jun 2009)
864 - Daniel Steinberg pointed out that Curl_FormInit() in formdata.c was not
865 initializing the fread callback pointer and this triggered a compiler
866 warning, also provided a friendly suggestion on how to fix it.
868 Daniel Stenberg (8 Jun 2009)
869 - Claes Jakobsson provided a patch for libcurl-NSS that fixed a bad refcount
870 issue with client certs that caused issues like segfaults.
871 http://curl.haxx.se/mail/lib-2009-05/0316.html
873 - Triggered by bug report #2798852 and the patch in there, I fixed configure
874 to detect gnutls build options with pkg-config only and not libgnutls-config
875 anymore since GnuTLS has stopped distributing that tool. If an explicit path
876 is given to configure, we will instead guess on how to link and use that
877 lib. I did not use the patch from the bug report.
879 Yang Tse (8 Jun 2009)
880 - Igor Novoseltsev adjusted Makefile.vxworks to get sources and headers
881 included from Makefile.inc, and provided docs\INSTALL VxWorks section.
883 - I removed buildconf.bat from release and daily snapshot archives. This
884 file is only for CVS tree checkout builds.
886 Daniel Stenberg (8 Jun 2009)
887 - Eric Wong fixed --no-buffer to actually switch off output buffering. Been
890 Bill Hoffman (6 Jun 2009)
891 - Added some cmake docs and fixed socklen_t in the build.
893 Yang Tse (5 Jun 2009)
894 - John E. Malmberg provided VMS specific patch: "This fixes an existing bug
895 in urlglob.c where it was not converting the Curl Unix exit code to a VMS
896 DCL compatible exit code. This fix required the enhancement described next.
897 This also adds an enhancement to main.c so that when curl is run under a
898 Unix shell like Bash on VMS, it will return the standard Unix exit codes
899 and messages." And another patch for docs/examples.
901 I introduced os-specific.c and os-specific.h for use in curl tool code
902 and adjusted John E. Malmberg's patch placement to use these new files
903 as an effort to prevent main.c from growing ad infinitum. Code already
904 existing in main.c which is OS specific should be moved into these files.
906 Daniel Stenberg (4 June 2009)
907 - Setting the Content-Length: header from your app when you do a POST or PUT
908 is almost always a VERY BAD IDEA. Yet there are still apps out there doing
909 this, and now recently it triggered a bug/side-effect in libcurl as when
910 libcurl sends a POST or PUT with NTLM, it sends an empty post first when it
911 knows it will just get a 401/407 back. If the app then replaced the
912 Content-Length header, it caused the server to wait for input that libcurl
913 wouldn't send. Aaron Oneal reported this problem in bug report #2799008
914 (http://curl.haxx.se/bug/view.cgi?id=2799008) and helped us verify the fix.
916 Yang Tse (4 Jun 2009)
917 - Igor Novoseltsev provided patches and information, that after some
918 adjustments to better fit curl's way of doing things, have resulted
919 in the posibility of building libcurl for VxWorks.
921 Daniel Fandrich (2 June 2009)
922 - Checked in a Google Android make file. To use it, you must first
923 create a config.h file by running configure in the Android environment,
924 which doesn't seem to be easy to do. If no easy way can be found, a
925 static config-android.h may need to be created and checked in to the
928 Daniel Stenberg (1 June 2009)
929 - Claes Jakobsson fixed the configure script to better find and use NSS
932 Yang Tse (1 Jun 2009)
933 - John E. Malmberg provided a VMS specific clean-up for curl.h, and pointed
934 out that the configure script was failing to detect the timeval struct on
935 VMS when building with _XOPEN_SOURCE_EXTENDED undefined due to definition
936 taking place in socket.h instead of time.h. I have adjusted configure
937 script to also include this header when checking struct timeval.
939 Daniel Stenberg (27 May 2009)
940 - Frank McGeough provided a small OpenSSL #include fix to make libcurl compile
941 fine with Nokia 5th edition 1.0 SDK for Symbian.
943 - Andre Guibert de Bruet found a call to a OpenSSL function that didn't check
944 for a failure properly.
946 - Mike Crowe pointed out that setting CURLOPT_USERPWD to NULL used to clear
947 the auth credentials back in 7.19.0 and earlier while now you have to set ""
948 to get the same effect. His patch brings back the ability to use NULL.
950 - Claes Jakobsson fixed libcurl-NSS to build fine even without the
951 PK11_CreateGenericObject() function.
953 Daniel Stenberg (25 May 2009)
954 - bug report #2796358 (http://curl.haxx.se/bug/view.cgi?id=2796358) pointed
955 out that the cookie parser would leak memory when it parses cookies that are
956 received with domain, path etc set multiple times in the same header. While
957 such a cookie is questionable, they occur in the wild and libcurl no longer
958 leaks memory for them. I added such a header to test case 8.
960 Daniel Fandrich (22 May 2009)
961 - Removed some obsolete digest code that caused a valgrind error in test 551.
963 Daniel Fandrich (20 May 2009)
964 - Added "non-existing host" test keywords to make it easy to skip those
965 tests on machines that have broken DNS configurations (such as
966 those configured to use OpenDNS).
968 Daniel Stenberg (19 May 2009)
969 - Kamil Dudka brought the patch from the Redhat bug entry
970 https://bugzilla.redhat.com/show_bug.cgi?id=427966 which was libcurl closing
971 a bad file descriptor when closing down the FTP data connection. Caolan
972 McNamara seems to be the original author of it.
974 Version 7.19.5 (18 May 2009)
976 Daniel Stenberg (17 May 2009)
977 - James Bursa posted a patch to the mailing list that fixed a problem with
978 no_proxy which made it not skip the proxy if the URL entered contained a
979 user name. I added test case 1101 to verify.
981 Daniel Stenberg (11 May 2009)
982 - Balint Szilakszi reported a memory leak when libcurl did gzip decompression
983 of streams that had some parts (legitimately) missing. We now provide and use
984 a proper cleanup function for the content encoding submodule.
985 http://curl.haxx.se/mail/lib-2009-05/0092.html
987 - Kamil Dudka provided a fix for libcurl-NSS reported by Michael Cronenworth
988 at https://bugzilla.redhat.com/show_bug.cgi?id=453612#c12
990 If an incorrect password is given while loading a private key, libcurl ends
991 up in an infinite loop consuming memory. The bug is critical.
993 - I fixed the problem with doing NTLM, POST and then following a 302 redirect,
994 as reported by Ebenezer Ikonne (on curl-users) and Laurent Rabret (on
995 curl-library). The transfer was mistakenly marked to get more data to send
996 but since it didn't actually have that, it just hung there...
998 Daniel Stenberg (10 May 2009)
999 - Andre Guibert de Bruet correctly pointed out an over-alloc with one wasted
1000 byte in the digest code.
1002 Yang Tse (9 May 2009)
1003 - Removed DOS and TPF package's subdirectory Makefile.am, it was only used
1004 to include some files in the distribution tarball serving no other purpose.
1005 Files from the DOS and TPF subdirectories are now included in the EXTRA_DIST
1006 of the Makefile in the parent subdirectory.
1008 Yang Tse (8 May 2009)
1009 - Changed host name literal in several tests to one under the haxx.se domain.
1011 - Renamed vc6 workspace and project files to avoid filename clash when used
1012 for conversion to later VS versions.
1014 Daniel Stenberg (8 May 2009)
1015 - Constantine Sapuntzakis fixed bug report #2784055
1016 (http://curl.haxx.se/bug/view.cgi?id=2784055) identifying a problem to
1017 connect to SOCKS proxies when using the multi interface. It turned out to
1018 almost not work at all previously. We need to wait for the TCP connect to
1019 be properly verified before doing the SOCKS magic.
1021 There's still a flaw in the FTP code for this.
1023 Daniel Stenberg (7 May 2009)
1024 - Made the SO_SNDBUF setting for the data connection socket for ftp uploads as
1025 well. See change 28 Apr 2009.
1027 Yang Tse (7 May 2009)
1028 - Fixed an issue affecting FTP transfers, introduced with the transfer.c
1029 patch committed May 4.
1031 Daniel Stenberg (7 May 2009)
1032 - Man page *roff problems fixed thanks to input from Colin Watson. Problems
1033 reported in the Debian package.
1035 - Vijay G filed bug report #2723236
1036 (http://curl.haxx.se/bug/view.cgi?id=2723236) identifying a problem with
1037 libcurl's TFTP code and its lack of dealing with the OACK packet.
1039 Yang Tse (5 May 2009)
1040 - Fixed the --ftp-port address of test #251 to the CLIENTIP address, and
1041 reverted the change affecting test suite harness committed 4 May.
1043 Daniel Stenberg (5 May 2009)
1044 - Inspired by Michael Smith's session id fix for OpenSSL, I did the
1045 corresponding fix in the GnuTLS code: make sure to store the new session id
1046 in case the previous re-used one is rejected.
1048 Daniel Stenberg (4 May 2009)
1049 - Michael Smith posted bug report #2786255
1050 (http://curl.haxx.se/bug/view.cgi?id=2786255) with a patch, identifying how
1051 libcurl did not deal with SSL session ids properly if the server rejected a
1052 re-use of one. Starting now, it will forget the rejected one and remember
1053 the new. This change was for OpenSSL only, it is likely that other SSL lib
1054 code needs similar fixes.
1056 Yang Tse (4 May 2009)
1057 - Applied David McCreedy's "transfer.c fixes for CURL_DO_LINEEND_CONV and
1058 non-ASCII platform HTTP requests" patch addressing two HTTP PUT problems:
1059 1) On non-ASCII platforms not all of the protocol portions of the PUT are
1060 being translated to ASCII. 2) On all platforms the line endings of part of
1061 the protocol portions are mangled from CRLF to CRCRLF if data->set.crlf or
1062 data->set.prefer_ascii are set (depending on CURL_DO_LINEEND_CONV).
1064 - Applied David McCreedy's patch to fix test suite harness to allow test FTP
1065 server and client on different machines, providing FTP client address when
1066 running the FTP test server.
1068 Daniel Fandrich (3 May 2009)
1069 - Added and disabled test case 563 which shows KNOWN_BUGS #59. The bug
1070 report failed to mention that a proxy must be used to reproduce it.
1072 Yang Tse (2 May 2009)
1073 - Use a build-time configured curl_socklen_t data type instead of socklen_t.
1075 Yang Tse (1 May 2009)
1076 - Applied David McCreedy's patches "TPF-platform specific changes to various
1077 files" and "http.c fix to Curl_proxyCONNECT for non-ASCII platforms", the
1078 former with minor edits.
1080 Daniel Stenberg (30 Apr 2009)
1081 - I was going to fix issue #59 in KNOWN_BUGS
1083 If the CURLOPT_PORT option is used on an FTP URL like
1084 "ftp://example.com/file;type=A" the ";type=A" is stripped off.
1086 I added test case 562 to verify, only to find out that I couldn't repeat
1087 this bug so I hereby consider it not a bug anymore!
1089 Daniel Stenberg (29 Apr 2009)
1090 - Based on bug report #2723219 (http://curl.haxx.se/bug/view.cgi?id=2723219)
1091 I've now made TFTP "connections" not being kept for re-use within libcurl.
1092 TFTP is UDP-based so the benefit was really low (if even existing) to begin
1093 with so instead of tracking down to fix this problem we instead removed the
1094 re-use. I also enabled test case 1099 that I wrote a few days ago to verify
1095 that this change fixes the reported problem.
1097 Daniel Stenberg (28 Apr 2009)
1098 - Constantine Sapuntzakis filed bug report #2783090
1099 (http://curl.haxx.se/bug/view.cgi?id=2783090) pointing out that on windows
1100 we need to grow the SO_SNDBUF buffer somewhat to get really good upload
1101 speeds. http://support.microsoft.com/kb/823764 has the details. Friends
1102 confirmed that simply adding 32 to CURL_MAX_WRITE_SIZE is enough.
1104 - Bug report #2709004 (http://curl.haxx.se/bug/view.cgi?id=2709004) by Tim
1105 Chen pointed out how curl couldn't upload with resume when reading from a
1108 This ended up with the introduction of a new return code for the
1109 CURLOPT_SEEKFUNCTION callback that basically says that the seek failed but
1110 that libcurl may try to resolve the situation anyway. In our case this means
1111 libcurl will attempt to instead read that much data from the stream instead
1112 of seeking and that way curl can now upload with resume when data is read
1115 Daniel Stenberg (26 Apr 2009)
1116 - Bug report #2779733 (http://curl.haxx.se/bug/view.cgi?id=2779733) by Sven
1117 Wegener pointed out that CURLINFO_APPCONNECT_TIME didn't work with the multi
1118 interface and provided a patch that fixed the problem!
1120 Daniel Stenberg (24 Apr 2009)
1121 - Kamil Dudka fixed another NSS-related leak when client certs were used.
1123 - Bug report #2779245 (http://curl.haxx.se/bug/view.cgi?id=2779245) by Rainer
1124 Koenig pointed out that the man page didn't tell that the *_proxy
1125 environment variables can be specified lower case or UPPER CASE and the
1126 lower case takes precedence,
1128 Daniel Fandrich (21 Apr 2009)
1129 - Added new libcurl source files to Amiga, RiscOS and VC6 build files.
1131 Yang Tse (21 Apr 2009)
1132 - Moved potential inclusion of system's malloc.h and memory.h header files to
1133 setup_once.h. Inclusion of each header file is based on the definition of
1134 NEED_MALLOC_H and NEED_MEMORY_H respectively.
1136 Renamed libcurl's memory.h to curl_memory.h
1138 Daniel Stenberg (20 Apr 2009)
1139 - Leanic Lefever reported a crash and did some detailed research on why and
1140 how it occurs (http://curl.haxx.se/mail/lib-2009-04/0289.html). The
1141 conclusion was that if an error is detected and Curl_done() is called for
1142 the connection, ftp_done() could at times return another error code that
1143 then would take precedence and that new code confused existing logic that
1144 works for the first error code (CURLE_SEND_ERROR) only.
1146 - Gisle Vanem noticed that --libtool would produce bogus strings at times for
1147 OBJECTPOINT options. Now we've introduced a new function - my_setopt_str -
1148 within the app for setting plain string options to avoid the risk of this
1151 Daniel Stenberg (17 Apr 2009)
1152 - Pramod Sharma reported and tracked down a bug when doing FTP over a HTTP
1153 proxy. libcurl would then wrongly close the connection after each
1154 request. In his case it had the weird side-effect that it killed NTLM auth
1155 for the proxy causing an inifinite loop!
1157 I added test case 1098 to verify this fix. The test case does however not
1158 properly verify that the transfers are done persistently - as I couldn't
1159 think of a clever way to achieve it right now - but you need to read the
1160 stderr output after a test run to see that it truly did the right thing.
1162 Daniel Stenberg (13 Apr 2009)
1163 - bug report #2727981 (http://curl.haxx.se/bug/view.cgi?id=2727981) by Martin
1164 Storsjö pointed out how setting CURLOPT_NOBODY to 0 could be downright
1165 confusing as it set the method to either GET or HEAD. The example he showed
1168 curl_easy_setopt(curl, CURLOPT_PUT, 1);
1169 curl_easy_setopt(curl, CURLOPT_NOBODY, 0);
1171 The new way doesn't alter the method until the request is about to start. If
1172 CURLOPT_NOBODY is then 1 the HTTP request will be HEAD. If CURLOPT_NOBODY is
1173 0 and the request happens to have been set to HEAD, it will then instead be
1174 set to GET. I believe this will be less surprising to users, and hopefully
1175 not hit any existing users badly.
1177 - Toshio Kuratomi reported a memory leak problem with libcurl+NSS that turned
1178 out to be leaking cacerts. Kamil Dudka helped me complete the fix. The issue
1179 is found in Redhat's bug tracker:
1180 https://bugzilla.redhat.com/show_bug.cgi?id=453612
1182 There are still memory leaks present, but they seem to have other reasons.
1184 Daniel Fandrich (11 Apr 2009)
1185 - Added new libcurl source files to Symbian OS build files.
1186 - Improved Symbian support for SSL.
1188 Yang Tse (10 Apr 2009)
1189 - Daniel Johnson improved the MacOSX-Framework shell script to now perform all
1190 the steps required to build a Mac OS X four way fat ppc/i386/ppc64/x86_64
1191 libcurl.framework. Four way fat framework requires OS X 10.5 SDK or later.
1193 Yang Tse (8 Apr 2009)
1194 - Removed Sun compilers preprocessor block from curlbuild.h.dist, this also
1195 removes it from the curlbuild.h file originally distributed by the cURL
1196 project as this file is intended for systems not capable of running the
1197 configure script. For those who have been building curl out of the source
1198 code curl distribution tarball provided by curl.haxx.se the change implies
1199 nothing. Previous change in this area committed 2 Apr becomes irrelevant.
1201 Daniel Stenberg (6 Apr 2009)
1202 - I clarified in the docs that CURLOPT_SEEKFUNCTION should return 0 on success
1203 and 1 on fatal errors. Previously it only mentioned non-zero on fatal
1204 errors. This is a slight change in meaning, but it follows what we've done
1205 elsewhere before and it opens up for LOTS of more useful return codes
1206 whenever we can think of them...
1208 Yang Tse (2 Apr 2009)
1209 - Fix curl_off_t definition for builds done using Sun compilers and a
1210 non-configured libcurl. In this case curl_off_t data type was gated
1211 to the off_t data type which depends on the _FILE_OFFSET_BITS. This
1212 configuration is exactly the unwanted configuration for our curl_off_t
1213 data type which must not depend on such setting. This breaks ABI for
1214 libcurl libraries built with Sun compilers which were built without
1215 having run the configure script with _FILE_OFFSET_BITS different than
1216 64 and using the ILP32 data model.
1218 Daniel Stenberg (1 Apr 2009)
1219 - Andre Guibert de Bruet fixed a NULL pointer use in an infof() call if a
1220 strdup() call failed.
1222 Daniel Fandrich (31 Mar 2009)
1223 - Properly return an error code in curl_easy_recv (reported by Jim Freeman).
1225 Daniel Stenberg (18 Mar 2009)
1226 - Kamil Dudka brought a patch that enables 6 additional crypto algorithms when
1227 NSS is used. These ciphers were added in NSS 3.4 and require to be enabled
1230 Daniel Stenberg (13 Mar 2009)
1231 - Use libssh2_version() to present the libssh2 version in case the libssh2
1232 library is found to support it.
1234 Yang Tse (12 Mar 2009)
1235 - Added missing Curl_read() return code checking in TELNET transfers.
1237 - Pierre Brico found and fixed TELNET transfers not being aborted upon
1238 a write callback failure.
1240 Daniel Stenberg (11 Mar 2009)
1241 - Kamil Dudka made the curl tool properly call curl_global_init() before any
1242 other libcurl function.
1244 Yang Tse (11 Mar 2009)
1245 - Added missing TELNET timeout support for Windows builds. This issue was
1246 reported by Pierre Brico.
1248 Daniel Stenberg (9 Mar 2009)
1249 - Frank Hempel found out a bug and provided the fix:
1251 curl_easy_duphandle did not necessarily duplicate the CURLOPT_COOKIEFILE
1252 option. It only enabled the cookie engine in the destination handle if
1253 data->cookies is not NULL (where data is the source handle). In case of a
1254 newly initialized handle which just had the cookie support enabled by a
1255 curl_easy_setopt(handle, CURL_COOKIEFILE, "")-call, handle->cookies was
1256 still NULL because the setopt-call only appends the value to
1257 data->change.cookielist, hence duplicating this handle would not have the
1258 cookie engine switched on.
1260 We also concluded that the slist-functionality would be suitable for being
1261 put in its own module rather than simply hanging out in lib/sendf.c so I
1262 created lib/slist.[ch] for them.
1264 - Andreas Farber made the 'buildconf' script check for the presence of m4
1265 scripts to make it detect a bad checkout earlier. People with older
1266 checkouts who don't do cvs update with the -d option won't get the new dirs
1267 and then will get funny outputs that can be a bit hard to understand and
1270 Daniel Stenberg (8 Mar 2009)
1271 - Andre Guibert de Bruet found and fixed a code segment in ssluse.c where the
1272 allocation of the memory BIO was not being properly checked.
1274 - Andre Guibert de Bruet fixed the gnutls-using code: There are a few places
1275 in the gnutls code where we were checking for negative values for errors,
1276 when the man pages state that GNUTLS_E_SUCCESS is returned on success and
1277 other values indicate error conditions.
1279 - Bill Egert pointed out (http://curl.haxx.se/bug/view.cgi?id=2671602) that
1280 curl didn't use sprintf() in a way that is documented to work in POSIX but
1281 since we use our own printf() code (from libcurl) that shouldn't be a
1282 problem. Nonetheless I modified the code to not rely on such particular
1283 features and to not cause further raised eyebrowse with no good reason.
1285 Daniel Fandrich (5 Mar 2009)
1286 - Expanded the security section of the libcurl-tutorial man page to cover
1287 more issues for authors to consider when writing robust libcurl-using
1290 Yang Tse (5 Mar 2009)
1291 - Fixed NTLM authentication memory leak on SSPI enabled Windows builds. This
1292 issue was noticed by Chris Deidun.
1294 Daniel Fandrich (4 Mar 2009)
1295 - Fixed a problem with m4 quoting in the OpenSSL configure check reported
1298 Daniel Stenberg (3 Mar 2009)
1299 - David James brought a patch that make libcurl close (all) dead connections
1300 whenever you attempt to open a new connection.
1302 1. After cleaning up a dead connection, "continue" instead of
1303 returning FALSE. This ensures that we clean up all dead connections,
1304 rather than just cleaning up the first dead connection.
1305 2. Move up the cleanup for dead connections so that it occurs for
1306 all connections, rather than just the connections which have the same
1307 preferences as our current new connection.
1309 Version 7.19.4 (3 March 2009)
1311 Daniel Stenberg (3 Mar 2009)
1312 - David Kierznowski notified us about a security flaw
1313 (http://curl.haxx.se/docs/adv_20090303.html also known as CVE-2009-0037) in
1314 which previous libcurl versions (by design) can be tricked to access an
1315 arbitrary local/different file instead of a remote one when
1316 CURLOPT_FOLLOWLOCATION is enabled. This flaw is now fixed in this release
1317 together this the addition of two new setopt options for controlling this
1320 o CURLOPT_REDIR_PROTOCOLS controls what protocols libcurl is allowed to
1321 follow to when CURLOPT_FOLLOWLOCATION is enabled. By default, this option
1322 excludes the FILE and SCP protocols and thus you nee to explicitly allow
1323 them in your app if you really want that behavior.
1325 o CURLOPT_PROTOCOLS controls what protocol(s) libcurl is allowed to fetch
1326 using the primary URL option. This is useful if you want to allow a user or
1327 other outsiders control what URL to pass to libcurl and yet not allow all
1328 protocols libcurl may have been built to support.
1330 Daniel Stenberg (27 Feb 2009)
1331 - Senthil Raja Velu reported a problem when CURLOPT_INTERFACE and
1332 CURLOPT_LOCALPORT were used together (the local port bind failed), and
1333 Markus Koetter provided the fix!
1335 Daniel Stenberg (25 Feb 2009)
1336 - As Daniel Fandrich figured out, we must do the GnuTLS initing in the
1337 curl_global_init() function to properly maintain the performing functions
1338 thread-safe. We've previously (28 April 2007) moved the init to a later time
1339 just to avoid it to fail very early when libgcrypt dislikes the situation,
1340 but that move was bad and the fix should rather be in libgcrypt or
1343 Daniel Stenberg (24 Feb 2009)
1344 - Brian J. Murrell found out that Negotiate proxy authentication didn't work.
1345 It happened because the code used the struct for server-based auth all the
1346 time for both proxy and server auth which of course was wrong.
1348 Daniel Stenberg (23 Feb 2009)
1349 - After a bug reported by James Cheng I've made curl_easy_getinfo() for
1350 CURLINFO_CONTENT_LENGTH_DOWNLOAD and CURLINFO_CONTENT_LENGTH_UPLOAD return
1351 -1 if the sizes aren't know. Previously these returned 0, make it impossible
1352 to detect the difference between actually zero and unknown.
1354 Yang Tse (23 Feb 2009)
1355 - Daniel Johnson provided a shell script that will perform all the steps needed
1356 to build a Mac OS X fat ppc/i386 or ppc64/x86_64 libcurl.framework
1358 Daniel Stenberg (23 Feb 2009)
1359 - I renamed everything in the windows builds files that used the name 'curllib'
1360 to the proper 'libcurl' as clearly this caused confusion.
1362 Yang Tse (20 Feb 2009)
1363 - Do not halt compilation when using VS2008 to build a Windows 2000 target.
1365 Daniel Stenberg (20 Feb 2009)
1366 - Linus Nielsen Feltzing reported and helped me repeat and fix a problem with
1367 FTP with the multi interface: when a transfer fails, like when aborted by a
1368 write callback, the control connection was wrongly closed and thus not
1371 This change is also an attempt to cleanup the code somewhat in this area, as
1372 now the FTP code attempts to keep (better) track on pending responses
1373 necessary to get read in ftp_done().
1375 Daniel Stenberg (19 Feb 2009)
1376 - Patrik Thunstrom reported a problem and helped me repeat it. It turned out
1377 libcurl did a superfluous 1000ms wait when doing SFTP downloads!
1379 We read data with libssh2 while doing the "DO" operation for SFTP and then
1380 when we were about to start getting data for the actual file part, the
1381 "TRANSFER" part, we waited for socket action (in 1000ms) before doing a
1382 libssh2-read. But in this case libssh2 had already read and buffered the
1383 data so we ended up always just waiting 1000ms before we get working on the
1386 Patrick Monnerat (18 Feb 2009)
1387 - FTP downloads (i.e.: RETR) ending with code 550 now return error
1388 CURLE_REMOTE_FILE_NOT_FOUND instead of CURLE_FTP_COULDNT_RETR_FILE.
1390 Daniel Stenberg (17 Feb 2009)
1391 - Kamil Dudka made NSS-powered builds compile and run again!
1393 - A second follow-up change by Andre Guibert de Bruet to fix a related memory
1394 leak like that fixed on the 14th. When zlib returns failure, we need to
1395 cleanup properly before returning error.
1397 - CURLOPT_FTP_CREATE_MISSING_DIRS can now be set to 2 in addition to 1 for
1398 plain FTP connections, and it will then allow MKD to fail once and retry the
1399 CWD afterwards. This is especially useful if you're doing many simultanoes
1400 connections against the same server and they all have this option enabled,
1401 as then CWD may first fail but then another connection does MKD before this
1402 connection and thus MKD fails but trying CWD works! The numbers can
1403 (should?) now be set with the convenience enums now called
1404 CURLFTP_CREATE_DIR and CURLFTP_CREATE_DIR_RETRY.
1406 Tests has proven that if you're making an application that uploads a set of
1407 files to an ftp server, you will get a noticable gain in speed if you're
1408 using multiple connections and this option will be then be very useful.
1410 Daniel Stenberg (14 Feb 2009)
1411 - Andre Guibert de Bruet found and fixed a memory leak in the content encoding
1412 code, which could happen on libz errors.
1414 Daniel Fandrich (12 Feb 2009)
1415 - Added support for Digest and NTLM authentication using GnuTLS.
1417 Daniel Stenberg (11 Feb 2009)
1418 - CURLINFO_CONDITION_UNMET was added to allow an application to get to know if
1419 the condition in the previous request was unmet. This is typically a time
1420 condition set with CURLOPT_TIMECONDITION and was previously not possible to
1421 reliably figure out. From bug report #2565128
1422 (http://curl.haxx.se/bug/view.cgi?id=2565128) filed by Jocelyn Jaubert.
1424 Daniel Fandrich (4 Feb 2009)
1425 - Don't add the standard /usr/lib or /usr/include paths to LDFLAGS and CPPFLAGS
1426 (respectively) when --with-ssl=/usr is used (patch based on FreeBSD).
1428 - Added an explicit buffer limit check in msdosify() (patch based on FreeBSD).
1429 This couldn't ever overflow in curl, but might if the code were used
1430 elsewhere or under different conditions.
1432 Daniel Stenberg (3 Feb 2009)
1433 - Hidemoto Nakada provided a small fix that makes it possible to get the
1434 CURLINFO_CONTENT_LENGTH_DOWNLOAD size from file:// "transfers" with
1435 CURLOPT_NOBODY set true.
1437 Daniel Stenberg (2 Feb 2009)
1438 - Patrick Scott found a rather large memory leak when using the multi
1439 interface and setting CURLMOPT_MAXCONNECTS to something less than the number
1440 of handles you add to the multi handle. All the connections that didn't fit
1441 in the cache would not be properly disconnected nor freed!
1443 - Craig A West brought us: libcurl now defaults to do CONNECT with HTTP
1444 version 1.1 instead of 1.0 like before. This change also introduces the new
1445 proxy type for libcurl called 'CURLPROXY_HTTP_1_0' that then allows apps to
1446 switch (back) to CONNECT 1.0 requests. The curl tool also got a --proxy1.0
1447 option that works exactly like --proxy but sets CURLPROXY_HTTP_1_0.
1449 I updated all test cases cases that use CONNECT and I tried to do some using
1450 --proxy1.0 and some updated to do CONNECT 1.1 to get both versions run.
1452 Daniel Stenberg (31 Jan 2009)
1453 - When building with c-ares 1.6.1 (not yet released) or later and IPv6 support
1454 enabled, we can now take advantage of its brand new AF_UNSPEC support in
1455 ares_gethostbyname(). This makes test case 241 finally run fine for me with
1456 this setup since it now parses the "::1 ip6-localhost" line fine in my
1459 Daniel Stenberg (30 Jan 2009)
1460 - Scott Cantor filed bug report #2550061
1461 (http://curl.haxx.se/bug/view.cgi?id=2550061) mentioning that I failed to
1462 properly make sure that the VC9 makefiles got included in the latest
1463 release. I've now fixed the release script and verified it so next release
1464 will hopefully include them properly!
1466 Daniel Fandrich (30 Jan 2009)
1467 - Fixed --disable-proxy for FTP and SOCKS. Thanks to Daniel Egger for
1470 Yang Tse (29 Jan 2009)
1471 - Introduced curl_sspi.c and curl_sspi.h for the implementation of functions
1472 Curl_sspi_global_init() and Curl_sspi_global_cleanup() which previously were
1473 named Curl_ntlm_global_init() and Curl_ntlm_global_cleanup() in http_ntlm.c
1474 Also adjusted socks_sspi.c to remove the link-time dependency on the Windows
1475 SSPI library using it now in the same way as it was done in http_ntlm.c.
1477 Daniel Stenberg (28 Jan 2009)
1478 - Markus Moeller introduced two new options to libcurl:
1479 CURLOPT_SOCKS5_GSSAPI_SERVICE and CURLOPT_SOCKS5_GSSAPI_NEC to allow libcurl
1480 to do GSS-style authentication with SOCKS5 proxies. The curl tool got the
1481 options called --socks5-gssapi-service and --socks5-gssapi-nec to enable
1484 Daniel Stenberg (26 Jan 2009)
1485 - Chad Monroe provided the new CURLOPT_TFTP_BLKSIZE option that allows an app
1486 to set desired block size to use for TFTP transfers instead of the default
1489 - The "-no_ticket" option was introduced in Openssl0.9.8j. It's a flag to
1490 disable "rfc4507bis session ticket support". rfc4507bis was later turned
1491 into the proper RFC5077 it seems: http://tools.ietf.org/html/rfc5077
1493 The enabled extension concerns the session management. I wonder how often
1494 libcurl stops a connection and then resumes a TLS session. also, sending the
1495 session data is some overhead. .I suggest that you just use your proposed
1496 patch (which explicitly disables TICKET).
1498 If someone writes an application with libcurl and openssl who wants to
1499 enable the feature, one can do this in the SSL callback.
1501 Sharad Gupta brought this to my attention. Peter Sylvester helped me decide
1502 on the proper action.
1504 - Alexey Borzov filed bug report #2535504
1505 (http://curl.haxx.se/bug/view.cgi?id=2535504) pointing out that realms with
1506 quoted quotation marks in HTTP Digest headers didn't work. I've now added
1507 test case 1095 that verifies my fix.
1509 - Craig A West brought CURLOPT_NOPROXY and the corresponding --noproxy option.
1510 They basically offer the same thing the NO_PROXY environment variable only
1511 offered previously: list a set of host names that shall not use the proxy
1512 even if one is specified.
1514 Daniel Fandrich (20 Jan 2009)
1515 - Call setlocale() for libtest tests to test the effects of locale-induced
1516 libc changes on libcurl.
1518 - Fixed a couple more locale-dependent toupper conversions, mainly for
1519 clarity. This does fix one problem that causes ;type=i FTP URLs
1520 to fail in the Turkish locale when CURLOPT_PROXY_TRANSFER_MODE is
1521 used (test case 561)
1523 - Added tests 561 and 1091 through 1094 to test various combinations
1524 of ;type= and ;mode= URLs that could potentially fail in the Turkish
1527 Daniel Stenberg (20 Jan 2009)
1528 - Lisa Xu pointed out that the ssh.obj file was missing from the
1529 lib/Makefile.vc6 file (and thus from the vc8 and vc9 ones too).
1531 Version 7.19.3 (19 January 2009)
1533 Daniel Stenberg (16 Jan 2009)
1534 - Andrew de los Reyes fixed curlbuild.h for "generic" gcc builds on PPC, both
1537 Daniel Stenberg (15 Jan 2009)
1538 - Tim Ansell fixed a compiler warning in lib/cookie.c
1540 Daniel Stenberg (14 Jan 2009)
1541 - Grant Erickson fixed timeouts for TFTP such that specifying a
1542 connect-timeout, a max-time or both options work correctly and as expected
1543 by passing the correct boolean value to Curl_timeleft via the
1544 'duringconnect' parameter.
1546 With this small change, curl TFTP now behaves as expected (and likely as
1547 originally-designed):
1549 1) For non-existent or unreachable dotted IP addresses:
1551 a) With no options, follows the default curl 300s timeout...
1552 b) With --connect-timeout only, follows that value...
1553 c) With --max-time only, follows that value...
1554 d) With both --connect-timeout and --max-time, follows the smaller value...
1556 and times out with a "curl: (7) Couldn't connect to server" error.
1558 2) For transfers to/from a valid host:
1560 a) With no options, follows default curl 300s timeout for the
1561 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1562 timeout for the remainder of the transfer...
1564 b) With --connect-time only, follows that value for the
1565 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1566 timeout for the remainder of the transfer...
1568 c) With --max-time only, follows that value for the first
1569 XRQ/DATA/ACK transaction and for the remainder of the
1572 d) With both --connect-timeout and --max-time, follows the former
1573 for the first XRQ/DATA/ACK transaction and the latter for the
1574 remainder of the transfer...
1576 and times out with a "curl: (28) Timeout was reached" error as
1579 Daniel Stenberg (13 Jan 2009)
1580 - Michael Wallner fixed a NULL pointer deref when calling
1581 curl_easy_setup(curl, CURLOPT_COOKIELIST, "SESS") on a CURL handle with no
1584 - Stefan Teleman brought a patch to fix the default curlbuild.h file for the
1587 Daniel Stenberg (12 Jan 2009)
1588 - Based on bug report #2498665 (http://curl.haxx.se/bug/view.cgi?id=2498665)
1589 by Daniel Black, I've now added magic to the configure script that makes it
1590 use pkg-config to detect gnutls details as well if the existing method
1591 (using libgnutls-config) fails. While doing this, I cleaned up and unified
1592 the pkg-config usage when detecting openssl and nss as well.
1594 Daniel Stenberg (11 Jan 2009)
1595 - Karl Moerder brought the patch that creates vc9 Makefiles, and I made
1596 'maketgz' now use the actual makefile targets to do the VC8 and VC9
1599 Daniel Stenberg (10 Jan 2009)
1600 - Emil Romanus fixed:
1602 When using the multi interface over HTTP and the server returns a Location
1603 header, the running easy handle will get stuck in the CURLM_STATE_PERFORM
1604 state, leaving the external event loop stuck waiting for data from the
1605 ingoing socket (when using the curl_multi_socket_action stuff). While this
1606 bug was pretty hard to find, it seems to require only a one-line fix. The
1607 break statement on line 1374 in multi.c caused the function to skip the call
1610 How to reproduce this bug? Well, that's another question. evhiperfifo.c in
1611 the examples directory chokes on this bug only _sometimes_, probably
1612 depending on how fast the URLs are added. One way of testing the bug out is
1613 writing to hiper.fifo from more than one source at the same time.
1615 Daniel Fandrich (7 Jan 2009)
1616 - Unified much of the SessionHandle initialization done in Curl_open() and
1617 curl_easy_reset() by creating Curl_init_userdefined(). This had the side
1618 effect of fixing curl_easy_reset() so it now also resets
1619 CURLOPT_FTP_FILEMETHOD and CURLOPT_SSL_SESSIONID_CACHE
1621 Daniel Stenberg (7 Jan 2009)
1622 - Rob Crittenden did once again provide an NSS update:
1624 I have to jump through a few hoops now with the NSS library initialization
1625 since another part of an application may have already initialized NSS by the
1626 time Curl gets invoked. This patch is more careful to only shutdown the NSS
1627 library if Curl did the initialization.
1629 It also adds in a bit of code to set the default ciphers if the app that
1630 call NSS_Init* did not call NSS_SetDomesticPolicy() or set specific
1631 ciphers. One might argue that this lets other application developers get
1632 lazy and/or they aren't using the NSS API correctly, and you'd be right.
1633 But still, this will avoid terribly difficult-to-trace crashes and is
1636 Daniel Stenberg (1 Jan 2009)
1637 - 'reconf' is removed since we rather have users use 'buildconf'
1639 Daniel Stenberg (31 Dec 2008)
1640 - Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
1641 out that 'reconf' didn't properly point out the m4 subdirectory when running
1644 Daniel Stenberg (29 Dec 2008)
1645 - Phil Lisiecki filed bug report #2413067
1646 (http://curl.haxx.se/bug/view.cgi?id=2413067) that identified a problem that
1647 would cause libcurl to mark a DNS cache entry "in use" eternally if the
1648 subsequence TCP connect failed. It would thus never get pruned and refreshed
1649 as it should've been.
1651 Phil provided his own patch to this problem that while it seemed to work
1652 wasn't complete and thus I wrote my own fix to the problem.
1654 Daniel Stenberg (28 Dec 2008)
1655 - Peter Korsgaard fixed building libcurl with "configure --with-ssl
1658 - Anthony Bryan fixed more language and spelling flaws in man pages.
1660 Daniel Stenberg (22 Dec 2008)
1661 - Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
1662 on file indexes beyond 2 or 4GB.
1664 - Anthony Bryan provided a set of patches that cleaned up manual language,
1665 corrected spellings and more.
1667 Daniel Stenberg (20 Dec 2008)
1668 - Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
1669 pipelining, as libcurl could then easily get confused and A) work on the
1670 handle that was not "first in queue" on a pipeline, or even B) tell the app
1671 to REMOVE a socket while it was in use by a second handle in a pipeline. Both
1672 errors caused hanging or stalling applications.
1674 Daniel Stenberg (19 Dec 2008)
1675 - curl_multi_timeout() could return a timeout value of 0 even though nothing
1676 was actually ready to get done, as the internal time resolution is higher
1677 than the returned millisecond timer. Therefore it could cause applications
1678 running on fast processors to do short bursts of busy-loops.
1679 curl_multi_timeout() will now only return 0 if the timeout is actually
1682 - Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
1683 now has an improved ability to do right when the multi interface (both
1684 "regular" and multi_socket) is used for SCP and SFTP transfers. This should
1685 result in (much) less busy-loop situations and thus less CPU usage with no
1688 Daniel Stenberg (17 Dec 2008)
1689 - SCP and SFTP with the multi interface had the same flaw: the 'DONE'
1690 operation didn't complete properly if the EAGAIN equivalent was returned but
1691 libcurl would simply continue with a half-completed close operation
1692 performed. This ruined persistent connection re-use and cause some
1693 SSH-protocol errors in general. The correction is unfortunately adding a
1694 blocking function - doing it entirely non-blocking should be considered for
1697 Gisle Vanem (16 Dec 2008)
1698 - Added the possibility to use the Watt-32 tcp/ip stack under Windows.
1699 The change simply involved adding a USE_WATT32 section in the
1700 config-win32.h files (under ./lib and ./src). This section disables
1701 the use of any Winsock headers.
1703 Daniel Stenberg (16 Dec 2008)
1704 - libssh2_sftp_last_error() was wrongly used at some places in libcurl which
1705 made libcurl sometimes not properly abort problematic SFTP transfers.
1707 Daniel Stenberg (12 Dec 2008)
1708 - More work with Igor Novoseltsev to first fix the remaining stuff for
1709 removing easy handles from multi handles when the easy handle is/was within
1710 a HTTP pipeline. His bug report #2351653
1711 (http://curl.haxx.se/bug/view.cgi?id=2351653) was also related and was
1712 eventually fixed by a patch by Igor himself.
1714 Yang Tse (12 Dec 2008)
1715 - Patrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
1716 OS/400 compilations with IPv6 enabled.
1718 Daniel Stenberg (12 Dec 2008)
1719 - Mark Karpeles filed bug report #2416182 titled "crash in ConnectionExists
1720 when using duphandle+curl_mutli"
1721 (http://curl.haxx.se/bug/view.cgi?id=2416182) which showed that
1722 curl_easy_duphandle() wrongly also copied the pointer to the connection
1723 cache, which was plain wrong and caused a segfault if the handle would be
1724 used in a different multi handle than the handle it was duplicated from.
1726 Daniel Stenberg (11 Dec 2008)
1727 - Keshav Krity found out that libcurl failed to deal with dotted IPv6
1728 addresses if they were very long (>39 letters) due to a too strict address
1729 validity parser. It now accepts addresses up to 45 bytes long.
1731 Daniel Stenberg (11 Dec 2008)
1732 - Internet Explorer had a broken HTTP digest authentication before v7 and
1733 there are servers "out there" that relies on the client doing this broken
1734 Digest authentication. Apache even comes with an option to work with such
1737 The difference is only for URLs that contain a query-part (a '?'-letter and
1738 text to the right of it).
1740 libcurl now supports this quirk, and you enable it by setting the
1741 CURLAUTH_DIGEST_IE bit in the bitmask you pass to the CURLOPT_HTTPAUTH or
1742 CURLOPT_PROXYAUTH options. They are thus individually controlled to server
1745 (note that there's no way to activate this with the curl tool yet)
1747 Daniel Fandrich (9 Dec 2008)
1748 - Added test cases 1089 and 1090 to test --write-out after a redirect to
1749 test a report that the size didn't work, but these test cases pass.
1751 - Documented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.
1753 Daniel Stenberg (9 Dec 2008)
1754 - Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
1755 particular state for the control connection like it did before for implicit
1756 FTPS (libcurl assumed such control connections to be encrypted while some
1757 FTPS servers such as FileZilla assumes such connections to be clear
1758 mode). Use the CURLOPT_USE_SSL option to set your desired level.
1760 Daniel Stenberg (8 Dec 2008)
1761 - Fred Machado posted about a weird FTP problem on the curl-users list and when
1762 researching it, it turned out he got a 550 response back from a SIZE command
1763 and then I fell over the text in RFC3659 that says:
1765 The presence of the 550 error response to a SIZE command MUST NOT be taken
1766 by the client as an indication that the file cannot be transferred in the
1767 current MODE and TYPE.
1769 In other words: the change I did on September 30th 2008 and that has been
1770 included in the last two releases were a regression and a bad idea. We MUST
1771 NOT take a 550 response from SIZE as a hint that the file doesn't exist.
1773 - Christian Krause filed bug #2221237
1774 (http://curl.haxx.se/bug/view.cgi?id=2221237) that identified an infinite
1775 loop during GSS authentication given some specific conditions. With his
1776 patience and great feedback I managed to narrow down the problem and
1777 eventually fix it although I can't test any of this myself!
1779 Daniel Fandrich (3 Dec 2008)
1780 - Fixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
1781 support (e.g. Minix)
1783 Daniel Stenberg (3 Dec 2008)
1784 - Igor Novoseltsev filed bug #2351645
1785 (http://curl.haxx.se/bug/view.cgi?id=2351645) that identified a problem with
1786 the multi interface that occured if you removed an easy handle while in
1787 progress and the handle was used in a HTTP pipeline.
1789 - Pawel Kierski pointed out a mistake in the cookie code that could lead to a
1790 bad fclose() after a fatal error had occured.
1791 (http://curl.haxx.se/bug/view.cgi?id=2382219)
1793 Daniel Fandrich (25 Nov 2008)
1794 - If a HTTP request is Basic and num is already >=1000, the HTTP test
1795 server adds 1 to num to get the data section to return. This allows
1796 testing authentication negotiations using the Basic authentication
1799 - Added tests 1087 and 1088 to test Basic authentication on a redirect
1800 with and without --location-trusted
1802 Daniel Stenberg (24 Nov 2008)
1803 - Based on a patch by Vlad Grachov, libcurl now uses a new libssh2 0.19
1804 function when built to support SCP and SFTP that helps the library to know
1805 in which direction a particular libssh2 operation would return EAGAIN so
1806 that libcurl knows what socket conditions to wait for before trying the
1807 function call again. Previously (and still when using libssh2 0.18 or
1808 earlier), libcurl will busy-loop in this situation when the easy interface
1811 Daniel Fandrich (20 Nov 2008)
1812 - Automatically detect OpenBSD's CA cert bundle.
1814 Daniel Stenberg (19 Nov 2008)
1815 - I removed the default use of "Pragma: no-cache" from libcurl when a proxy is
1816 used. It has been used since forever but it was never a good idea to use
1817 unless explicitly asked for.
1819 - Josef Wolf's extension that allows a $TESTDIR/gdbinit$testnum file that when
1820 you use runtests.pl -g, will be sourced by gdb to allow additional fancy or
1821 whatever you see fit
1823 - Christian Krause reported and fixed a memory leak that would occur with HTTP
1824 GSS/kerberos authentication (http://curl.haxx.se/bug/view.cgi?id=2284386)
1826 - Andreas Wurf and Markus Koetter helped me analyze a problem that Andreas got
1827 when uploading files to a single FTP server using multiple easy handle
1828 handles with the multi interface. Occasionally a handle would stall in
1831 The problem turned out to be a side-effect of the ConnectionExists()
1832 function's eagerness to re-use a handle for HTTP pipelining so it would
1833 select it even if already being in use, due to an inadequate check for its
1834 chances of being used for pipelnining.
1836 Daniel Fandrich (17 Nov 2008)
1837 - Added more compiler warning options for gcc 4.3
1839 Yang Tse (17 Nov 2008)
1840 - Fix a remaining problem in the inet_pton() runtime configure check. And
1841 fix internal Curl_inet_pton() failures to reject certain malformed literals.
1843 - Make configure script check if ioctl with the SIOCGIFADDR command can be
1844 used, and define HAVE_IOCTL_SIOCGIFADDR if appropriate.
1846 Daniel Stenberg (16 Nov 2008)
1847 - Christian Krause fixed a build failure when building with gss support
1848 enabled and FTP disabled.
1850 - Added check for NULL returns from strdup() in src/main.c and lib/formdata.c
1851 - reported by Jim Meyering also prevent buffer overflow on MSDOS when you do
1852 for example -O on a url with a file name part longer than PATH_MAX letters
1854 - lib/nss.c fixes based on the report by Jim Meyering: I went over and added
1855 checks for return codes for all calls to malloc and strdup that were
1856 missing. I also changed a few malloc(13) to use arrays on the stack and a
1857 few malloc(PATH_MAX) to instead use aprintf() to lower memory use.
1859 - I fixed a memory leak in Curl_nss_connect() when CURLOPT_ISSUERCERT is
1862 Daniel Fandrich (14 Nov 2008)
1863 - Added .xml as one of the few common file extensions known by the multipart
1866 - Added some #ifdefs around header files and change the EAGAIN test to
1867 fix compilation on Cell (reported by Jeff Curley).
1869 Yang Tse (14 Nov 2008)
1870 - Fixed several configure script issues affecting checks for inet_ntoa_r(),
1871 inet_ntop(), inet_pton(), getifaddrs(), fcntl() and getaddrinfo().
1873 Yang Tse (13 Nov 2008)
1874 - Refactored configure script detection of functions used to set sockets into
1875 non-blocking mode, and decouple function detection from function capability.