5 \___|\___/|_| \_\_____|
9 Daniel Stenberg (29 Apr 2010)
10 - Ben Greear made telnet a lot better/easier to use by an application:
12 The main change is to allow input from user-specified methods, when they are
13 specified with CURLOPT_READFUNCTION. All calls to fflush(stdout) in
14 telnet.c were removed, which makes using 'curl telnet://foo.com' painful
15 since prompts and other data are not always returned to the user promptly.
16 Use 'curl --no-buffer telnet://foo.com' instead. In general, the user
17 should have their CURLOPT_WRITEFUNCTION do a fflush for interactive use.
19 Also fix assumption that reading from stdin never returns < 0.
20 Old code could crash in that case.
22 Call progress functions in telnet main loop.
24 Daniel Stenberg (26 Apr 2010)
25 - Make use of the libssh2_init/exit functions that libssh2 added in version
26 1.2.5. Using them will improve how libcurl works in threaded situations when
27 SCP and SFTP are transfered.
29 Daniel Stenberg (25 Apr 2010)
30 - Based on work by Kamil Dudka, I've introduced the new configure option
31 --enable-threaded-resolver. When used, the configure script will check for
32 pthreads and if around, it will build libcurl to use pthreads to do name
33 resolving in a threaded manner. Note that this is just a fix to offer an
34 option that can enable the code that already included. The threader resolver
35 code was mostly added on Jan 26 2010.
37 Daniel Stenberg (24 Apr 2010)
38 - Alex Bligh introduced the --proto and -proto-redir options that limit what
39 protocols curl accepts for the requests and when following redirects.
41 Kamil Dudka (24 Apr 2010)
42 - Fixed test536 in order to not fail with threaded DNS resolver and tweaked
43 comments in certain examples using curl_multi_fdset().
45 - Fixed SSL handshake timeout underflow in libcurl-NSS, which caused test405
46 to hang on a slow machine.
48 Daniel Stenberg (21 Apr 2010)
49 - The -O option caused curl to crash on windows and DOS due to the tool
50 writing out of boundary memory.
52 Yang Tse (20 Apr 2010)
53 - Ruslan Gazizov detected that MSVC makefiles were using wsock32.lib instead
54 of ws2_32.lib, this generated linking issues on MSVC IPv6 enabled builds
55 that were done using those makefiles.
57 Daniel Stenberg (19 Apr 2010)
58 - -J/--remote-header-name didn't strip trailing carriage returns or linefeeds
59 properly, so they could be used in the file name.
61 Daniel Stenberg (16 Apr 2010)
62 - Jerome Vouillon made the GnuTLS SSL handshake phase non-blocking.
64 - The recent overhaul of the SSL recv function made the GnuTLS specific code
65 treat a zero returned from gnutls_record_recv() as an error, and this caused
66 our HTTPS test cases to fail. We leave it to upper layer code to detect if
67 an EOF is a problem or not.
69 - I reverted the resolver fix from yesterday and instead removed all uses of
70 AI_CANONNAME all over libcurl and made the only user of that info (krb5.c)
71 use the host name from the URL instead. No reverse resolving is a good
74 - Paul Howarth made configure properly detect GSS "on ancient Linux distros"
75 by editing in which order we use headers to detect GSS.
77 Daniel Stenberg (15 Apr 2010)
78 - Rainer Canavan filed bug report #2987196 that identified libcurl doing
79 unnecesary reverse name lookups in many cases when built to use IPv4 and
80 getaddrinfo(). The logic for ipv6 is now used for ipv4 too.
82 (http://curl.haxx.se/bug/view.cgi?id=2963679)
84 Version 7.20.1 (14 April 2010)
86 Daniel Stenberg (9 Apr 2010)
87 - Prefixing the FTP quote commands with an asterisk really only worked for the
88 postquote actions. This is now fixed and test case 227 has been extended to
91 Kamil Dudka (4 Apr 2010)
92 - Eliminated a race condition in Curl_resolv_timeout().
94 - Refactorized interface of Curl_ssl_recv()/Curl_ssl_send().
96 - libcurl-NSS now provides more accurate messages and error codes in case of
97 client certificate problem. Either during connection, or transfer phase.
99 Daniel Stenberg (1 Apr 2010)
100 - Matt Wixson found and fixed a bug in the SCP/SFTP area where the code
101 treated a 0 return code from libssh2 to be the same as EAGAIN while in
102 reality it isn't. The problem caused a hang in SFTP transfers from a
105 Daniel Stenberg (28 Mar 2010)
106 - Ben Greear: If you pass a URL to pop3 that does not contain a message ID as
107 part of the URL, it would previously ask for 'INBOX' which just causes the
108 pop3 server to return an error.
110 Now libcurl treats en empty message ID as a request for LIST (list of pop3
111 message IDs). User's code could then parse this and download individual
114 Daniel Stenberg (27 Mar 2010)
115 - Ben Greear brought a patch that from now on allows all protocols to specify
116 name and user within the URL, in the same manner HTTP and FTP have been
117 allowed to in the past - although far from all of the libcurl supported
118 protocls actually have that feature in their URL definition spec.
120 Daniel Stenberg (26 Mar 2010)
121 - Ben Greear brought code that makes the rate limiting code for the easy
122 interface a bit smoother as it introduces sub-second sleeps during it and it
123 also takes the buffer sizes into account.
125 Daniel Stenberg (24 Mar 2010)
126 - Bob Richmond: There's an annoying situation where libcurl will read new HTTP
127 response data from a socket, then check if it's a timeout if one is set. If
128 the last packet received constitutes the end of the response body, libcurl
129 still treats it as a timeout condition and reports a message like:
131 "Operation timed out after 3000 milliseconds with 876 out of 876 bytes
134 It should only a timeout if the timer lapsed and we DIDN'T receive the end
135 of the response body yet.
137 - Christopher Conroy fixed a problem with RTSP and GET_PARAMETER reported
138 to us by Massimo Callegari. There's a new test case 572 that verifies this
141 - The 'ares' subtree has been removed from the source repository. It was
142 always a separate project that sort of piggybacked on the curl project since
143 the dawn of times and now the time has come for it to go stand on its own
144 legs and continue living its own life. All details on c-ares and its new
145 source code repository is found at http://c-ares.haxx.se/
147 Daniel Stenberg (23 Mar 2010)
148 - Kenny To filed the bug report #2963679 with patch to fix a problem he
149 experienced with doing multi interface HTTP POST over a proxy using
150 PROXYTUNNEL. He found a case where it would connect fine but bits.tcpconnect
151 was not set correct so libcurl didn't work properly.
153 (http://curl.haxx.se/bug/view.cgi?id=2963679)
155 - Akos Pasztory filed debian bug report #572276
156 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=572276 mentioning a problem
157 with a resource that returns chunked-encoded _and_ with a Content-Length
158 and libcurl failed to properly ignore the latter information.
160 - Hauke Duden provided an example program that made the multi interface crash.
161 His example simply used the multi interface and did first one FTP transfer
162 and after completion it used a second easy handle and did another FTP
163 transfer on the same FTP server.
165 This triggered a bug in the "delayed easy handle kill" system that curl
166 uses: when an FTP connection is left alive it must keep an easy handle
167 around internally - only for the purpose of having an easy handle when it
168 later disconnects it. The code assumed that when the easy handle was removed
169 and an internal reference was made, that version could be killed later on
170 when a new easy handle came using the same connection. This was wrong as
171 Hauke's example showed that the removed handle wasn't killed for real until
172 later. This caused a double close attempt => segfault.
174 Daniel Stenberg (22 Mar 2010)
175 - Thomas Lopatic fixed the alarm()-based DNS timeout:
177 Looking at the code of Curl_resolv_timeout() in hostip.c, I think that in
178 case of a timeout, the signal handler for SIGALRM never gets removed. I
179 think that in my case it gets executed at some point later on when execution
180 has long left Curl_resolv_timeout() or even the cURL library.
182 The code that is jumped to with siglongjmp() simply sets the error message
183 to "name lookup timed out" and then returns with CURLRESOLV_ERROR. I guess
184 that instead of simply returning without cleaning up, the code should have a
185 goto that jumps to the spot right after the call to Curl_resolv().
187 Kamil Dudka (22 Mar 2010)
188 - Douglas Steinwand contributed a patch fixing insufficient initialization in
189 Curl_clone_ssl_config()
191 Daniel Stenberg (21 Mar 2010)
192 - Ben Greear improved TFTP: the error code returning and the treatment
193 of TSIZE == 0 when uploading.
195 - We've switched from CVS to git. See http://curl.haxx.se/source.html
197 Kamil Dudka (19 Mar 2010)
198 - Improved Curl_read() to not ignore the error returned from Curl_ssl_recv().
200 Daniel Stenberg (15 Mar 2010)
201 - Constantine Sapuntzakis brought a patch:
203 The problem mentioned on Dec 10 2009
204 (http://curl.haxx.se/bug/view.cgi?id=2905220) was only partially fixed.
205 Partially because an easy handle can be associated with many connections in
206 the cache (e.g. if there is a redirect during the lifetime of the easy
207 handle). The previous patch only cleaned up the first one. The new fix now
208 removes the easy handle from all connections, not just the first one.
210 Daniel Stenberg (6 Mar 2010)
211 - Ben Greear brought a patch that fixed the rate limiting logic for TFTP when
212 the easy interface was used.
214 Daniel Stenberg (5 Mar 2010)
215 - Daniel Johnson provided fixes for building curl with the clang compiler.
217 Yang Tse (5 Mar 2010)
218 - Constantine Sapuntzakis detected and fixed a double free in builds done
219 with threaded resolver enabled (Windows default configuration) that would
220 get triggered when a curl handle is closed while doing DNS resolution.
222 Daniel Stenberg (2 Mar 2010)
223 - [Daniel Johnson] I've been trying to build libcurl with clang on Darwin and
224 ran into some issues with the GSSAPI tests in configure.ac. The tests first
225 try to determine the include dirs and libs and set CPPFLAGS and LIBS
226 accordingly. It then checks for the headers and finally sets LIBS a second
227 time, causing the libs to be included twice. The first setting of LIBS seems
228 redundant and should be left out, since the first part is otherwise just
229 about finding headers.
231 My second issue is that 'krb5-config --libs gssapi' on Darwin is less than
232 useless and returns junk that, while it happens to work with gcc, causes
233 clang to choke. For example, --libs returns $CFLAGS along with the libs,
234 which is really retarded. Simply setting 'LIBS="$LIBS -lgssapi_krb5
235 -lresolv"' on Darwin is sufficient.
237 - Based on patch provided by Jacob Moshenko, the transfer logic now properly
238 makes sure that when using sub-second timeouts, there's no final bad 1000ms
239 wait. Previously, a sub-second timeout would often make the elapsed time end
240 up the time rounded up to the nearest second (e.g. 1s for 200ms timeout)
242 - Andrei Benea filed bug report #2956698 and pointed out that the
243 CURLOPT_CERTINFO feature leaked memory due to a missing OpenSSL function
244 call. He provided the patch to fix it too.
246 http://curl.haxx.se/bug/view.cgi?id=2956698
248 - Markus Duft pointed out in bug #2961796 that even though Interix has a
249 poll() function it doesn't quite work the way we want it so we must disable
250 it, and he also provided a patch for it.
252 http://curl.haxx.se/bug/view.cgi?id=2961796
254 - Made the pingpong timeout code properly deal with the response timeout AND
255 the global timeout if set. Also, as was reported in the bug report #2956437
256 by Ryan Chan, the time stamp to use as basis for the per command timeout was
257 not set properly in the DONE phase for FTP (and not for SMTP) so I fixed
258 that just now. This was a regression compared to 7.19.7 due to the
259 conversion of FTP code over to the generic pingpong concepts.
261 http://curl.haxx.se/bug/view.cgi?id=2956437
263 Daniel Stenberg (1 Mar 2010)
264 - Ben Greear provided an update for TFTP that fixes upload.
266 - Wesley Miaw reported bug #2958179 which identified a case of looping during
267 OpenSSL based SSL handshaking even though the multi interface was used and
268 there was no good reason for it.
270 http://curl.haxx.se/bug/view.cgi?id=2958179
272 Daniel Stenberg (26 Feb 2010)
273 - Pat Ray in bug #2958474 pointed out an off-by-one case when receiving a
274 chunked-encoding trailer.
276 http://curl.haxx.se/bug/view.cgi?id=2958474
278 Daniel Fandrich (25 Feb 2010)
279 - Fixed a couple of out of memory leaks and a segfault in the SMTP & IMAP code.
281 Yang Tse (25 Feb 2010)
282 - I fixed bug report #2958074 indicating
283 (http://curl.haxx.se/bug/view.cgi?id=2958074) that curl on Windows with
284 option --trace-time did not use local time when timestamping trace lines.
285 This could also happen on other systems depending on time souurce.
287 Patrick Monnerat (22 Feb 2010)
288 - Proper handling of STARTTLS on SMTP, taking CURLUSESSL_TRY into account.
289 - SMTP falls back to RFC821 HELO when EHLO fails (and SSL is not required).
290 - Use of true local host name (i.e.: via gethostname()) when available, as
291 default argument to SMTP HELO/EHLO.
292 - Test case 804 for HELO fallback.
294 Daniel Stenberg (20 Feb 2010)
295 - Fixed the SMTP compliance by making sure RCPT TO addresses are specified
296 properly in angle brackets. Recipients provided with CURLOPT_MAIL_RCPT now
297 get angle bracket wrapping automatically by libcurl unless the recipient
298 starts with an angle bracket as then the app is assumed to deal with that
301 - I made the SMTP code expect a 250 response back from the server after the
302 full DATA has been sent, and I modified the test SMTP server to also send
303 that response. As usual, the DONE operation that is made after a completed
304 transfer is still not doable in a non-blocking way so this waiting for 250
305 is unfortunately made blockingly.
307 Yang Tse (14 Feb 2010)
308 - Overhauled test suite getpart() function. Fixing potential out of bounds
309 stack and memory overwrites triggered with huge test case definitions.
311 Daniel Stenberg (13 Feb 2010)
312 - Martin Hager reported and fixed a problem with a missing quote in libcurl.m4
314 (http://curl.haxx.se/bug/view.cgi?id=2951319)
316 - Tom Donovan fixed the CURL_FORMAT_* defines when building with cmake.
318 (http://curl.haxx.se/bug/view.cgi?id=2951269)
320 Daniel Stenberg (12 Feb 2010)
321 - Jack Zhang reported a problem with SMTP: we wrongly used multiple addresses
322 in the same RCPT TO line, when they should be sent in separate single
323 commands. I updated test case 802 to verify this.
325 - I also fixed a bad use of my_setopt_str() of CURLOPT_MAIL_RCPT in the curl
326 tool which made it try to output it as string for the --libcurl feature
327 which could lead to crashes.
329 Yang Tse (11 Feb 2010)
330 - Steven M. Schweda fixed VMS builder bad behavior when used in a batch job,
331 removed obsolete batch_compile.com and defines.com and updated VMS readme.
333 Version 7.20.0 (9 February 2010)
335 Daniel Stenberg (9 Feb 2010)
336 - When downloading compressed content over HTTP and the app asked libcurl to
337 automatically uncompress it with the CURLOPT_ENCODING option, libcurl could
338 wrongly provide the callback with more data than the maximum documented
339 amount. An application could thus get tricked into badness if the maximum
340 limit was trusted to be enforced by libcurl itself (as it is documented).
342 This is further detailed and explained in the libcurl security advisory
345 http://curl.haxx.se/docs/adv_20100209.html
347 Daniel Fandrich (3 Feb 2010)
348 - Changed the Watcom makefiles to make them easier to keep in sync with
349 Makefile.inc since that can't be included directly.
351 Yang Tse (2 Feb 2010)
352 - Symbol CURL_FORMAT_OFF_T now obsoleted, will be removed in a future release,
353 symbol will not be available when building with CURL_NO_OLDIES defined. Use
354 of CURL_FORMAT_CURL_OFF_T is preferred since 7.19.0
356 Daniel Stenberg (1 Feb 2010)
357 - Using the multi_socket API, it turns out at times it seemed to "forget"
358 connections (which caused a hang). It turned out to be an existing (7.19.7)
359 bug in libcurl (that's been around for a long time) and it happened like
362 The app calls curl_multi_add_handle() to add a new easy handle, libcurl will
363 then set it to timeout in 1 millisecond so libcurl will tell the app about
366 The app's timeout fires off that there's a timeout, the app calls libcurl as
367 we so often document it:
370 res = curl_multi_socket_action(... TIMEOUT ...);
371 } while(CURLM_CALL_MULTI_PERFORM == res);
373 And this is the problem number one:
375 When curl_multi_socket_action() is called with no specific handle, but only
376 a timeout-action, it will *only* perform actions within libcurl that are
377 marked to run at this time. In this case, the request would go from INIT to
378 CONNECT and return CURLM_CALL_MULTI_PERFORM. When the app then calls libcurl
379 again, there's no timer set for this handle so it remains in the CONNECT
380 state. The CONNECT state is a transitional state in libcurl so it reports no
381 sockets there, and thus libcurl never tells the app anything more about that
382 easy handle/connection.
384 libcurl _does_ set a 1ms timeout for the handle at the end of
385 multi_runsingle() if it returns CURLM_CALL_MULTI_PERFORM, but since the loop
386 is instant the new job is not ready to run at that point (and there's no
387 code that makes libcurl call the app to update the timout for this new
388 timeout). It will simply rely on that some other timeout will trigger later
389 on or that something else will update the timeout callback. This makes the
390 bug fairly hard to repeat.
392 The fix made to adress this issue:
394 We introduce a loop in lib/multi.c around all calls to multi_runsingle() and
395 simply check for CURLM_CALL_MULTI_PERFORM internally. This has the added
396 benefit that this goes in line with my long-term wishes to get rid of the
397 CURLM_CALL_MULTI_PERFORM all together from the public API.
399 The downside of this fix, is that the counter we return in 'running_handles'
400 in several of our public functions then gets a slightly new and possibly
401 confusing behavior during times:
403 If an app adds a handle that fails to connect (very quickly) it may just
404 as well never appear as a 'running_handle' with this fix. Previously it
405 would first bump the counter only to get it decreased again at next call.
406 Even I have used that change in handle counter to signal "end of a
407 transfer". The only *good* way to find the end of a individual transfer
408 is calling curl_multi_info_read() to see if it returns one.
410 Of course, if the app previously did the looping before it checked the
411 counter, it really shouldn't be any new effect.
413 Yang Tse (26 Jan 2010)
414 - Constantine Sapuntzakis' and Joshua Kwan's work done in the last four months
415 relative to the asynchronous DNS lookups, along with with some integration
416 adjustments I have done are finally committed to CVS.
418 Currently these enhancements will benefit builds done using c-ares on any
419 platform as well as Windows builds using the default threaded resolver.
421 This release does not make generally available POSIX threaded DNS lookups
422 yet. There is no configure option to enable this feature yet. It is possible
423 to experimantally try this feature running configure with compiler flags that
424 make simultaneous definition of preprocessor symbols USE_THREADS_POSIX and
425 HAVE_PTHREAD_H, as well as whatever reentrancy compiler flags and linker ones
426 are required to link and properly use pthread_* functions on each platform.
428 Daniel Stenberg (26 Jan 2010)
429 - Mike Crowe made libcurl return CURLE_COULDNT_RESOLVE_PROXY when it is the
430 proxy that cannot be resolved when using c-ares. This matches the behaviour
431 when not using c-ares.
433 Björn Stenberg (23 Jan 2010)
434 - Added a new flag: -J/--remote-header-name. This option tells the
435 -O/--remote-name option to use the server-specified Content-Disposition
436 filename instead of extracting a filename from the URL.
438 Daniel Stenberg (21 Jan 2010)
439 - Chris Conroy brought support for RTSP transfers, and with it comes 8(!) new
440 libcurl options for controlling what to get and how to receive posssibly
441 interleaved RTP data.
443 Daniel Stenberg (20 Jan 2010)
444 - As was pointed out on the http-state mailing list, the order of cookies in a
445 HTTP Cookie: header _needs_ to be sorted on the path length in the cases
446 where two cookies using the same name are set more than once using
447 (overlapping) paths. Realizing this, identically named cookies must be
448 sorted correctly. But detecting only identically named cookies and take care
449 of them individually is harder than just to blindly and unconditionally sort
450 all cookies based on their path lengths. All major browsers also already do
451 this, so this makes our behavior one step closer to them in the cookie area.
453 Test case 8 was the only one that broke due to this change and I updated it
456 Daniel Stenberg (19 Jan 2010)
457 - David McCreedy brought a fix and a new test case (129) to make libcurl work
458 again when downloading files over FTP using ASCII and it turns out that the
459 final size of the file is not the same as the initial size the server
460 reported. This is very common since servers don't take the newline
461 conversions into account.
463 Kamil Dudka (14 Jan 2010)
464 - Suppressed side effect of OpenSSL configure checks, which prevented NSS from
465 being properly detected under certain circumstances. It had been caused by
466 strange behavior of pkg-config when handling PKG_CONFIG_LIBDIR. pkg-config
467 distinguishes among empty and non-existent environment variable in that case.
469 Daniel Stenberg (12 Jan 2010)
470 - Gil Weber reported a peculiar flaw with the multi interface when doing SFTP
471 transfers: curl_multi_fdset() would return -1 and not set and file
472 descriptors several times during a transfer of a single file. It turned out
473 to be due to two different flaws now fixed. Gil's excellent recipe helped me
476 Daniel Stenberg (11 Jan 2010)
477 - Made sure that the progress callback is repeatedly called at a regular
478 interval even during very slow connects.
480 - The tests/runtests.pl script now checks to see if the test case that runs is
481 present in the tests/data/Makefile.am and outputs a notice message on the
482 screen if not. Each test file has to be included in that Makefile.am to get
483 included in release archives and forgetting to add files there is a common
484 mistake. This is an attempt to make it harder to forget.
486 Daniel Stenberg (9 Jan 2010)
487 - Johan van Selst found and fixed a OpenSSL session ref count leak:
489 ossl_connect_step3() increments an SSL session handle reference counter on
490 each call. When sessions are re-used this reference counter may be
491 incremented many times, but it will be decremented only once when done (by
492 Curl_ossl_session_free()); and the internal OpenSSL data will not be freed
493 if this reference count remains positive. When a session is re-used the
494 reference counter should be corrected by explicitly calling
495 SSL_SESSION_free() after each consecutive SSL_get1_session() to avoid
496 introducing a memory leak.
498 (http://curl.haxx.se/bug/view.cgi?id=2926284)
500 Daniel Stenberg (7 Jan 2010)
501 - Make sure the progress callback is called repeatedly even during very slow
502 name resolves when c-ares is used for resolving.
504 Claes Jakobsson (6 Jan 2010)
505 - Julien Chaffraix fixed so that the fragment part in an URL is not sent
506 to the server anymore.
508 Kamil Dudka (3 Jan 2010)
509 - Julien Chaffraix eliminated a duplicated initialization in singlesocket().
511 Daniel Stenberg (2 Jan 2010)
512 - Make curl support --ssl and --ssl-reqd instead of the previous FTP-specific
513 versions --ftp-ssl and --ftp-ssl-reqd as these options are now used to
514 control SSL/TLS for IMAP, POP3 and SMTP as well in addition to FTP. The old
515 option names are still working but the new ones are the ones listed and
518 Daniel Stenberg (1 Jan 2010)
519 - Ingmar Runge enhanced libcurl's FTP engine to support the PRET command. This
520 command is a special "hack" used by the drftpd server, but even though it is
521 a custom extension I've deemed it fine to add to libcurl since this server
522 seems to survive and people keep using it and want libcurl to support
523 it. The new libcurl option is named CURLOPT_FTP_USE_PRET, and it is also
524 usable from the curl tool with --ftp-pret. Using this option on a server
525 that doesn't support this command will make libcurl fail.
527 I added test cases 1107 and 1108 to verify the functionality.
529 The PRET command is documented at
530 http://www.drftpd.org/index.php/Distributed_PASV
532 Yang Tse (30 Dec 2009)
533 - Steven M. Schweda improved VMS build system, and Craig A. Berry helped
534 with the patch and testing.
536 Daniel Stenberg (26 Dec 2009)
537 - Renato Botelho and Peter Pentchev brought a patch that makes the libcurl
538 headers work correctly even on FreeBSD systems before v8.
540 (http://curl.haxx.se/bug/view.cgi?id=2916915)
542 Daniel Stenberg (17 Dec 2009)
543 - David Byron fixed Curl_ossl_cleanup to actually call ENGINE_cleanup when
546 - Follow-up fix for the proxy fix I did for Jon Nelson's bug. It turned out I
547 was a bit too quick and broke test case 1101 with that change. The order of
548 some of the setups is sensitive. I now changed it slightly again to make
549 sure we do them in this order:
551 1 - parse URL and figure out what protocol is used in the URL
552 2 - prepend protocol:// to URL if missing
553 3 - parse name+password off URL, which needs to know what protocol is used
554 (since only some allows for name+password in the URL)
555 4 - figure out if a proxy should be used set by an option
556 5 - if no proxy option, check proxy environment variables
557 6 - run the protocol-specific setup function, which needs to have the proxy
560 Daniel Stenberg (15 Dec 2009)
561 - Jon Nelson found a regression that turned out to be a flaw in how libcurl
562 detects and uses proxies based on the environment variables. If the proxy
563 was given as an explicit option it worked, but due to the setup order
564 mistake proxies would not be used fine for a few protocols when picked up
565 from '[protocol]_proxy'. Obviously this broke after 7.19.4. I now also added
566 test case 1106 that verifies this functionality.
568 (http://curl.haxx.se/bug/view.cgi?id=2913886)
570 Daniel Stenberg (12 Dec 2009)
571 - IMAP, POP3 and SMTP support and their TLS versions (including IMAPS, POP3S
572 and SMTPS) are now supported. The current state may not yet be solid, but
573 the foundation is in place and the test suite has some initial support for
574 these protocols. Work will now persue to make them nice libcurl citizens
577 The work with supporting these new protocols was sponsored by
578 networking4all.com - thanks!
580 Daniel Stenberg (10 Dec 2009)
581 - Siegfried Gyuricsko found out that the curl manual said --retry would retry
582 on FTP errors in the transient 5xx range. Transient FTP errors are in the
583 4xx range. The code itself only tried on 5xx errors that occured _at login_.
584 Now the retry code retries on all FTP transfer failures that ended with a
587 (http://curl.haxx.se/bug/view.cgi?id=2911279)
589 - Constantine Sapuntzakis figured out a case which would lead to libcurl
590 accessing alredy freed memory and thus crash when using HTTPS (with
591 OpenSSL), multi interface and the CURLOPT_DEBUGFUNCTION and a certain order
592 of cleaning things up. I fixed it.
594 (http://curl.haxx.se/bug/view.cgi?id=2905220)
596 Daniel Stenberg (7 Dec 2009)
597 - Martin Storsjo made libcurl use the Expect: 100-continue header for posts
598 with unknown size. Previously it was only used for posts with a known size
599 larger than 1024 bytes.
601 Daniel Stenberg (1 Dec 2009)
602 - If the Expect: 100-continue header has been set by the application through
603 curl_easy_setopt with CURLOPT_HTTPHEADER, the library should set
604 data->state.expect100header accordingly - the current code (in 7.19.7 at
605 least) doesn't handle this properly. Martin Storsjo provided the fix!
607 Yang Tse (28 Nov 2009)
608 - Added Diffie-Hellman parameters to several test harness certificate files in
609 PEM format. Required by several stunnel versions used by our test harness.
611 Daniel Stenberg (28 Nov 2009)
612 - Markus Koetter provided a polished and updated version of Chad Monroe's TFTP
613 rework patch that now integrates TFTP properly into libcurl so that it can
614 be used non-blocking with the multi interface and more. BLKSIZE also works.
616 The --tftp-blksize option was added to allow setting the TFTP BLKSIZE from
619 Daniel Stenberg (26 Nov 2009)
620 - Extended and fixed the change I did on Dec 11 for the the progress
621 meter/callback during FTP command/response sequences. It turned out it was
622 really lame before and now the progress meter SHOULD get called at least
625 Daniel Stenberg (23 Nov 2009)
626 - Bjorn Augustsson reported a bug which made curl not report any problems even
627 though it failed to write a very small download to disk (done in a single
628 fwrite call). It turned out to be because fwrite() returned success, but
629 there was insufficient error-checking for the fclose() call which tricked
630 curl to believe things were fine.
632 Yang Tse (23 Nov 2009)
633 - David Byron modified Makefile.dist vc8 and vc9 targets in order to allow
634 finer granularity control when generating src and lib makefiles.
636 Yang Tse (22 Nov 2009)
637 - I modified configure to force removal of the curlbuild.h file included in
638 distribution tarballs for use by non-configure systems. As intended, this
639 would get overwriten when doing in-tree builds. But VPATH builds would end
640 having two curlbuild.h files, one in the source tree and another in the
641 build tree. With the modification I introduced 5 Nov 2009 this could become
642 an issue when running libcurl's test suite.
644 Daniel Stenberg (20 Nov 2009)
645 - Constantine Sapuntzakis identified a write after close, as the sockets were
646 closed by libcurl before the SSL lib were shutdown and they may write to its
647 socket. Detected to at least happen with OpenSSL builds.
649 - Jad Chamcham pointed out a bug with connection re-use. If a connection had
650 CURLOPT_HTTPPROXYTUNNEL enabled over a proxy, a subsequent request using the
651 same proxy with the tunnel option disabled would still wrongly re-use that
652 previous connection and the outcome would only be badness.
654 Yang Tse (18 Nov 2009)
655 - I modified the memory tracking system to make it intolerant with zero sized
656 malloc(), calloc() and realloc() function calls.
658 Daniel Stenberg (17 Nov 2009)
659 - Constantine Sapuntzakis provided another fix for the DNS cache that could
660 end up with entries that wouldn't time-out:
662 1. Set up a first web server that redirects (307) to a http://server:port
664 2. Have curl connect to the first web server using curl multi
666 After the curl_easy_cleanup call, there will be curl dns entries hanging
667 around with in_use != 0.
669 (http://curl.haxx.se/bug/view.cgi?id=2891591)
671 - Marc Kleine-Budde fixed: curl saved the LDFLAGS set during configure into
672 its pkg-config file. So -Wl stuff ended up in the .pc file, which is really
673 bad, and breaks if there are multiple -Wl in our LDFLAGS (which are in
674 PTXdist). bug #2893592 (http://curl.haxx.se/bug/view.cgi?id=2893592)
676 Kamil Dudka (15 Nov 2009)
677 - David Byron improved the configure script to use pkg-config to find OpenSSL
678 (and in particular the list of required libraries) even if a path is given
679 as argument to --with-ssl
681 Yang Tse (15 Nov 2009)
682 - I removed enable-thread / disable-thread configure option. These were only
683 placebo options. The library is always built as thread safe as possible on
686 Claes Jakobsson (14 Nov 2009)
687 - curl-config now accepts '--configure' to see what arguments was
688 passed to the configure script when building curl.
690 Daniel Stenberg (14 Nov 2009)
691 - Claes Jakobsson restored the configure functionality to detect NSS when
692 --with-nss is set but not "yes".
694 I think we can still improve that to check for pkg-config in that path etc,
695 but at least this patch brings back the same functionality we had before.
697 - Camille Moncelier added support for the file type SSL_FILETYPE_ENGINE for
698 the client certificate. It also disable the key name test as some engines
699 can select a private key/cert automatically (When there is only one key
700 and/or certificate on the hardware device used by the engine)
702 Yang Tse (14 Nov 2009)
703 - Constantine Sapuntzakis provided the fix that ensures that an SSL connection
704 won't be reused unless protection level for peer and host verification match.
706 I refactored how preprocessor symbol _THREAD_SAFE definition is done.
708 Kamil Dudka (12 Nov 2009)
709 - Kevin Baughman provided a fix preventing libcurl-NSS from crash on doubly
710 closed NSPR descriptor. The issue was hard to find, reported several times
711 before and always closed unresolved. More info at the RH bug:
712 https://bugzilla.redhat.com/534176
714 - libcurl-NSS now tries to reconnect with TLS disabled in case it detects
715 a broken TLS server. However it does not happen if SSL version is selected
716 manually. The approach was originally taken from PSM. Kaspar Brand helped me
717 to complete the patch. Original bug reports:
718 https://bugzilla.redhat.com/525496
719 https://bugzilla.redhat.com/527771
721 Yang Tse (12 Nov 2009)
722 - I modified configure script to make the getaddrinfo function check also
723 verify if the function is thread safe.
725 Yang Tse (11 Nov 2009)
726 - Marco Maggi reported that compilation failed when configured --with-gssapi
727 and GNU GSS installed due to a missing mutual exclusion of header files in
728 the Kerberos 5 code path. He also verified that my patch worked for him.
730 Daniel Stenberg (11 Nov 2009)
731 - Constantine Sapuntzakis posted bug #2891595
732 (http://curl.haxx.se/bug/view.cgi?id=2891595) which identified how an entry
733 in the DNS cache would linger too long if the request that added it was in
734 use that long. He also provided the patch that now makes libcurl capable of
735 still doing a request while the DNS hash entry may get timed out.
737 - Christian Schmitz noticed that the progress meter/callback was not properly
738 used during the FTP connection phase (after the actual TCP connect), while
739 it of course should be. I also made the speed check get called correctly so
740 that really slow servers will trigger that properly too.
742 Kamil Dudka (5 Nov 2009)
743 - Dropped misleading timeouts in libcurl-NSS and made sure the SSL socket works
744 in non-blocking mode.
746 Yang Tse (5 Nov 2009)
747 - I removed leading 'curl' path on the 'curlbuild.h' include statement in
748 curl.h, adjusting auto-makefiles include path, to enhance portability to
749 OS's without an orthogonal directory tree structure such as OS/400.
751 Daniel Stenberg (4 Nov 2009)
752 - I fixed several problems with the transfer progress meter. It showed the
753 wrong percentage for small files, most notable for <1000 bytes and could
754 easily end up showing more than 100% at the end. It also didn't show any
755 percentage, transfer size or estimated transfer times when transferring
758 Version 7.19.7 (4 November 2009)
760 Daniel Stenberg (2 Nov 2009)
761 - As reported independent by both Stan van de Burgt and Didier Brisebourg,
762 CURLINFO_SIZE_DOWNLOAD (the -w variable size_download) didn't work when
763 getting data from ldap!
765 Daniel Stenberg (31 Oct 2009)
766 - Gabriel Kuri reported a problem with CURLINFO_CONTENT_LENGTH_DOWNLOAD if the
767 download was 0 bytes, as libcurl would then return the size as unknown (-1)
768 and not 0. I wrote a fix and test case 566 to verify it.
770 Daniel Stenberg (30 Oct 2009)
771 - Liza Alenchery mentioned a problem with re-used SCP connection when a bad
772 auth is used, as it caused a crash. I failed to repeat the issue, but still
773 made a change that now forces the TCP connection used for a freed SCP
774 session to get closed and not be re-used.
776 - "Tom" posted a bug report that mentioned how libcurl did wrong when doing a
777 POST using a read callback, with Digest authentication and
778 "Transfer-Encoding: chunked" enforced. I would then cause the first request
779 to be wrongly sent and then basically hang until the server closed the
780 connection. I fixed the problem and added test case 565 to verify it.
782 Daniel Stenberg (25 Oct 2009)
783 - Dima Barsky made the curl cookie parser accept cookies even with blank or
784 unparsable expiry dates and then treat them as session cookies - previously
785 libcurl would reject cookies with a date format it couldn't parse. Research
786 shows that the major browser treat such cookies as session cookies. I
787 modified test 8 and 31 to verify this.
789 Daniel Stenberg (21 Oct 2009)
790 - Attempt to use pkg-config for finding out libssh2 installation details
793 - A patch in bug report #2883177 (http://curl.haxx.se/bug/view.cgi?id=2883177)
794 by Johan van Selst introduced the --crlfile option to curl, which makes curl
795 tell libcurl about a file with CRL (certificate revocation list) data to
798 Daniel Stenberg (18 Oct 2009)
799 - Ray Dassen provided a patch in Debian's bug tracker (bug number #551461)
800 that now makes curl_getdate(3) actually handles RFC 822 formatted dates that
801 use the "single letter military timezones".
802 http://www.rfc-ref.org/RFC-TEXTS/822/chapter5.html has the details.
804 - Fixed memory leak in the SCP/SFTP code as it never freed the knownhosts
807 - John Dennis filed bug report #2873666
808 (http://curl.haxx.se/bug/view.cgi?id=2873666) which identified a problem
809 which made libcurl loop infinitely when given incorrect credentials when
810 using HTTP GSS negotiate authentication. He also provided a small and simple
813 - Kevin Baughman found a double close() problem with libcurl-NSS, as when
814 libcurl called NSS to close the SSL "session" it also closed the actual
817 Yang Tse (17 Oct 2009)
818 - Bug report #2866724 indicated
819 (http://curl.haxx.se/bug/view.cgi?id=2866724) that curl on Windows failed
820 when writing files whose file names originally contained characters which
821 are not valid for file names on Windows. Dan Fandrich provided an initial
822 patch and another revised one to fix this issue.
824 Daniel Stenberg (1 Oct 2009)
825 - Tom Mueller correctly reported in bug report #2870221
826 (http://curl.haxx.se/bug/view.cgi?id=2870221) that libcurl returned an
827 incorrect return code from the internal trynextip() function which caused
828 him grief. This is a regression that was introduced in 7.19.1 and I find it
829 strange it hasn't hit us harder, but I won't persue into figuring out
832 - Constantine Sapuntzakis: The current implementation will always set
833 SO_SNDBUF to CURL_WRITE_SIZE even if the SO_SNDBUF starts out larger. The
834 patch doesn't do a setsockopt if SO_SNDBUF is already greater than
835 CURL_WRITE_SIZE. This should help folks who have set up their computer with
838 Daniel Stenberg (27 Sep 2009)
839 - I introduced a maximum limit for received HTTP headers. It is controlled by
840 the define CURL_MAX_HTTP_HEADER which is even exposed in the public header
841 file to allow for users to fairly easy rebuild libcurl with a modified
842 limit. The rationale for a fixed limit is that libcurl is realloc()ing a
843 buffer to be able to put a full header into it, so that it can call the
844 header callback with the entire header, but that also risk getting it into
845 trouble if a server by mistake or willingly sends a header that is more or
846 less without an end. The limit is set to 100K.
848 Daniel Stenberg (26 Sep 2009)
849 - John P. McCaskey posted a bug report that showed how libcurl did wrong when
850 saving received cookies with no given path, if the path in the request had a
851 query part. That is means a question mark (?) and characters on the right
852 side of that. I wrote test case 1105 and fixed this problem.
854 Kamil Dudka (26 Sep 2009)
855 - Implemented a protocol independent way to specify blocking direction, used by
856 transfer.c for blocking. It is currently used only by SCP and SFTP protocols.
857 This enhancement resolves an issue with 100% CPU usage during SFTP upload,
860 Daniel Stenberg (25 Sep 2009)
861 - Chris Mumford filed bug report #2861587
862 (http://curl.haxx.se/bug/view.cgi?id=2861587) identifying that libcurl used
863 the OpenSSL function X509_load_crl_file() wrongly and failed if it would
864 load a CRL file with more than one certificate within. This is now fixed.
866 Daniel Stenberg (16 Sep 2009)
867 - Sven Anders reported that we introduced a cert verfication flaw for OpenSSL-
868 powered libcurl in 7.19.6. If there was a X509v3 Subject Alternative Name
869 field in the certficate it had to match and so even if non-DNS and non-IP
870 entry was present it caused the verification to fail.
872 Daniel Fandrich (15 Sep 2009)
873 - Moved the libssh2 checks after the SSL library checks. This helps when
874 statically linking since libssh2 needs the SSL library link flags to be
875 set up already to satisfy its dependencies. This wouldn't be necessary if
876 the libssh2 configure check was changed to use pkg-config since the
877 --static flag would add the dependencies automatically.
879 Yang Tse (14 Sep 2009)
880 - Revert Joshua Kwan's patch committed 11 Sep 2009.
882 Some systems poll function sets POLLHUP in revents without setting
883 POLLIN, and sets POLLERR without setting POLLIN and POLLOUT. In some
884 libcurl code execution paths this could trigger busy wait loops with
885 high CPU usage until a timeout condition aborted the loop.
887 The reverted patch addressed the above issue for a very specific case,
888 when awaiting c-ares to resolve. A libcurl-wide fix for Curl_poll now
891 Guenter Knauf (11 Sep 2009)
892 - Joshua Kwan provided a patch to pass POLLERR / POLLHUP back to c-ares.
893 This fixes a loop problem with high CPU usage.
895 Daniel Stenberg (10 Sep 2009)
896 - Claes Jakobsson fixed a problem with cookie expiry dates at exctly the epoch
897 start second "Thu Jan 1 00:00:00 GMT 1970" as the date parser then returns 0
898 which internally then is treated as a session cookie. That particular date
899 is now made to get the value of 1.
901 Daniel Stenberg (2 Sep 2009)
902 - Daniel Johnson found a flaw in the code converting sftp-errors to libcurl
905 Daniel Stenberg (1 Sep 2009)
906 - Peter Sylvester made a debug feature for Curl_resolv() that now will force
907 libcurl to resolve 'localhost' whatever name you use in the URL *if* you set
908 the --interface option to (exactly) "LocalHost". This will enable us to
909 write tests for custom hosts names but still use a local host server.
911 - configure now tries to use pkg-config for a number of sub-dependencies even
912 when cross-compiling. The key to success is then you properly setup
913 PKG_CONFIG_PATH before invoking configure.
915 I also improved how NSS is detected by trying nss-config if pkg-config isn't
916 present, and as a last resort just use the lib name and force the user to
917 setup the LIBS/LDFLAGS/CFLAGS etc properly. The previous last resort would
918 add a range of various libs that would almost never be quite correct.
920 Daniel Stenberg (31 Aug 2009)
921 - When using the multi interface with FTP and you asked for NOBODY, you did no
922 QUOTE commands and the request used the same path as the connection had
923 already changed to, it would decide that no commands would be necessary for
924 the "DO" action and that was not handled properly but libcurl would instead
927 Kamil Dudka (28 Aug 2009)
928 - Improved error message for not matching certificate subject name in
929 libcurl-NSS. Originally reported at:
930 https://bugzilla.redhat.com/show_bug.cgi?id=516056#c9
932 Patrick Monnerat (24 Aug 2009)
933 - Introduced a SYST-based test to properly set-up name format when dealing
934 with the OS/400 FTP server.
936 - Fixed an ftp_readresp() bug preventing detection of failing control socket
937 and causing FTP client to loop forever.
939 Daniel Stenberg (24 Aug 2009)
940 - Marc de Bruin pointed out that configure --with-gnutls=PATH didn't work
941 properly and provided a fix. http://curl.haxx.se/bug/view.cgi?id=2843008
943 - Eric Wong introduced support for the new option -T. (dot) that makes curl
944 read stdin in a non-blocking fashion. This also brings back -T- (minus) to
945 the previous blocking behavior since it could break stuff for people at
948 Michal Marek (21 Aug 2009)
949 - With CURLOPT_PROXY_TRANSFER_MODE, avoid sending invalid URLs like
950 ftp://example.com;type=i if the user specified ftp://example.com without the
953 Daniel Stenberg (21 Aug 2009)
954 - Andre Guibert de Bruet pointed out a missing return code check for a
955 strdup() that could lead to segfault if it returned NULL. I extended his
956 suggest patch to now have Curl_retry_request() return a regular return code
957 and better check that.
959 - Lots of good work by Krister Johansen, mostly related to pipelining:
961 Fix SIGSEGV on free'd easy_conn when pipe unexpectedly breaks
962 Fix data corruption issue with re-connected transfers
963 Fix use after free if we're completed but easy_conn not NULL
965 Kamil Dudka (13 Aug 2009)
966 - Changed NSS code to not ignore the value of ssl.verifyhost and produce more
967 verbose error messages. Originally reported at:
968 https://bugzilla.redhat.com/show_bug.cgi?id=516056
970 Daniel Stenberg (12 Aug 2009)
971 - Karl Moerder fixed the Makefile.vc* makefiles to include the new file
972 nonblock.c so that they work fine again
974 - I expanded test 517 with a bunch of more dates that originate from the
975 Chrome browser test suite. It turns out most of them get parsed the same
978 Version 7.19.6 (12 August 2009)
980 Daniel Stenberg (12 Aug 2009)
981 - Carsten Lange reported a bug and provided a patch for TFTP upload and the
982 sending of the TSIZE option. I don't like fixing bugs just hours before
983 a release, but since it was broken and the patch fixes this for him I decided
986 Daniel Stenberg (11 Aug 2009)
987 - Peter Sylvester made the HTTPS test server use specific certificates for
988 each test, so that the test suite can now be used to actually test the
989 verification of cert names etc. This made an error show up in the OpenSSL-
990 specific code where it would attempt to match the CN field even if a
991 subjectAltName exists that doesn't match. This is now fixed and verified
994 - Benbuck Nason posted the bug report #2835196
995 (http://curl.haxx.se/bug/view.cgi?id=2835196), fixing a few compiler
996 warnings when mixing ints and bools.
998 Daniel Fandrich (10 Aug 2009)
999 - Fixed a memory leak in the FTP code and an off-by-one heap buffer overflow.
1001 Daniel Fandrich (9 Aug 2009)
1002 - Fixed some memory leaks in the command-line tool that caused most of the
1003 torture tests to fail.
1005 Daniel Stenberg (2 Aug 2009)
1006 - Curt Bogmine reported a problem with SNI enabled on a particular server. We
1007 should introduce an option to disable SNI, but as we're in feature freeze
1008 now I've addressed the obvious bug here (pointed out by Peter Sylvester): we
1009 shouldn't try to enable SNI when SSLv2 or SSLv3 is explicitly selected.
1010 Code for OpenSSL and GnuTLS was fixed. NSS doesn't seem to have a particular
1011 option for SNI, or are we simply not using it?
1013 Daniel Stenberg (1 Aug 2009)
1014 - Scott Cantor posted the bug report #2829955
1015 (http://curl.haxx.se/bug/view.cgi?id=2829955) mentioning the recent SSL cert
1016 verification flaw found and exploited by Moxie Marlinspike. The presentation
1017 he did at Black Hat is available here:
1018 https://www.blackhat.com/html/bh-usa-09/bh-usa-09-archives.html#Marlinspike
1020 Apparently at least one CA allowed a subjectAltName or CN that contain a
1021 zero byte, and thus clients that assumed they would never have zero bytes
1022 were exploited to OK a certificate that didn't actually match the site. Like
1023 if the name in the cert was "example.com\0theatualsite.com", libcurl would
1024 happily verify that cert for example.com.
1026 libcurl now better uses the length of the extracted name, not using the zero
1027 termination for getting the string length.
1029 This fixing only made and needed in OpenSSL interfacing code.
1031 - Tanguy Fautre pointed out that OpenSSL's function RAND_screen() (present
1032 only in some OpenSSL installs - like on Windows) isn't thread-safe and we
1033 agreed that moving it to the global_init() function is a decent way to deal
1034 with this situation.
1036 - Alexander Beedie provided the patch for a noproxy problem: If I have set
1037 CURLOPT_NOPROXY to "*", or to a host that should not use a proxy, I actually
1038 could still end up using a proxy if a proxy environment variable was set.
1040 Daniel Stenberg (27 Jul 2009)
1041 - All the quote options (CURLOPT_QUOTE, CURLOPT_POSTQUOTE and
1042 CURLOPT_PREQUOTE) now accept a preceeding asterisk before the command to
1043 send when using FTP, as a sign that libcurl shall simply ignore the response
1044 from the server instead of treating it as an error. Not treating a 400+ FTP
1045 response code as an error means that failed commands will not abort the
1046 chain of commands, nor will they cause the connection to get disconnected.
1048 Daniel Stenberg (26 Jul 2009)
1049 - Johan van Selst posted bug report #2825989
1050 (http://curl.haxx.se/bug/view.cgi?id=2825989) pointing out that
1051 OpenSSL-powered libcurl didn't support the SHA-2 digest algorithm, and
1052 provided the solution too: to use OpenSSL_add_all_algorithms() in addition
1053 to the older SSLeay_* alternative. OpenSSL_add_all_algorithms was added in
1056 Daniel Stenberg (23 Jul 2009)
1057 - Added CURLOPT_SSH_KNOWNHOSTS, CURLOPT_SSH_KEYFUNCTION, CURLOPT_SSH_KEYDATA.
1058 They introduce known_host support for SSH keys to libcurl. See docs for
1059 details. Note that this feature depends on a new enough libssh2 version, to
1060 be supported in libssh2 1.2 and later (or current git repo at this time).
1062 Michal Marek (22 Jul 2009)
1063 - David Binderman found a memory and fd leak in lib/gtls.c:load_file()
1064 (https://bugzilla.novell.com/523919). When looking at the code, I found that
1065 also the ptr pointer can leak.
1067 Kamil Dudka (20 Jul 2009)
1068 - Claes Jakobsson improved the support for client certificates handling in
1069 NSS-powered libcurl. Now the client certificates can be selected
1070 automatically by a NSS built-in hook. Additionally pre-login to all PKCS11
1071 slots is no more performed. It used to cause problems with HW tokens.
1073 - Fixed reference counting for NSS client certificates. Now the PEM reader
1074 module should be always properly unloaded on Curl_nss_cleanup(). If the
1075 unload fails though, libcurl will try to reuse the already loaded instance.
1077 Daniel Fandrich (15 Jul 2009)
1078 - Added nonblock.c to the non-automake makefiles (note that the dependencies
1079 in the Watcom makefiles aren't quite correct).
1081 Michal Marek (15 Jul 2009)
1082 - Changed the description of CURLINFO_OS_ERRNO to make it clear that the
1083 errno is not reset on success.
1085 Guenter Knauf (14 Jul 2009)
1086 - renamed generated config.h to curl_config.h to avoid any future clashes
1087 with config.h from other projects.
1089 Daniel Stenberg (9 Jul 2009)
1090 - Eric Wong introduced curlx_nonblock() that the curl tool now (re-)uses for
1091 setting a file descriptor non-blocking. Used by the functionality Eric
1092 himself brough on June 15th.
1094 Daniel Stenberg (8 Jul 2009)
1095 - Constantine Sapuntzakis posted bug report #2813123
1096 (http://curl.haxx.se/bug/view.cgi?id=2813123) and an a patch that fixes the
1099 Url A is accessed using auth. Url A redirects to Url B (on a different
1100 server0. Url B reuses a persistent connection. Url B has auth, even though
1101 it's on a different server.
1103 Note: if Url B does not reuse a persistent connection, auth is not sent.
1107 data->state.first_host is not initialized becuase Curl_http_connect is not
1108 called when a connection is reused.
1112 move initialization of data->state.first_host to Curl_http. No code before
1113 Curl_http uses data->state.first_host anyway.
1115 Guenter Knauf (4 Jul 2009)
1116 - Markus Koetter provided a patch to avoid getnameinfo() usage which broke a
1117 couple of both IPv4 and IPv6 autobuilds.
1119 Daniel Stenberg (29 Jun 2009)
1120 - Markus Koetter made CURLOPT_FTPPORT (and curl's -P/--ftpport) support a port
1121 range if given colon-separated after the host name/address part. Like
1122 "192.168.0.1:2000-10000"
1124 - Modified the separators used for CURLOPT_CERTINFO in multi-part outputs. I
1125 don't know how they got wrong in the first place, but using this output
1126 format makes it possible to quite easily separate the string into an array
1129 Daniel Fandrich (16 June 2009)
1130 - Added a few more compiler warning options for gcc.
1132 Daniel Stenberg (16 Jun 2009)
1133 - Reuven Wachtfogel made curl -o - properly produce a binary output on windows
1134 (no newline translations). Use -B/--use-ascii if you rather get the ascii
1137 Michal Marek (16 Jun 2009)
1138 - When doing non-anonymous ftp via http proxies and the password is not
1139 provided in the url, add it there (squid needs this).
1141 Daniel Stenberg (15 Jun 2009)
1142 - Eric Wong's patch:
1144 This allows curl(1) to be used as a client-side tunnel for arbitrary stream
1145 protocols by abusing chunked transfer encoding in both the HTTP request and
1146 HTTP response. This requires server support for sending a response while a
1147 request is still being read, of course.
1149 If attempting to read from stdin returns EAGAIN, then we pause our sender.
1150 This leaves curl to attempt to read from the socket while reading from stdin
1151 (and thus sending) is paused.
1153 This change was needed to allow successfully tunneling the git protocol over
1154 HTTP (--no-buffer is needed, as well).
1156 Patrick Monnerat (15 Jun 2009)
1157 - Replaced use of standard C library rand()/srand() by our own pseudo-random
1160 Yang Tse (11 Jun 2009)
1161 - I adapted testcurl script to allow building test harness programs when
1162 cross-compiling for a *-*-mingw* host.
1164 Daniel Stenberg (10 Jun 2009)
1165 - Fabian Keil ran clang on the (lib)curl code, found a bunch of warnings and
1166 contributed a range of patches to fix them.
1168 Yang Tse (10 Jun 2009)
1169 - I introduced configure script option --enable-curldebug which now allows
1170 the decoupled enabling or disabling of the curl debug memory tracking
1171 feature from the --enable-debug option which no longer controls this.
1173 curl --version will list 'Debug' feature for debug enabled builds, and
1174 will list 'TrackMemory' feature for curl debug memory tracking capable
1175 builds. These features are independent and can be controlled when running
1176 the configure script. When --enable-debug is given both features will be
1177 enabled, unless some restriction prevents memory tracking from being used.
1179 Internally, definition of preprocessor symbol DEBUGBUILD restricts code
1180 which is only compiled for debug enabled builds. And symbol CURLDEBUG is
1181 used to differentiate code which is _only_ used for memory tracking.
1183 Yang Tse (9 Jun 2009)
1184 - Daniel Steinberg pointed out that Curl_FormInit() in formdata.c was not
1185 initializing the fread callback pointer and this triggered a compiler
1186 warning, also provided a friendly suggestion on how to fix it.
1188 Daniel Stenberg (8 Jun 2009)
1189 - Claes Jakobsson provided a patch for libcurl-NSS that fixed a bad refcount
1190 issue with client certs that caused issues like segfaults.
1191 http://curl.haxx.se/mail/lib-2009-05/0316.html
1193 - Triggered by bug report #2798852 and the patch in there, I fixed configure
1194 to detect gnutls build options with pkg-config only and not libgnutls-config
1195 anymore since GnuTLS has stopped distributing that tool. If an explicit path
1196 is given to configure, we will instead guess on how to link and use that
1197 lib. I did not use the patch from the bug report.
1199 Yang Tse (8 Jun 2009)
1200 - Igor Novoseltsev adjusted Makefile.vxworks to get sources and headers
1201 included from Makefile.inc, and provided docs\INSTALL VxWorks section.
1203 - I removed buildconf.bat from release and daily snapshot archives. This
1204 file is only for CVS tree checkout builds.
1206 Daniel Stenberg (8 Jun 2009)
1207 - Eric Wong fixed --no-buffer to actually switch off output buffering. Been
1210 Bill Hoffman (6 Jun 2009)
1211 - Added some cmake docs and fixed socklen_t in the build.
1213 Yang Tse (5 Jun 2009)
1214 - John E. Malmberg provided VMS specific patch: "This fixes an existing bug
1215 in urlglob.c where it was not converting the Curl Unix exit code to a VMS
1216 DCL compatible exit code. This fix required the enhancement described next.
1217 This also adds an enhancement to main.c so that when curl is run under a
1218 Unix shell like Bash on VMS, it will return the standard Unix exit codes
1219 and messages." And another patch for docs/examples.
1221 I introduced os-specific.c and os-specific.h for use in curl tool code
1222 and adjusted John E. Malmberg's patch placement to use these new files
1223 as an effort to prevent main.c from growing ad infinitum. Code already
1224 existing in main.c which is OS specific should be moved into these files.
1226 Daniel Stenberg (4 June 2009)
1227 - Setting the Content-Length: header from your app when you do a POST or PUT
1228 is almost always a VERY BAD IDEA. Yet there are still apps out there doing
1229 this, and now recently it triggered a bug/side-effect in libcurl as when
1230 libcurl sends a POST or PUT with NTLM, it sends an empty post first when it
1231 knows it will just get a 401/407 back. If the app then replaced the
1232 Content-Length header, it caused the server to wait for input that libcurl
1233 wouldn't send. Aaron Oneal reported this problem in bug report #2799008
1234 (http://curl.haxx.se/bug/view.cgi?id=2799008) and helped us verify the fix.
1236 Yang Tse (4 Jun 2009)
1237 - Igor Novoseltsev provided patches and information, that after some
1238 adjustments to better fit curl's way of doing things, have resulted
1239 in the posibility of building libcurl for VxWorks.
1241 Daniel Fandrich (2 June 2009)
1242 - Checked in a Google Android make file. To use it, you must first
1243 create a config.h file by running configure in the Android environment,
1244 which doesn't seem to be easy to do. If no easy way can be found, a
1245 static config-android.h may need to be created and checked in to the
1246 libcurl source tree.
1248 Daniel Stenberg (1 June 2009)
1249 - Claes Jakobsson fixed the configure script to better find and use NSS
1252 Yang Tse (1 Jun 2009)
1253 - John E. Malmberg provided a VMS specific clean-up for curl.h, and pointed
1254 out that the configure script was failing to detect the timeval struct on
1255 VMS when building with _XOPEN_SOURCE_EXTENDED undefined due to definition
1256 taking place in socket.h instead of time.h. I have adjusted configure
1257 script to also include this header when checking struct timeval.
1259 Daniel Stenberg (27 May 2009)
1260 - Frank McGeough provided a small OpenSSL #include fix to make libcurl compile
1261 fine with Nokia 5th edition 1.0 SDK for Symbian.
1263 - Andre Guibert de Bruet found a call to a OpenSSL function that didn't check
1264 for a failure properly.
1266 - Mike Crowe pointed out that setting CURLOPT_USERPWD to NULL used to clear
1267 the auth credentials back in 7.19.0 and earlier while now you have to set ""
1268 to get the same effect. His patch brings back the ability to use NULL.
1270 - Claes Jakobsson fixed libcurl-NSS to build fine even without the
1271 PK11_CreateGenericObject() function.
1273 Daniel Stenberg (25 May 2009)
1274 - bug report #2796358 (http://curl.haxx.se/bug/view.cgi?id=2796358) pointed
1275 out that the cookie parser would leak memory when it parses cookies that are
1276 received with domain, path etc set multiple times in the same header. While
1277 such a cookie is questionable, they occur in the wild and libcurl no longer
1278 leaks memory for them. I added such a header to test case 8.
1280 Daniel Fandrich (22 May 2009)
1281 - Removed some obsolete digest code that caused a valgrind error in test 551.
1283 Daniel Fandrich (20 May 2009)
1284 - Added "non-existing host" test keywords to make it easy to skip those
1285 tests on machines that have broken DNS configurations (such as
1286 those configured to use OpenDNS).
1288 Daniel Stenberg (19 May 2009)
1289 - Kamil Dudka brought the patch from the Redhat bug entry
1290 https://bugzilla.redhat.com/show_bug.cgi?id=427966 which was libcurl closing
1291 a bad file descriptor when closing down the FTP data connection. Caolan
1292 McNamara seems to be the original author of it.
1294 Version 7.19.5 (18 May 2009)
1296 Daniel Stenberg (17 May 2009)
1297 - James Bursa posted a patch to the mailing list that fixed a problem with
1298 no_proxy which made it not skip the proxy if the URL entered contained a
1299 user name. I added test case 1101 to verify.
1301 Daniel Stenberg (11 May 2009)
1302 - Balint Szilakszi reported a memory leak when libcurl did gzip decompression
1303 of streams that had some parts (legitimately) missing. We now provide and use
1304 a proper cleanup function for the content encoding submodule.
1305 http://curl.haxx.se/mail/lib-2009-05/0092.html
1307 - Kamil Dudka provided a fix for libcurl-NSS reported by Michael Cronenworth
1308 at https://bugzilla.redhat.com/show_bug.cgi?id=453612#c12
1310 If an incorrect password is given while loading a private key, libcurl ends
1311 up in an infinite loop consuming memory. The bug is critical.
1313 - I fixed the problem with doing NTLM, POST and then following a 302 redirect,
1314 as reported by Ebenezer Ikonne (on curl-users) and Laurent Rabret (on
1315 curl-library). The transfer was mistakenly marked to get more data to send
1316 but since it didn't actually have that, it just hung there...
1318 Daniel Stenberg (10 May 2009)
1319 - Andre Guibert de Bruet correctly pointed out an over-alloc with one wasted
1320 byte in the digest code.
1322 Yang Tse (9 May 2009)
1323 - Removed DOS and TPF package's subdirectory Makefile.am, it was only used
1324 to include some files in the distribution tarball serving no other purpose.
1325 Files from the DOS and TPF subdirectories are now included in the EXTRA_DIST
1326 of the Makefile in the parent subdirectory.
1328 Yang Tse (8 May 2009)
1329 - Changed host name literal in several tests to one under the haxx.se domain.
1331 - Renamed vc6 workspace and project files to avoid filename clash when used
1332 for conversion to later VS versions.
1334 Daniel Stenberg (8 May 2009)
1335 - Constantine Sapuntzakis fixed bug report #2784055
1336 (http://curl.haxx.se/bug/view.cgi?id=2784055) identifying a problem to
1337 connect to SOCKS proxies when using the multi interface. It turned out to
1338 almost not work at all previously. We need to wait for the TCP connect to
1339 be properly verified before doing the SOCKS magic.
1341 There's still a flaw in the FTP code for this.
1343 Daniel Stenberg (7 May 2009)
1344 - Made the SO_SNDBUF setting for the data connection socket for ftp uploads as
1345 well. See change 28 Apr 2009.
1347 Yang Tse (7 May 2009)
1348 - Fixed an issue affecting FTP transfers, introduced with the transfer.c
1349 patch committed May 4.
1351 Daniel Stenberg (7 May 2009)
1352 - Man page *roff problems fixed thanks to input from Colin Watson. Problems
1353 reported in the Debian package.
1355 - Vijay G filed bug report #2723236
1356 (http://curl.haxx.se/bug/view.cgi?id=2723236) identifying a problem with
1357 libcurl's TFTP code and its lack of dealing with the OACK packet.
1359 Yang Tse (5 May 2009)
1360 - Fixed the --ftp-port address of test #251 to the CLIENTIP address, and
1361 reverted the change affecting test suite harness committed 4 May.
1363 Daniel Stenberg (5 May 2009)
1364 - Inspired by Michael Smith's session id fix for OpenSSL, I did the
1365 corresponding fix in the GnuTLS code: make sure to store the new session id
1366 in case the previous re-used one is rejected.
1368 Daniel Stenberg (4 May 2009)
1369 - Michael Smith posted bug report #2786255
1370 (http://curl.haxx.se/bug/view.cgi?id=2786255) with a patch, identifying how
1371 libcurl did not deal with SSL session ids properly if the server rejected a
1372 re-use of one. Starting now, it will forget the rejected one and remember
1373 the new. This change was for OpenSSL only, it is likely that other SSL lib
1374 code needs similar fixes.
1376 Yang Tse (4 May 2009)
1377 - Applied David McCreedy's "transfer.c fixes for CURL_DO_LINEEND_CONV and
1378 non-ASCII platform HTTP requests" patch addressing two HTTP PUT problems:
1379 1) On non-ASCII platforms not all of the protocol portions of the PUT are
1380 being translated to ASCII. 2) On all platforms the line endings of part of
1381 the protocol portions are mangled from CRLF to CRCRLF if data->set.crlf or
1382 data->set.prefer_ascii are set (depending on CURL_DO_LINEEND_CONV).
1384 - Applied David McCreedy's patch to fix test suite harness to allow test FTP
1385 server and client on different machines, providing FTP client address when
1386 running the FTP test server.
1388 Daniel Fandrich (3 May 2009)
1389 - Added and disabled test case 563 which shows KNOWN_BUGS #59. The bug
1390 report failed to mention that a proxy must be used to reproduce it.
1392 Yang Tse (2 May 2009)
1393 - Use a build-time configured curl_socklen_t data type instead of socklen_t.
1395 Yang Tse (1 May 2009)
1396 - Applied David McCreedy's patches "TPF-platform specific changes to various
1397 files" and "http.c fix to Curl_proxyCONNECT for non-ASCII platforms", the
1398 former with minor edits.
1400 Daniel Stenberg (30 Apr 2009)
1401 - I was going to fix issue #59 in KNOWN_BUGS
1403 If the CURLOPT_PORT option is used on an FTP URL like
1404 "ftp://example.com/file;type=A" the ";type=A" is stripped off.
1406 I added test case 562 to verify, only to find out that I couldn't repeat
1407 this bug so I hereby consider it not a bug anymore!
1409 Daniel Stenberg (29 Apr 2009)
1410 - Based on bug report #2723219 (http://curl.haxx.se/bug/view.cgi?id=2723219)
1411 I've now made TFTP "connections" not being kept for re-use within libcurl.
1412 TFTP is UDP-based so the benefit was really low (if even existing) to begin
1413 with so instead of tracking down to fix this problem we instead removed the
1414 re-use. I also enabled test case 1099 that I wrote a few days ago to verify
1415 that this change fixes the reported problem.
1417 Daniel Stenberg (28 Apr 2009)
1418 - Constantine Sapuntzakis filed bug report #2783090
1419 (http://curl.haxx.se/bug/view.cgi?id=2783090) pointing out that on windows
1420 we need to grow the SO_SNDBUF buffer somewhat to get really good upload
1421 speeds. http://support.microsoft.com/kb/823764 has the details. Friends
1422 confirmed that simply adding 32 to CURL_MAX_WRITE_SIZE is enough.
1424 - Bug report #2709004 (http://curl.haxx.se/bug/view.cgi?id=2709004) by Tim
1425 Chen pointed out how curl couldn't upload with resume when reading from a
1428 This ended up with the introduction of a new return code for the
1429 CURLOPT_SEEKFUNCTION callback that basically says that the seek failed but
1430 that libcurl may try to resolve the situation anyway. In our case this means
1431 libcurl will attempt to instead read that much data from the stream instead
1432 of seeking and that way curl can now upload with resume when data is read
1435 Daniel Stenberg (26 Apr 2009)
1436 - Bug report #2779733 (http://curl.haxx.se/bug/view.cgi?id=2779733) by Sven
1437 Wegener pointed out that CURLINFO_APPCONNECT_TIME didn't work with the multi
1438 interface and provided a patch that fixed the problem!
1440 Daniel Stenberg (24 Apr 2009)
1441 - Kamil Dudka fixed another NSS-related leak when client certs were used.
1443 - Bug report #2779245 (http://curl.haxx.se/bug/view.cgi?id=2779245) by Rainer
1444 Koenig pointed out that the man page didn't tell that the *_proxy
1445 environment variables can be specified lower case or UPPER CASE and the
1446 lower case takes precedence,
1448 Daniel Fandrich (21 Apr 2009)
1449 - Added new libcurl source files to Amiga, RiscOS and VC6 build files.
1451 Yang Tse (21 Apr 2009)
1452 - Moved potential inclusion of system's malloc.h and memory.h header files to
1453 setup_once.h. Inclusion of each header file is based on the definition of
1454 NEED_MALLOC_H and NEED_MEMORY_H respectively.
1456 Renamed libcurl's memory.h to curl_memory.h
1458 Daniel Stenberg (20 Apr 2009)
1459 - Leanic Lefever reported a crash and did some detailed research on why and
1460 how it occurs (http://curl.haxx.se/mail/lib-2009-04/0289.html). The
1461 conclusion was that if an error is detected and Curl_done() is called for
1462 the connection, ftp_done() could at times return another error code that
1463 then would take precedence and that new code confused existing logic that
1464 works for the first error code (CURLE_SEND_ERROR) only.
1466 - Gisle Vanem noticed that --libtool would produce bogus strings at times for
1467 OBJECTPOINT options. Now we've introduced a new function - my_setopt_str -
1468 within the app for setting plain string options to avoid the risk of this
1471 Daniel Stenberg (17 Apr 2009)
1472 - Pramod Sharma reported and tracked down a bug when doing FTP over a HTTP
1473 proxy. libcurl would then wrongly close the connection after each
1474 request. In his case it had the weird side-effect that it killed NTLM auth
1475 for the proxy causing an inifinite loop!
1477 I added test case 1098 to verify this fix. The test case does however not
1478 properly verify that the transfers are done persistently - as I couldn't
1479 think of a clever way to achieve it right now - but you need to read the
1480 stderr output after a test run to see that it truly did the right thing.
1482 Daniel Stenberg (13 Apr 2009)
1483 - bug report #2727981 (http://curl.haxx.se/bug/view.cgi?id=2727981) by Martin
1484 Storsjö pointed out how setting CURLOPT_NOBODY to 0 could be downright
1485 confusing as it set the method to either GET or HEAD. The example he showed
1488 curl_easy_setopt(curl, CURLOPT_PUT, 1);
1489 curl_easy_setopt(curl, CURLOPT_NOBODY, 0);
1491 The new way doesn't alter the method until the request is about to start. If
1492 CURLOPT_NOBODY is then 1 the HTTP request will be HEAD. If CURLOPT_NOBODY is
1493 0 and the request happens to have been set to HEAD, it will then instead be
1494 set to GET. I believe this will be less surprising to users, and hopefully
1495 not hit any existing users badly.
1497 - Toshio Kuratomi reported a memory leak problem with libcurl+NSS that turned
1498 out to be leaking cacerts. Kamil Dudka helped me complete the fix. The issue
1499 is found in Redhat's bug tracker:
1500 https://bugzilla.redhat.com/show_bug.cgi?id=453612
1502 There are still memory leaks present, but they seem to have other reasons.
1504 Daniel Fandrich (11 Apr 2009)
1505 - Added new libcurl source files to Symbian OS build files.
1506 - Improved Symbian support for SSL.
1508 Yang Tse (10 Apr 2009)
1509 - Daniel Johnson improved the MacOSX-Framework shell script to now perform all
1510 the steps required to build a Mac OS X four way fat ppc/i386/ppc64/x86_64
1511 libcurl.framework. Four way fat framework requires OS X 10.5 SDK or later.
1513 Yang Tse (8 Apr 2009)
1514 - Removed Sun compilers preprocessor block from curlbuild.h.dist, this also
1515 removes it from the curlbuild.h file originally distributed by the cURL
1516 project as this file is intended for systems not capable of running the
1517 configure script. For those who have been building curl out of the source
1518 code curl distribution tarball provided by curl.haxx.se the change implies
1519 nothing. Previous change in this area committed 2 Apr becomes irrelevant.
1521 Daniel Stenberg (6 Apr 2009)
1522 - I clarified in the docs that CURLOPT_SEEKFUNCTION should return 0 on success
1523 and 1 on fatal errors. Previously it only mentioned non-zero on fatal
1524 errors. This is a slight change in meaning, but it follows what we've done
1525 elsewhere before and it opens up for LOTS of more useful return codes
1526 whenever we can think of them...
1528 Yang Tse (2 Apr 2009)
1529 - Fix curl_off_t definition for builds done using Sun compilers and a
1530 non-configured libcurl. In this case curl_off_t data type was gated
1531 to the off_t data type which depends on the _FILE_OFFSET_BITS. This
1532 configuration is exactly the unwanted configuration for our curl_off_t
1533 data type which must not depend on such setting. This breaks ABI for
1534 libcurl libraries built with Sun compilers which were built without
1535 having run the configure script with _FILE_OFFSET_BITS different than
1536 64 and using the ILP32 data model.
1538 Daniel Stenberg (1 Apr 2009)
1539 - Andre Guibert de Bruet fixed a NULL pointer use in an infof() call if a
1540 strdup() call failed.
1542 Daniel Fandrich (31 Mar 2009)
1543 - Properly return an error code in curl_easy_recv (reported by Jim Freeman).
1545 Daniel Stenberg (18 Mar 2009)
1546 - Kamil Dudka brought a patch that enables 6 additional crypto algorithms when
1547 NSS is used. These ciphers were added in NSS 3.4 and require to be enabled
1550 Daniel Stenberg (13 Mar 2009)
1551 - Use libssh2_version() to present the libssh2 version in case the libssh2
1552 library is found to support it.
1554 Yang Tse (12 Mar 2009)
1555 - Added missing Curl_read() return code checking in TELNET transfers.
1557 - Pierre Brico found and fixed TELNET transfers not being aborted upon
1558 a write callback failure.
1560 Daniel Stenberg (11 Mar 2009)
1561 - Kamil Dudka made the curl tool properly call curl_global_init() before any
1562 other libcurl function.
1564 Yang Tse (11 Mar 2009)
1565 - Added missing TELNET timeout support for Windows builds. This issue was
1566 reported by Pierre Brico.
1568 Daniel Stenberg (9 Mar 2009)
1569 - Frank Hempel found out a bug and provided the fix:
1571 curl_easy_duphandle did not necessarily duplicate the CURLOPT_COOKIEFILE
1572 option. It only enabled the cookie engine in the destination handle if
1573 data->cookies is not NULL (where data is the source handle). In case of a
1574 newly initialized handle which just had the cookie support enabled by a
1575 curl_easy_setopt(handle, CURL_COOKIEFILE, "")-call, handle->cookies was
1576 still NULL because the setopt-call only appends the value to
1577 data->change.cookielist, hence duplicating this handle would not have the
1578 cookie engine switched on.
1580 We also concluded that the slist-functionality would be suitable for being
1581 put in its own module rather than simply hanging out in lib/sendf.c so I
1582 created lib/slist.[ch] for them.
1584 - Andreas Farber made the 'buildconf' script check for the presence of m4
1585 scripts to make it detect a bad checkout earlier. People with older
1586 checkouts who don't do cvs update with the -d option won't get the new dirs
1587 and then will get funny outputs that can be a bit hard to understand and
1590 Daniel Stenberg (8 Mar 2009)
1591 - Andre Guibert de Bruet found and fixed a code segment in ssluse.c where the
1592 allocation of the memory BIO was not being properly checked.
1594 - Andre Guibert de Bruet fixed the gnutls-using code: There are a few places
1595 in the gnutls code where we were checking for negative values for errors,
1596 when the man pages state that GNUTLS_E_SUCCESS is returned on success and
1597 other values indicate error conditions.
1599 - Bill Egert pointed out (http://curl.haxx.se/bug/view.cgi?id=2671602) that
1600 curl didn't use sprintf() in a way that is documented to work in POSIX but
1601 since we use our own printf() code (from libcurl) that shouldn't be a
1602 problem. Nonetheless I modified the code to not rely on such particular
1603 features and to not cause further raised eyebrowse with no good reason.
1605 Daniel Fandrich (5 Mar 2009)
1606 - Expanded the security section of the libcurl-tutorial man page to cover
1607 more issues for authors to consider when writing robust libcurl-using
1610 Yang Tse (5 Mar 2009)
1611 - Fixed NTLM authentication memory leak on SSPI enabled Windows builds. This
1612 issue was noticed by Chris Deidun.
1614 Daniel Fandrich (4 Mar 2009)
1615 - Fixed a problem with m4 quoting in the OpenSSL configure check reported
1618 Daniel Stenberg (3 Mar 2009)
1619 - David James brought a patch that make libcurl close (all) dead connections
1620 whenever you attempt to open a new connection.
1622 1. After cleaning up a dead connection, "continue" instead of
1623 returning FALSE. This ensures that we clean up all dead connections,
1624 rather than just cleaning up the first dead connection.
1625 2. Move up the cleanup for dead connections so that it occurs for
1626 all connections, rather than just the connections which have the same
1627 preferences as our current new connection.
1629 Version 7.19.4 (3 March 2009)
1631 Daniel Stenberg (3 Mar 2009)
1632 - David Kierznowski notified us about a security flaw
1633 (http://curl.haxx.se/docs/adv_20090303.html also known as CVE-2009-0037) in
1634 which previous libcurl versions (by design) can be tricked to access an
1635 arbitrary local/different file instead of a remote one when
1636 CURLOPT_FOLLOWLOCATION is enabled. This flaw is now fixed in this release
1637 together this the addition of two new setopt options for controlling this
1640 o CURLOPT_REDIR_PROTOCOLS controls what protocols libcurl is allowed to
1641 follow to when CURLOPT_FOLLOWLOCATION is enabled. By default, this option
1642 excludes the FILE and SCP protocols and thus you nee to explicitly allow
1643 them in your app if you really want that behavior.
1645 o CURLOPT_PROTOCOLS controls what protocol(s) libcurl is allowed to fetch
1646 using the primary URL option. This is useful if you want to allow a user or
1647 other outsiders control what URL to pass to libcurl and yet not allow all
1648 protocols libcurl may have been built to support.
1650 Daniel Stenberg (27 Feb 2009)
1651 - Senthil Raja Velu reported a problem when CURLOPT_INTERFACE and
1652 CURLOPT_LOCALPORT were used together (the local port bind failed), and
1653 Markus Koetter provided the fix!
1655 Daniel Stenberg (25 Feb 2009)
1656 - As Daniel Fandrich figured out, we must do the GnuTLS initing in the
1657 curl_global_init() function to properly maintain the performing functions
1658 thread-safe. We've previously (28 April 2007) moved the init to a later time
1659 just to avoid it to fail very early when libgcrypt dislikes the situation,
1660 but that move was bad and the fix should rather be in libgcrypt or
1663 Daniel Stenberg (24 Feb 2009)
1664 - Brian J. Murrell found out that Negotiate proxy authentication didn't work.
1665 It happened because the code used the struct for server-based auth all the
1666 time for both proxy and server auth which of course was wrong.
1668 Daniel Stenberg (23 Feb 2009)
1669 - After a bug reported by James Cheng I've made curl_easy_getinfo() for
1670 CURLINFO_CONTENT_LENGTH_DOWNLOAD and CURLINFO_CONTENT_LENGTH_UPLOAD return
1671 -1 if the sizes aren't know. Previously these returned 0, make it impossible
1672 to detect the difference between actually zero and unknown.
1674 Yang Tse (23 Feb 2009)
1675 - Daniel Johnson provided a shell script that will perform all the steps needed
1676 to build a Mac OS X fat ppc/i386 or ppc64/x86_64 libcurl.framework
1678 Daniel Stenberg (23 Feb 2009)
1679 - I renamed everything in the windows builds files that used the name 'curllib'
1680 to the proper 'libcurl' as clearly this caused confusion.
1682 Yang Tse (20 Feb 2009)
1683 - Do not halt compilation when using VS2008 to build a Windows 2000 target.
1685 Daniel Stenberg (20 Feb 2009)
1686 - Linus Nielsen Feltzing reported and helped me repeat and fix a problem with
1687 FTP with the multi interface: when a transfer fails, like when aborted by a
1688 write callback, the control connection was wrongly closed and thus not
1691 This change is also an attempt to cleanup the code somewhat in this area, as
1692 now the FTP code attempts to keep (better) track on pending responses
1693 necessary to get read in ftp_done().
1695 Daniel Stenberg (19 Feb 2009)
1696 - Patrik Thunstrom reported a problem and helped me repeat it. It turned out
1697 libcurl did a superfluous 1000ms wait when doing SFTP downloads!
1699 We read data with libssh2 while doing the "DO" operation for SFTP and then
1700 when we were about to start getting data for the actual file part, the
1701 "TRANSFER" part, we waited for socket action (in 1000ms) before doing a
1702 libssh2-read. But in this case libssh2 had already read and buffered the
1703 data so we ended up always just waiting 1000ms before we get working on the
1706 Patrick Monnerat (18 Feb 2009)
1707 - FTP downloads (i.e.: RETR) ending with code 550 now return error
1708 CURLE_REMOTE_FILE_NOT_FOUND instead of CURLE_FTP_COULDNT_RETR_FILE.
1710 Daniel Stenberg (17 Feb 2009)
1711 - Kamil Dudka made NSS-powered builds compile and run again!
1713 - A second follow-up change by Andre Guibert de Bruet to fix a related memory
1714 leak like that fixed on the 14th. When zlib returns failure, we need to
1715 cleanup properly before returning error.
1717 - CURLOPT_FTP_CREATE_MISSING_DIRS can now be set to 2 in addition to 1 for
1718 plain FTP connections, and it will then allow MKD to fail once and retry the
1719 CWD afterwards. This is especially useful if you're doing many simultanoes
1720 connections against the same server and they all have this option enabled,
1721 as then CWD may first fail but then another connection does MKD before this
1722 connection and thus MKD fails but trying CWD works! The numbers can
1723 (should?) now be set with the convenience enums now called
1724 CURLFTP_CREATE_DIR and CURLFTP_CREATE_DIR_RETRY.
1726 Tests has proven that if you're making an application that uploads a set of
1727 files to an ftp server, you will get a noticable gain in speed if you're
1728 using multiple connections and this option will be then be very useful.
1730 Daniel Stenberg (14 Feb 2009)
1731 - Andre Guibert de Bruet found and fixed a memory leak in the content encoding
1732 code, which could happen on libz errors.
1734 Daniel Fandrich (12 Feb 2009)
1735 - Added support for Digest and NTLM authentication using GnuTLS.
1737 Daniel Stenberg (11 Feb 2009)
1738 - CURLINFO_CONDITION_UNMET was added to allow an application to get to know if
1739 the condition in the previous request was unmet. This is typically a time
1740 condition set with CURLOPT_TIMECONDITION and was previously not possible to
1741 reliably figure out. From bug report #2565128
1742 (http://curl.haxx.se/bug/view.cgi?id=2565128) filed by Jocelyn Jaubert.
1744 Daniel Fandrich (4 Feb 2009)
1745 - Don't add the standard /usr/lib or /usr/include paths to LDFLAGS and CPPFLAGS
1746 (respectively) when --with-ssl=/usr is used (patch based on FreeBSD).
1748 - Added an explicit buffer limit check in msdosify() (patch based on FreeBSD).
1749 This couldn't ever overflow in curl, but might if the code were used
1750 elsewhere or under different conditions.
1752 Daniel Stenberg (3 Feb 2009)
1753 - Hidemoto Nakada provided a small fix that makes it possible to get the
1754 CURLINFO_CONTENT_LENGTH_DOWNLOAD size from file:// "transfers" with
1755 CURLOPT_NOBODY set true.
1757 Daniel Stenberg (2 Feb 2009)
1758 - Patrick Scott found a rather large memory leak when using the multi
1759 interface and setting CURLMOPT_MAXCONNECTS to something less than the number
1760 of handles you add to the multi handle. All the connections that didn't fit
1761 in the cache would not be properly disconnected nor freed!
1763 - Craig A West brought us: libcurl now defaults to do CONNECT with HTTP
1764 version 1.1 instead of 1.0 like before. This change also introduces the new
1765 proxy type for libcurl called 'CURLPROXY_HTTP_1_0' that then allows apps to
1766 switch (back) to CONNECT 1.0 requests. The curl tool also got a --proxy1.0
1767 option that works exactly like --proxy but sets CURLPROXY_HTTP_1_0.
1769 I updated all test cases cases that use CONNECT and I tried to do some using
1770 --proxy1.0 and some updated to do CONNECT 1.1 to get both versions run.
1772 Daniel Stenberg (31 Jan 2009)
1773 - When building with c-ares 1.6.1 (not yet released) or later and IPv6 support
1774 enabled, we can now take advantage of its brand new AF_UNSPEC support in
1775 ares_gethostbyname(). This makes test case 241 finally run fine for me with
1776 this setup since it now parses the "::1 ip6-localhost" line fine in my
1779 Daniel Stenberg (30 Jan 2009)
1780 - Scott Cantor filed bug report #2550061
1781 (http://curl.haxx.se/bug/view.cgi?id=2550061) mentioning that I failed to
1782 properly make sure that the VC9 makefiles got included in the latest
1783 release. I've now fixed the release script and verified it so next release
1784 will hopefully include them properly!
1786 Daniel Fandrich (30 Jan 2009)
1787 - Fixed --disable-proxy for FTP and SOCKS. Thanks to Daniel Egger for
1790 Yang Tse (29 Jan 2009)
1791 - Introduced curl_sspi.c and curl_sspi.h for the implementation of functions
1792 Curl_sspi_global_init() and Curl_sspi_global_cleanup() which previously were
1793 named Curl_ntlm_global_init() and Curl_ntlm_global_cleanup() in http_ntlm.c
1794 Also adjusted socks_sspi.c to remove the link-time dependency on the Windows
1795 SSPI library using it now in the same way as it was done in http_ntlm.c.
1797 Daniel Stenberg (28 Jan 2009)
1798 - Markus Moeller introduced two new options to libcurl:
1799 CURLOPT_SOCKS5_GSSAPI_SERVICE and CURLOPT_SOCKS5_GSSAPI_NEC to allow libcurl
1800 to do GSS-style authentication with SOCKS5 proxies. The curl tool got the
1801 options called --socks5-gssapi-service and --socks5-gssapi-nec to enable
1804 Daniel Stenberg (26 Jan 2009)
1805 - Chad Monroe provided the new CURLOPT_TFTP_BLKSIZE option that allows an app
1806 to set desired block size to use for TFTP transfers instead of the default
1809 - The "-no_ticket" option was introduced in Openssl0.9.8j. It's a flag to
1810 disable "rfc4507bis session ticket support". rfc4507bis was later turned
1811 into the proper RFC5077 it seems: http://tools.ietf.org/html/rfc5077
1813 The enabled extension concerns the session management. I wonder how often
1814 libcurl stops a connection and then resumes a TLS session. also, sending the
1815 session data is some overhead. .I suggest that you just use your proposed
1816 patch (which explicitly disables TICKET).
1818 If someone writes an application with libcurl and openssl who wants to
1819 enable the feature, one can do this in the SSL callback.
1821 Sharad Gupta brought this to my attention. Peter Sylvester helped me decide
1822 on the proper action.
1824 - Alexey Borzov filed bug report #2535504
1825 (http://curl.haxx.se/bug/view.cgi?id=2535504) pointing out that realms with
1826 quoted quotation marks in HTTP Digest headers didn't work. I've now added
1827 test case 1095 that verifies my fix.
1829 - Craig A West brought CURLOPT_NOPROXY and the corresponding --noproxy option.
1830 They basically offer the same thing the NO_PROXY environment variable only
1831 offered previously: list a set of host names that shall not use the proxy
1832 even if one is specified.
1834 Daniel Fandrich (20 Jan 2009)
1835 - Call setlocale() for libtest tests to test the effects of locale-induced
1836 libc changes on libcurl.
1838 - Fixed a couple more locale-dependent toupper conversions, mainly for
1839 clarity. This does fix one problem that causes ;type=i FTP URLs
1840 to fail in the Turkish locale when CURLOPT_PROXY_TRANSFER_MODE is
1841 used (test case 561)
1843 - Added tests 561 and 1091 through 1094 to test various combinations
1844 of ;type= and ;mode= URLs that could potentially fail in the Turkish
1847 Daniel Stenberg (20 Jan 2009)
1848 - Lisa Xu pointed out that the ssh.obj file was missing from the
1849 lib/Makefile.vc6 file (and thus from the vc8 and vc9 ones too).
1851 Version 7.19.3 (19 January 2009)
1853 Daniel Stenberg (16 Jan 2009)
1854 - Andrew de los Reyes fixed curlbuild.h for "generic" gcc builds on PPC, both
1857 Daniel Stenberg (15 Jan 2009)
1858 - Tim Ansell fixed a compiler warning in lib/cookie.c
1860 Daniel Stenberg (14 Jan 2009)
1861 - Grant Erickson fixed timeouts for TFTP such that specifying a
1862 connect-timeout, a max-time or both options work correctly and as expected
1863 by passing the correct boolean value to Curl_timeleft via the
1864 'duringconnect' parameter.
1866 With this small change, curl TFTP now behaves as expected (and likely as
1867 originally-designed):
1869 1) For non-existent or unreachable dotted IP addresses:
1871 a) With no options, follows the default curl 300s timeout...
1872 b) With --connect-timeout only, follows that value...
1873 c) With --max-time only, follows that value...
1874 d) With both --connect-timeout and --max-time, follows the smaller value...
1876 and times out with a "curl: (7) Couldn't connect to server" error.
1878 2) For transfers to/from a valid host:
1880 a) With no options, follows default curl 300s timeout for the
1881 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1882 timeout for the remainder of the transfer...
1884 b) With --connect-time only, follows that value for the
1885 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1886 timeout for the remainder of the transfer...
1888 c) With --max-time only, follows that value for the first
1889 XRQ/DATA/ACK transaction and for the remainder of the
1892 d) With both --connect-timeout and --max-time, follows the former
1893 for the first XRQ/DATA/ACK transaction and the latter for the
1894 remainder of the transfer...
1896 and times out with a "curl: (28) Timeout was reached" error as
1899 Daniel Stenberg (13 Jan 2009)
1900 - Michael Wallner fixed a NULL pointer deref when calling
1901 curl_easy_setup(curl, CURLOPT_COOKIELIST, "SESS") on a CURL handle with no
1904 - Stefan Teleman brought a patch to fix the default curlbuild.h file for the
1907 Daniel Stenberg (12 Jan 2009)
1908 - Based on bug report #2498665 (http://curl.haxx.se/bug/view.cgi?id=2498665)
1909 by Daniel Black, I've now added magic to the configure script that makes it
1910 use pkg-config to detect gnutls details as well if the existing method
1911 (using libgnutls-config) fails. While doing this, I cleaned up and unified
1912 the pkg-config usage when detecting openssl and nss as well.
1914 Daniel Stenberg (11 Jan 2009)
1915 - Karl Moerder brought the patch that creates vc9 Makefiles, and I made
1916 'maketgz' now use the actual makefile targets to do the VC8 and VC9
1919 Daniel Stenberg (10 Jan 2009)
1920 - Emil Romanus fixed:
1922 When using the multi interface over HTTP and the server returns a Location
1923 header, the running easy handle will get stuck in the CURLM_STATE_PERFORM
1924 state, leaving the external event loop stuck waiting for data from the
1925 ingoing socket (when using the curl_multi_socket_action stuff). While this
1926 bug was pretty hard to find, it seems to require only a one-line fix. The
1927 break statement on line 1374 in multi.c caused the function to skip the call
1930 How to reproduce this bug? Well, that's another question. evhiperfifo.c in
1931 the examples directory chokes on this bug only _sometimes_, probably
1932 depending on how fast the URLs are added. One way of testing the bug out is
1933 writing to hiper.fifo from more than one source at the same time.
1935 Daniel Fandrich (7 Jan 2009)
1936 - Unified much of the SessionHandle initialization done in Curl_open() and
1937 curl_easy_reset() by creating Curl_init_userdefined(). This had the side
1938 effect of fixing curl_easy_reset() so it now also resets
1939 CURLOPT_FTP_FILEMETHOD and CURLOPT_SSL_SESSIONID_CACHE
1941 Daniel Stenberg (7 Jan 2009)
1942 - Rob Crittenden did once again provide an NSS update:
1944 I have to jump through a few hoops now with the NSS library initialization
1945 since another part of an application may have already initialized NSS by the
1946 time Curl gets invoked. This patch is more careful to only shutdown the NSS
1947 library if Curl did the initialization.
1949 It also adds in a bit of code to set the default ciphers if the app that
1950 call NSS_Init* did not call NSS_SetDomesticPolicy() or set specific
1951 ciphers. One might argue that this lets other application developers get
1952 lazy and/or they aren't using the NSS API correctly, and you'd be right.
1953 But still, this will avoid terribly difficult-to-trace crashes and is
1956 Daniel Stenberg (1 Jan 2009)
1957 - 'reconf' is removed since we rather have users use 'buildconf'
1959 Daniel Stenberg (31 Dec 2008)
1960 - Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
1961 out that 'reconf' didn't properly point out the m4 subdirectory when running
1964 Daniel Stenberg (29 Dec 2008)
1965 - Phil Lisiecki filed bug report #2413067
1966 (http://curl.haxx.se/bug/view.cgi?id=2413067) that identified a problem that
1967 would cause libcurl to mark a DNS cache entry "in use" eternally if the
1968 subsequence TCP connect failed. It would thus never get pruned and refreshed
1969 as it should've been.
1971 Phil provided his own patch to this problem that while it seemed to work
1972 wasn't complete and thus I wrote my own fix to the problem.
1974 Daniel Stenberg (28 Dec 2008)
1975 - Peter Korsgaard fixed building libcurl with "configure --with-ssl
1978 - Anthony Bryan fixed more language and spelling flaws in man pages.
1980 Daniel Stenberg (22 Dec 2008)
1981 - Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
1982 on file indexes beyond 2 or 4GB.
1984 - Anthony Bryan provided a set of patches that cleaned up manual language,
1985 corrected spellings and more.
1987 Daniel Stenberg (20 Dec 2008)
1988 - Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
1989 pipelining, as libcurl could then easily get confused and A) work on the
1990 handle that was not "first in queue" on a pipeline, or even B) tell the app
1991 to REMOVE a socket while it was in use by a second handle in a pipeline. Both
1992 errors caused hanging or stalling applications.
1994 Daniel Stenberg (19 Dec 2008)
1995 - curl_multi_timeout() could return a timeout value of 0 even though nothing
1996 was actually ready to get done, as the internal time resolution is higher
1997 than the returned millisecond timer. Therefore it could cause applications
1998 running on fast processors to do short bursts of busy-loops.
1999 curl_multi_timeout() will now only return 0 if the timeout is actually
2002 - Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
2003 now has an improved ability to do right when the multi interface (both
2004 "regular" and multi_socket) is used for SCP and SFTP transfers. This should
2005 result in (much) less busy-loop situations and thus less CPU usage with no
2008 Daniel Stenberg (17 Dec 2008)
2009 - SCP and SFTP with the multi interface had the same flaw: the 'DONE'
2010 operation didn't complete properly if the EAGAIN equivalent was returned but
2011 libcurl would simply continue with a half-completed close operation
2012 performed. This ruined persistent connection re-use and cause some
2013 SSH-protocol errors in general. The correction is unfortunately adding a
2014 blocking function - doing it entirely non-blocking should be considered for
2017 Gisle Vanem (16 Dec 2008)
2018 - Added the possibility to use the Watt-32 tcp/ip stack under Windows.
2019 The change simply involved adding a USE_WATT32 section in the
2020 config-win32.h files (under ./lib and ./src). This section disables
2021 the use of any Winsock headers.
2023 Daniel Stenberg (16 Dec 2008)
2024 - libssh2_sftp_last_error() was wrongly used at some places in libcurl which
2025 made libcurl sometimes not properly abort problematic SFTP transfers.
2027 Daniel Stenberg (12 Dec 2008)
2028 - More work with Igor Novoseltsev to first fix the remaining stuff for
2029 removing easy handles from multi handles when the easy handle is/was within
2030 a HTTP pipeline. His bug report #2351653
2031 (http://curl.haxx.se/bug/view.cgi?id=2351653) was also related and was
2032 eventually fixed by a patch by Igor himself.
2034 Yang Tse (12 Dec 2008)
2035 - Patrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
2036 OS/400 compilations with IPv6 enabled.
2038 Daniel Stenberg (12 Dec 2008)
2039 - Mark Karpeles filed bug report #2416182 titled "crash in ConnectionExists
2040 when using duphandle+curl_mutli"
2041 (http://curl.haxx.se/bug/view.cgi?id=2416182) which showed that
2042 curl_easy_duphandle() wrongly also copied the pointer to the connection
2043 cache, which was plain wrong and caused a segfault if the handle would be
2044 used in a different multi handle than the handle it was duplicated from.
2046 Daniel Stenberg (11 Dec 2008)
2047 - Keshav Krity found out that libcurl failed to deal with dotted IPv6
2048 addresses if they were very long (>39 letters) due to a too strict address
2049 validity parser. It now accepts addresses up to 45 bytes long.
2051 Daniel Stenberg (11 Dec 2008)
2052 - Internet Explorer had a broken HTTP digest authentication before v7 and
2053 there are servers "out there" that relies on the client doing this broken
2054 Digest authentication. Apache even comes with an option to work with such
2057 The difference is only for URLs that contain a query-part (a '?'-letter and
2058 text to the right of it).
2060 libcurl now supports this quirk, and you enable it by setting the
2061 CURLAUTH_DIGEST_IE bit in the bitmask you pass to the CURLOPT_HTTPAUTH or
2062 CURLOPT_PROXYAUTH options. They are thus individually controlled to server
2065 (note that there's no way to activate this with the curl tool yet)
2067 Daniel Fandrich (9 Dec 2008)
2068 - Added test cases 1089 and 1090 to test --write-out after a redirect to
2069 test a report that the size didn't work, but these test cases pass.
2071 - Documented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.
2073 Daniel Stenberg (9 Dec 2008)
2074 - Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
2075 particular state for the control connection like it did before for implicit
2076 FTPS (libcurl assumed such control connections to be encrypted while some
2077 FTPS servers such as FileZilla assumes such connections to be clear
2078 mode). Use the CURLOPT_USE_SSL option to set your desired level.
2080 Daniel Stenberg (8 Dec 2008)
2081 - Fred Machado posted about a weird FTP problem on the curl-users list and when
2082 researching it, it turned out he got a 550 response back from a SIZE command
2083 and then I fell over the text in RFC3659 that says:
2085 The presence of the 550 error response to a SIZE command MUST NOT be taken
2086 by the client as an indication that the file cannot be transferred in the
2087 current MODE and TYPE.
2089 In other words: the change I did on September 30th 2008 and that has been
2090 included in the last two releases were a regression and a bad idea. We MUST
2091 NOT take a 550 response from SIZE as a hint that the file doesn't exist.
2093 - Christian Krause filed bug #2221237
2094 (http://curl.haxx.se/bug/view.cgi?id=2221237) that identified an infinite
2095 loop during GSS authentication given some specific conditions. With his
2096 patience and great feedback I managed to narrow down the problem and
2097 eventually fix it although I can't test any of this myself!
2099 Daniel Fandrich (3 Dec 2008)
2100 - Fixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
2101 support (e.g. Minix)
2103 Daniel Stenberg (3 Dec 2008)
2104 - Igor Novoseltsev filed bug #2351645
2105 (http://curl.haxx.se/bug/view.cgi?id=2351645) that identified a problem with
2106 the multi interface that occured if you removed an easy handle while in
2107 progress and the handle was used in a HTTP pipeline.
2109 - Pawel Kierski pointed out a mistake in the cookie code that could lead to a
2110 bad fclose() after a fatal error had occured.
2111 (http://curl.haxx.se/bug/view.cgi?id=2382219)
2113 Daniel Fandrich (25 Nov 2008)
2114 - If a HTTP request is Basic and num is already >=1000, the HTTP test
2115 server adds 1 to num to get the data section to return. This allows
2116 testing authentication negotiations using the Basic authentication
2119 - Added tests 1087 and 1088 to test Basic authentication on a redirect
2120 with and without --location-trusted
2122 Daniel Stenberg (24 Nov 2008)
2123 - Based on a patch by Vlad Grachov, libcurl now uses a new libssh2 0.19
2124 function when built to support SCP and SFTP that helps the library to know
2125 in which direction a particular libssh2 operation would return EAGAIN so
2126 that libcurl knows what socket conditions to wait for before trying the
2127 function call again. Previously (and still when using libssh2 0.18 or
2128 earlier), libcurl will busy-loop in this situation when the easy interface
2131 Daniel Fandrich (20 Nov 2008)
2132 - Automatically detect OpenBSD's CA cert bundle.
2134 Daniel Stenberg (19 Nov 2008)
2135 - I removed the default use of "Pragma: no-cache" from libcurl when a proxy is
2136 used. It has been used since forever but it was never a good idea to use
2137 unless explicitly asked for.
2139 - Josef Wolf's extension that allows a $TESTDIR/gdbinit$testnum file that when
2140 you use runtests.pl -g, will be sourced by gdb to allow additional fancy or
2141 whatever you see fit
2143 - Christian Krause reported and fixed a memory leak that would occur with HTTP
2144 GSS/kerberos authentication (http://curl.haxx.se/bug/view.cgi?id=2284386)
2146 - Andreas Wurf and Markus Koetter helped me analyze a problem that Andreas got
2147 when uploading files to a single FTP server using multiple easy handle
2148 handles with the multi interface. Occasionally a handle would stall in
2151 The problem turned out to be a side-effect of the ConnectionExists()
2152 function's eagerness to re-use a handle for HTTP pipelining so it would
2153 select it even if already being in use, due to an inadequate check for its
2154 chances of being used for pipelnining.
2156 Daniel Fandrich (17 Nov 2008)
2157 - Added more compiler warning options for gcc 4.3
2159 Yang Tse (17 Nov 2008)
2160 - Fix a remaining problem in the inet_pton() runtime configure check. And
2161 fix internal Curl_inet_pton() failures to reject certain malformed literals.
2163 - Make configure script check if ioctl with the SIOCGIFADDR command can be
2164 used, and define HAVE_IOCTL_SIOCGIFADDR if appropriate.
2166 Daniel Stenberg (16 Nov 2008)
2167 - Christian Krause fixed a build failure when building with gss support
2168 enabled and FTP disabled.
2170 - Added check for NULL returns from strdup() in src/main.c and lib/formdata.c
2171 - reported by Jim Meyering also prevent buffer overflow on MSDOS when you do
2172 for example -O on a url with a file name part longer than PATH_MAX letters
2174 - lib/nss.c fixes based on the report by Jim Meyering: I went over and added
2175 checks for return codes for all calls to malloc and strdup that were
2176 missing. I also changed a few malloc(13) to use arrays on the stack and a
2177 few malloc(PATH_MAX) to instead use aprintf() to lower memory use.
2179 - I fixed a memory leak in Curl_nss_connect() when CURLOPT_ISSUERCERT is
2182 Daniel Fandrich (14 Nov 2008)
2183 - Added .xml as one of the few common file extensions known by the multipart
2186 - Added some #ifdefs around header files and change the EAGAIN test to
2187 fix compilation on Cell (reported by Jeff Curley).
2189 Yang Tse (14 Nov 2008)
2190 - Fixed several configure script issues affecting checks for inet_ntoa_r(),
2191 inet_ntop(), inet_pton(), getifaddrs(), fcntl() and getaddrinfo().
2193 Yang Tse (13 Nov 2008)
2194 - Refactored configure script detection of functions used to set sockets into
2195 non-blocking mode, and decouple function detection from function capability.