5 \___|\___/|_| \_\_____|
9 Daniel Stenberg (20 May 2010)
10 - Tanguy Fautre brought a fix to allow curl to build with Microsoft VC10.
12 Daniel Stenberg (18 May 2010)
13 - Eric Mertens posted bug report #3003005 pointing out that the libcurl TFTP
14 code was not sending the timeout option properly to the server, and
17 (http://curl.haxx.se/bug/view.cgi?id=3003005)
19 Kamil Dudka (16 May 2010)
20 - Pavel Raiskup introduced a new option CURLOPT_FNMATCH_DATA in order to pass
21 a custom data pointer to the callback specified by CURLOPT_FNMATCH_FUNCTION.
23 Daniel Stenberg (14 May 2010)
24 - John-Mark Bell filed bug #3000052 that identified a problem (with an
25 associated patch) with the OpenSSL handshake state machine when the multi
28 Performing an https request using a curl multi handle and using select or
29 epoll to wait for events results in a hang. It appears that the cause is the
30 fix for bug #2958179, which makes ossl_connect_common unconditionally return
31 from the step 2 loop when fetching from a multi handle.
33 When ossl_connect_step2 has completed, it updates connssl->connecting_state
34 to ssl_connect_3. ossl_connect_common will then return to the caller, as a
35 multi handle is in use. Eventually, the client code will call
36 curl_multi_fdset to obtain an updated fdset to select or epoll on. For https
37 requests, curl_multi_fdset will cause https_getsock to be called.
38 https_getsock will only return a socket handle if the connecting_state is
39 ssl_connect_2_reading or ssl_connect_2_writing. Therefore, the client will
40 never obtain a valid fdset, and thus not drive the multi handle, resulting
43 (http://curl.haxx.se/bug/view.cgi?id=3000052)
45 - Sebastian V reported bug #3000056 identifying a problem with redirect
46 following. It showed that when curl followed redirects it didn't properly
47 ignore the response body of the 30X response if that response was using
48 compressed Content-Encoding!
50 (http://curl.haxx.se/bug/view.cgi?id=3000056)
52 Daniel Stenberg (12 May 2010)
53 - Howard Chu brought support for RTMP. This is powered by the underlying
54 librtmp library. It supports a range of variations and "sub-protocols"
55 within the RTMP family.
57 - Pavel Raiskup brought support for FTP directory wildcard matching to allow
58 selective downloading. To provide that, a set of new options were added:
61 CURLOPT_CHUNK_BGN_FUNCTION
62 CURLOPT_CHUNK_END_FUNCTION
64 CURLOPT_FNMATCH_FUNCTION
66 There were also a set of new tests added (574 - 577) to verify this.
68 Kamil Dudka (11 May 2010)
69 - CRL support in libcurl-NSS has been completely broken. Now it works. Original
70 bug report: https://bugzilla.redhat.com/581926
72 Daniel Stenberg (7 May 2010)
73 - Dirk Manske reported a regression. When connecting with the multi interface,
74 there were situations where libcurl wouldn't store connect time correctly as
75 it used to (and is documented to) do.
77 Using his fine sample program we could repeat it, and I wrote up test case
78 573 using that code. The problem does not easily show itself using the local
81 The fix, also as suggested by Dirk, is a bit on the ugly side as it adds yet
82 another call to Curl_verboseconnect() and setting the TIMER_CONNECT time.
83 That situation is subject for some closer inspection in the future.
85 - Howard Chu split the I/O handling functions into private handlers.
87 Howard Chu brought the bulk work of this patch that properly moves out the
88 sending and recving of data to the parts of the code that are properly
89 responsible for the various ways of doing so.
91 Daniel Stenberg assisted with polishing a few bits and fixed some minor
92 flaws in the original patch.
94 Another upside of this patch is that we now abuse CURLcodes less with the
95 "magic" -1 return codes and instead use CURLE_AGAIN more consistently.
97 Daniel Stenberg (5 May 2010)
98 - Hoi-Ho Chan introduced support for using the PolarSSL library. You control
99 this with the new configure option --with-polarssl.
101 Daniel Stenberg (29 Apr 2010)
102 - Ben Greear made telnet a lot better/easier to use by an application:
104 The main change is to allow input from user-specified methods, when they are
105 specified with CURLOPT_READFUNCTION. All calls to fflush(stdout) in
106 telnet.c were removed, which makes using 'curl telnet://foo.com' painful
107 since prompts and other data are not always returned to the user promptly.
108 Use 'curl --no-buffer telnet://foo.com' instead. In general, the user
109 should have their CURLOPT_WRITEFUNCTION do a fflush for interactive use.
111 Also fix assumption that reading from stdin never returns < 0.
112 Old code could crash in that case.
114 Call progress functions in telnet main loop.
116 Daniel Stenberg (26 Apr 2010)
117 - Make use of the libssh2_init/exit functions that libssh2 added in version
118 1.2.5. Using them will improve how libcurl works in threaded situations when
119 SCP and SFTP are transfered.
121 Daniel Stenberg (25 Apr 2010)
122 - Based on work by Kamil Dudka, I've introduced the new configure option
123 --enable-threaded-resolver. When used, the configure script will check for
124 pthreads and if around, it will build libcurl to use pthreads to do name
125 resolving in a threaded manner. Note that this is just a fix to offer an
126 option that can enable the code that already included. The threader resolver
127 code was mostly added on Jan 26 2010.
129 Daniel Stenberg (24 Apr 2010)
130 - Alex Bligh introduced the --proto and -proto-redir options that limit what
131 protocols curl accepts for the requests and when following redirects.
133 Kamil Dudka (24 Apr 2010)
134 - Fixed test536 in order to not fail with threaded DNS resolver and tweaked
135 comments in certain examples using curl_multi_fdset().
137 - Fixed SSL handshake timeout underflow in libcurl-NSS, which caused test405
138 to hang on a slow machine.
140 Daniel Stenberg (21 Apr 2010)
141 - The -O option caused curl to crash on windows and DOS due to the tool
142 writing out of boundary memory.
144 Yang Tse (20 Apr 2010)
145 - Ruslan Gazizov detected that MSVC makefiles were using wsock32.lib instead
146 of ws2_32.lib, this generated linking issues on MSVC IPv6 enabled builds
147 that were done using those makefiles.
149 Daniel Stenberg (19 Apr 2010)
150 - -J/--remote-header-name didn't strip trailing carriage returns or linefeeds
151 properly, so they could be used in the file name.
153 Daniel Stenberg (16 Apr 2010)
154 - Jerome Vouillon made the GnuTLS SSL handshake phase non-blocking.
156 - The recent overhaul of the SSL recv function made the GnuTLS specific code
157 treat a zero returned from gnutls_record_recv() as an error, and this caused
158 our HTTPS test cases to fail. We leave it to upper layer code to detect if
159 an EOF is a problem or not.
161 - I reverted the resolver fix from yesterday and instead removed all uses of
162 AI_CANONNAME all over libcurl and made the only user of that info (krb5.c)
163 use the host name from the URL instead. No reverse resolving is a good
166 - Paul Howarth made configure properly detect GSS "on ancient Linux distros"
167 by editing in which order we use headers to detect GSS.
169 Daniel Stenberg (15 Apr 2010)
170 - Rainer Canavan filed bug report #2987196 that identified libcurl doing
171 unnecesary reverse name lookups in many cases when built to use IPv4 and
172 getaddrinfo(). The logic for ipv6 is now used for ipv4 too.
174 (http://curl.haxx.se/bug/view.cgi?id=2963679)
176 Version 7.20.1 (14 April 2010)
178 Daniel Stenberg (9 Apr 2010)
179 - Prefixing the FTP quote commands with an asterisk really only worked for the
180 postquote actions. This is now fixed and test case 227 has been extended to
183 Kamil Dudka (4 Apr 2010)
184 - Eliminated a race condition in Curl_resolv_timeout().
186 - Refactorized interface of Curl_ssl_recv()/Curl_ssl_send().
188 - libcurl-NSS now provides more accurate messages and error codes in case of
189 client certificate problem. Either during connection, or transfer phase.
191 Daniel Stenberg (1 Apr 2010)
192 - Matt Wixson found and fixed a bug in the SCP/SFTP area where the code
193 treated a 0 return code from libssh2 to be the same as EAGAIN while in
194 reality it isn't. The problem caused a hang in SFTP transfers from a
197 Daniel Stenberg (28 Mar 2010)
198 - Ben Greear: If you pass a URL to pop3 that does not contain a message ID as
199 part of the URL, it would previously ask for 'INBOX' which just causes the
200 pop3 server to return an error.
202 Now libcurl treats en empty message ID as a request for LIST (list of pop3
203 message IDs). User's code could then parse this and download individual
206 Daniel Stenberg (27 Mar 2010)
207 - Ben Greear brought a patch that from now on allows all protocols to specify
208 name and user within the URL, in the same manner HTTP and FTP have been
209 allowed to in the past - although far from all of the libcurl supported
210 protocls actually have that feature in their URL definition spec.
212 Daniel Stenberg (26 Mar 2010)
213 - Ben Greear brought code that makes the rate limiting code for the easy
214 interface a bit smoother as it introduces sub-second sleeps during it and it
215 also takes the buffer sizes into account.
217 Daniel Stenberg (24 Mar 2010)
218 - Bob Richmond: There's an annoying situation where libcurl will read new HTTP
219 response data from a socket, then check if it's a timeout if one is set. If
220 the last packet received constitutes the end of the response body, libcurl
221 still treats it as a timeout condition and reports a message like:
223 "Operation timed out after 3000 milliseconds with 876 out of 876 bytes
226 It should only a timeout if the timer lapsed and we DIDN'T receive the end
227 of the response body yet.
229 - Christopher Conroy fixed a problem with RTSP and GET_PARAMETER reported
230 to us by Massimo Callegari. There's a new test case 572 that verifies this
233 - The 'ares' subtree has been removed from the source repository. It was
234 always a separate project that sort of piggybacked on the curl project since
235 the dawn of times and now the time has come for it to go stand on its own
236 legs and continue living its own life. All details on c-ares and its new
237 source code repository is found at http://c-ares.haxx.se/
239 Daniel Stenberg (23 Mar 2010)
240 - Kenny To filed the bug report #2963679 with patch to fix a problem he
241 experienced with doing multi interface HTTP POST over a proxy using
242 PROXYTUNNEL. He found a case where it would connect fine but bits.tcpconnect
243 was not set correct so libcurl didn't work properly.
245 (http://curl.haxx.se/bug/view.cgi?id=2963679)
247 - Akos Pasztory filed debian bug report #572276
248 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=572276 mentioning a problem
249 with a resource that returns chunked-encoded _and_ with a Content-Length
250 and libcurl failed to properly ignore the latter information.
252 - Hauke Duden provided an example program that made the multi interface crash.
253 His example simply used the multi interface and did first one FTP transfer
254 and after completion it used a second easy handle and did another FTP
255 transfer on the same FTP server.
257 This triggered a bug in the "delayed easy handle kill" system that curl
258 uses: when an FTP connection is left alive it must keep an easy handle
259 around internally - only for the purpose of having an easy handle when it
260 later disconnects it. The code assumed that when the easy handle was removed
261 and an internal reference was made, that version could be killed later on
262 when a new easy handle came using the same connection. This was wrong as
263 Hauke's example showed that the removed handle wasn't killed for real until
264 later. This caused a double close attempt => segfault.
266 Daniel Stenberg (22 Mar 2010)
267 - Thomas Lopatic fixed the alarm()-based DNS timeout:
269 Looking at the code of Curl_resolv_timeout() in hostip.c, I think that in
270 case of a timeout, the signal handler for SIGALRM never gets removed. I
271 think that in my case it gets executed at some point later on when execution
272 has long left Curl_resolv_timeout() or even the cURL library.
274 The code that is jumped to with siglongjmp() simply sets the error message
275 to "name lookup timed out" and then returns with CURLRESOLV_ERROR. I guess
276 that instead of simply returning without cleaning up, the code should have a
277 goto that jumps to the spot right after the call to Curl_resolv().
279 Kamil Dudka (22 Mar 2010)
280 - Douglas Steinwand contributed a patch fixing insufficient initialization in
281 Curl_clone_ssl_config()
283 Daniel Stenberg (21 Mar 2010)
284 - Ben Greear improved TFTP: the error code returning and the treatment
285 of TSIZE == 0 when uploading.
287 - We've switched from CVS to git. See http://curl.haxx.se/source.html
289 Kamil Dudka (19 Mar 2010)
290 - Improved Curl_read() to not ignore the error returned from Curl_ssl_recv().
292 Daniel Stenberg (15 Mar 2010)
293 - Constantine Sapuntzakis brought a patch:
295 The problem mentioned on Dec 10 2009
296 (http://curl.haxx.se/bug/view.cgi?id=2905220) was only partially fixed.
297 Partially because an easy handle can be associated with many connections in
298 the cache (e.g. if there is a redirect during the lifetime of the easy
299 handle). The previous patch only cleaned up the first one. The new fix now
300 removes the easy handle from all connections, not just the first one.
302 Daniel Stenberg (6 Mar 2010)
303 - Ben Greear brought a patch that fixed the rate limiting logic for TFTP when
304 the easy interface was used.
306 Daniel Stenberg (5 Mar 2010)
307 - Daniel Johnson provided fixes for building curl with the clang compiler.
309 Yang Tse (5 Mar 2010)
310 - Constantine Sapuntzakis detected and fixed a double free in builds done
311 with threaded resolver enabled (Windows default configuration) that would
312 get triggered when a curl handle is closed while doing DNS resolution.
314 Daniel Stenberg (2 Mar 2010)
315 - [Daniel Johnson] I've been trying to build libcurl with clang on Darwin and
316 ran into some issues with the GSSAPI tests in configure.ac. The tests first
317 try to determine the include dirs and libs and set CPPFLAGS and LIBS
318 accordingly. It then checks for the headers and finally sets LIBS a second
319 time, causing the libs to be included twice. The first setting of LIBS seems
320 redundant and should be left out, since the first part is otherwise just
321 about finding headers.
323 My second issue is that 'krb5-config --libs gssapi' on Darwin is less than
324 useless and returns junk that, while it happens to work with gcc, causes
325 clang to choke. For example, --libs returns $CFLAGS along with the libs,
326 which is really retarded. Simply setting 'LIBS="$LIBS -lgssapi_krb5
327 -lresolv"' on Darwin is sufficient.
329 - Based on patch provided by Jacob Moshenko, the transfer logic now properly
330 makes sure that when using sub-second timeouts, there's no final bad 1000ms
331 wait. Previously, a sub-second timeout would often make the elapsed time end
332 up the time rounded up to the nearest second (e.g. 1s for 200ms timeout)
334 - Andrei Benea filed bug report #2956698 and pointed out that the
335 CURLOPT_CERTINFO feature leaked memory due to a missing OpenSSL function
336 call. He provided the patch to fix it too.
338 http://curl.haxx.se/bug/view.cgi?id=2956698
340 - Markus Duft pointed out in bug #2961796 that even though Interix has a
341 poll() function it doesn't quite work the way we want it so we must disable
342 it, and he also provided a patch for it.
344 http://curl.haxx.se/bug/view.cgi?id=2961796
346 - Made the pingpong timeout code properly deal with the response timeout AND
347 the global timeout if set. Also, as was reported in the bug report #2956437
348 by Ryan Chan, the time stamp to use as basis for the per command timeout was
349 not set properly in the DONE phase for FTP (and not for SMTP) so I fixed
350 that just now. This was a regression compared to 7.19.7 due to the
351 conversion of FTP code over to the generic pingpong concepts.
353 http://curl.haxx.se/bug/view.cgi?id=2956437
355 Daniel Stenberg (1 Mar 2010)
356 - Ben Greear provided an update for TFTP that fixes upload.
358 - Wesley Miaw reported bug #2958179 which identified a case of looping during
359 OpenSSL based SSL handshaking even though the multi interface was used and
360 there was no good reason for it.
362 http://curl.haxx.se/bug/view.cgi?id=2958179
364 Daniel Stenberg (26 Feb 2010)
365 - Pat Ray in bug #2958474 pointed out an off-by-one case when receiving a
366 chunked-encoding trailer.
368 http://curl.haxx.se/bug/view.cgi?id=2958474
370 Daniel Fandrich (25 Feb 2010)
371 - Fixed a couple of out of memory leaks and a segfault in the SMTP & IMAP code.
373 Yang Tse (25 Feb 2010)
374 - I fixed bug report #2958074 indicating
375 (http://curl.haxx.se/bug/view.cgi?id=2958074) that curl on Windows with
376 option --trace-time did not use local time when timestamping trace lines.
377 This could also happen on other systems depending on time souurce.
379 Patrick Monnerat (22 Feb 2010)
380 - Proper handling of STARTTLS on SMTP, taking CURLUSESSL_TRY into account.
381 - SMTP falls back to RFC821 HELO when EHLO fails (and SSL is not required).
382 - Use of true local host name (i.e.: via gethostname()) when available, as
383 default argument to SMTP HELO/EHLO.
384 - Test case 804 for HELO fallback.
386 Daniel Stenberg (20 Feb 2010)
387 - Fixed the SMTP compliance by making sure RCPT TO addresses are specified
388 properly in angle brackets. Recipients provided with CURLOPT_MAIL_RCPT now
389 get angle bracket wrapping automatically by libcurl unless the recipient
390 starts with an angle bracket as then the app is assumed to deal with that
393 - I made the SMTP code expect a 250 response back from the server after the
394 full DATA has been sent, and I modified the test SMTP server to also send
395 that response. As usual, the DONE operation that is made after a completed
396 transfer is still not doable in a non-blocking way so this waiting for 250
397 is unfortunately made blockingly.
399 Yang Tse (14 Feb 2010)
400 - Overhauled test suite getpart() function. Fixing potential out of bounds
401 stack and memory overwrites triggered with huge test case definitions.
403 Daniel Stenberg (13 Feb 2010)
404 - Martin Hager reported and fixed a problem with a missing quote in libcurl.m4
406 (http://curl.haxx.se/bug/view.cgi?id=2951319)
408 - Tom Donovan fixed the CURL_FORMAT_* defines when building with cmake.
410 (http://curl.haxx.se/bug/view.cgi?id=2951269)
412 Daniel Stenberg (12 Feb 2010)
413 - Jack Zhang reported a problem with SMTP: we wrongly used multiple addresses
414 in the same RCPT TO line, when they should be sent in separate single
415 commands. I updated test case 802 to verify this.
417 - I also fixed a bad use of my_setopt_str() of CURLOPT_MAIL_RCPT in the curl
418 tool which made it try to output it as string for the --libcurl feature
419 which could lead to crashes.
421 Yang Tse (11 Feb 2010)
422 - Steven M. Schweda fixed VMS builder bad behavior when used in a batch job,
423 removed obsolete batch_compile.com and defines.com and updated VMS readme.
425 Version 7.20.0 (9 February 2010)
427 Daniel Stenberg (9 Feb 2010)
428 - When downloading compressed content over HTTP and the app asked libcurl to
429 automatically uncompress it with the CURLOPT_ENCODING option, libcurl could
430 wrongly provide the callback with more data than the maximum documented
431 amount. An application could thus get tricked into badness if the maximum
432 limit was trusted to be enforced by libcurl itself (as it is documented).
434 This is further detailed and explained in the libcurl security advisory
437 http://curl.haxx.se/docs/adv_20100209.html
439 Daniel Fandrich (3 Feb 2010)
440 - Changed the Watcom makefiles to make them easier to keep in sync with
441 Makefile.inc since that can't be included directly.
443 Yang Tse (2 Feb 2010)
444 - Symbol CURL_FORMAT_OFF_T now obsoleted, will be removed in a future release,
445 symbol will not be available when building with CURL_NO_OLDIES defined. Use
446 of CURL_FORMAT_CURL_OFF_T is preferred since 7.19.0
448 Daniel Stenberg (1 Feb 2010)
449 - Using the multi_socket API, it turns out at times it seemed to "forget"
450 connections (which caused a hang). It turned out to be an existing (7.19.7)
451 bug in libcurl (that's been around for a long time) and it happened like
454 The app calls curl_multi_add_handle() to add a new easy handle, libcurl will
455 then set it to timeout in 1 millisecond so libcurl will tell the app about
458 The app's timeout fires off that there's a timeout, the app calls libcurl as
459 we so often document it:
462 res = curl_multi_socket_action(... TIMEOUT ...);
463 } while(CURLM_CALL_MULTI_PERFORM == res);
465 And this is the problem number one:
467 When curl_multi_socket_action() is called with no specific handle, but only
468 a timeout-action, it will *only* perform actions within libcurl that are
469 marked to run at this time. In this case, the request would go from INIT to
470 CONNECT and return CURLM_CALL_MULTI_PERFORM. When the app then calls libcurl
471 again, there's no timer set for this handle so it remains in the CONNECT
472 state. The CONNECT state is a transitional state in libcurl so it reports no
473 sockets there, and thus libcurl never tells the app anything more about that
474 easy handle/connection.
476 libcurl _does_ set a 1ms timeout for the handle at the end of
477 multi_runsingle() if it returns CURLM_CALL_MULTI_PERFORM, but since the loop
478 is instant the new job is not ready to run at that point (and there's no
479 code that makes libcurl call the app to update the timout for this new
480 timeout). It will simply rely on that some other timeout will trigger later
481 on or that something else will update the timeout callback. This makes the
482 bug fairly hard to repeat.
484 The fix made to adress this issue:
486 We introduce a loop in lib/multi.c around all calls to multi_runsingle() and
487 simply check for CURLM_CALL_MULTI_PERFORM internally. This has the added
488 benefit that this goes in line with my long-term wishes to get rid of the
489 CURLM_CALL_MULTI_PERFORM all together from the public API.
491 The downside of this fix, is that the counter we return in 'running_handles'
492 in several of our public functions then gets a slightly new and possibly
493 confusing behavior during times:
495 If an app adds a handle that fails to connect (very quickly) it may just
496 as well never appear as a 'running_handle' with this fix. Previously it
497 would first bump the counter only to get it decreased again at next call.
498 Even I have used that change in handle counter to signal "end of a
499 transfer". The only *good* way to find the end of a individual transfer
500 is calling curl_multi_info_read() to see if it returns one.
502 Of course, if the app previously did the looping before it checked the
503 counter, it really shouldn't be any new effect.
505 Yang Tse (26 Jan 2010)
506 - Constantine Sapuntzakis' and Joshua Kwan's work done in the last four months
507 relative to the asynchronous DNS lookups, along with with some integration
508 adjustments I have done are finally committed to CVS.
510 Currently these enhancements will benefit builds done using c-ares on any
511 platform as well as Windows builds using the default threaded resolver.
513 This release does not make generally available POSIX threaded DNS lookups
514 yet. There is no configure option to enable this feature yet. It is possible
515 to experimantally try this feature running configure with compiler flags that
516 make simultaneous definition of preprocessor symbols USE_THREADS_POSIX and
517 HAVE_PTHREAD_H, as well as whatever reentrancy compiler flags and linker ones
518 are required to link and properly use pthread_* functions on each platform.
520 Daniel Stenberg (26 Jan 2010)
521 - Mike Crowe made libcurl return CURLE_COULDNT_RESOLVE_PROXY when it is the
522 proxy that cannot be resolved when using c-ares. This matches the behaviour
523 when not using c-ares.
525 Björn Stenberg (23 Jan 2010)
526 - Added a new flag: -J/--remote-header-name. This option tells the
527 -O/--remote-name option to use the server-specified Content-Disposition
528 filename instead of extracting a filename from the URL.
530 Daniel Stenberg (21 Jan 2010)
531 - Chris Conroy brought support for RTSP transfers, and with it comes 8(!) new
532 libcurl options for controlling what to get and how to receive posssibly
533 interleaved RTP data.
535 Daniel Stenberg (20 Jan 2010)
536 - As was pointed out on the http-state mailing list, the order of cookies in a
537 HTTP Cookie: header _needs_ to be sorted on the path length in the cases
538 where two cookies using the same name are set more than once using
539 (overlapping) paths. Realizing this, identically named cookies must be
540 sorted correctly. But detecting only identically named cookies and take care
541 of them individually is harder than just to blindly and unconditionally sort
542 all cookies based on their path lengths. All major browsers also already do
543 this, so this makes our behavior one step closer to them in the cookie area.
545 Test case 8 was the only one that broke due to this change and I updated it
548 Daniel Stenberg (19 Jan 2010)
549 - David McCreedy brought a fix and a new test case (129) to make libcurl work
550 again when downloading files over FTP using ASCII and it turns out that the
551 final size of the file is not the same as the initial size the server
552 reported. This is very common since servers don't take the newline
553 conversions into account.
555 Kamil Dudka (14 Jan 2010)
556 - Suppressed side effect of OpenSSL configure checks, which prevented NSS from
557 being properly detected under certain circumstances. It had been caused by
558 strange behavior of pkg-config when handling PKG_CONFIG_LIBDIR. pkg-config
559 distinguishes among empty and non-existent environment variable in that case.
561 Daniel Stenberg (12 Jan 2010)
562 - Gil Weber reported a peculiar flaw with the multi interface when doing SFTP
563 transfers: curl_multi_fdset() would return -1 and not set and file
564 descriptors several times during a transfer of a single file. It turned out
565 to be due to two different flaws now fixed. Gil's excellent recipe helped me
568 Daniel Stenberg (11 Jan 2010)
569 - Made sure that the progress callback is repeatedly called at a regular
570 interval even during very slow connects.
572 - The tests/runtests.pl script now checks to see if the test case that runs is
573 present in the tests/data/Makefile.am and outputs a notice message on the
574 screen if not. Each test file has to be included in that Makefile.am to get
575 included in release archives and forgetting to add files there is a common
576 mistake. This is an attempt to make it harder to forget.
578 Daniel Stenberg (9 Jan 2010)
579 - Johan van Selst found and fixed a OpenSSL session ref count leak:
581 ossl_connect_step3() increments an SSL session handle reference counter on
582 each call. When sessions are re-used this reference counter may be
583 incremented many times, but it will be decremented only once when done (by
584 Curl_ossl_session_free()); and the internal OpenSSL data will not be freed
585 if this reference count remains positive. When a session is re-used the
586 reference counter should be corrected by explicitly calling
587 SSL_SESSION_free() after each consecutive SSL_get1_session() to avoid
588 introducing a memory leak.
590 (http://curl.haxx.se/bug/view.cgi?id=2926284)
592 Daniel Stenberg (7 Jan 2010)
593 - Make sure the progress callback is called repeatedly even during very slow
594 name resolves when c-ares is used for resolving.
596 Claes Jakobsson (6 Jan 2010)
597 - Julien Chaffraix fixed so that the fragment part in an URL is not sent
598 to the server anymore.
600 Kamil Dudka (3 Jan 2010)
601 - Julien Chaffraix eliminated a duplicated initialization in singlesocket().
603 Daniel Stenberg (2 Jan 2010)
604 - Make curl support --ssl and --ssl-reqd instead of the previous FTP-specific
605 versions --ftp-ssl and --ftp-ssl-reqd as these options are now used to
606 control SSL/TLS for IMAP, POP3 and SMTP as well in addition to FTP. The old
607 option names are still working but the new ones are the ones listed and
610 Daniel Stenberg (1 Jan 2010)
611 - Ingmar Runge enhanced libcurl's FTP engine to support the PRET command. This
612 command is a special "hack" used by the drftpd server, but even though it is
613 a custom extension I've deemed it fine to add to libcurl since this server
614 seems to survive and people keep using it and want libcurl to support
615 it. The new libcurl option is named CURLOPT_FTP_USE_PRET, and it is also
616 usable from the curl tool with --ftp-pret. Using this option on a server
617 that doesn't support this command will make libcurl fail.
619 I added test cases 1107 and 1108 to verify the functionality.
621 The PRET command is documented at
622 http://www.drftpd.org/index.php/Distributed_PASV
624 Yang Tse (30 Dec 2009)
625 - Steven M. Schweda improved VMS build system, and Craig A. Berry helped
626 with the patch and testing.
628 Daniel Stenberg (26 Dec 2009)
629 - Renato Botelho and Peter Pentchev brought a patch that makes the libcurl
630 headers work correctly even on FreeBSD systems before v8.
632 (http://curl.haxx.se/bug/view.cgi?id=2916915)
634 Daniel Stenberg (17 Dec 2009)
635 - David Byron fixed Curl_ossl_cleanup to actually call ENGINE_cleanup when
638 - Follow-up fix for the proxy fix I did for Jon Nelson's bug. It turned out I
639 was a bit too quick and broke test case 1101 with that change. The order of
640 some of the setups is sensitive. I now changed it slightly again to make
641 sure we do them in this order:
643 1 - parse URL and figure out what protocol is used in the URL
644 2 - prepend protocol:// to URL if missing
645 3 - parse name+password off URL, which needs to know what protocol is used
646 (since only some allows for name+password in the URL)
647 4 - figure out if a proxy should be used set by an option
648 5 - if no proxy option, check proxy environment variables
649 6 - run the protocol-specific setup function, which needs to have the proxy
652 Daniel Stenberg (15 Dec 2009)
653 - Jon Nelson found a regression that turned out to be a flaw in how libcurl
654 detects and uses proxies based on the environment variables. If the proxy
655 was given as an explicit option it worked, but due to the setup order
656 mistake proxies would not be used fine for a few protocols when picked up
657 from '[protocol]_proxy'. Obviously this broke after 7.19.4. I now also added
658 test case 1106 that verifies this functionality.
660 (http://curl.haxx.se/bug/view.cgi?id=2913886)
662 Daniel Stenberg (12 Dec 2009)
663 - IMAP, POP3 and SMTP support and their TLS versions (including IMAPS, POP3S
664 and SMTPS) are now supported. The current state may not yet be solid, but
665 the foundation is in place and the test suite has some initial support for
666 these protocols. Work will now persue to make them nice libcurl citizens
669 The work with supporting these new protocols was sponsored by
670 networking4all.com - thanks!
672 Daniel Stenberg (10 Dec 2009)
673 - Siegfried Gyuricsko found out that the curl manual said --retry would retry
674 on FTP errors in the transient 5xx range. Transient FTP errors are in the
675 4xx range. The code itself only tried on 5xx errors that occured _at login_.
676 Now the retry code retries on all FTP transfer failures that ended with a
679 (http://curl.haxx.se/bug/view.cgi?id=2911279)
681 - Constantine Sapuntzakis figured out a case which would lead to libcurl
682 accessing alredy freed memory and thus crash when using HTTPS (with
683 OpenSSL), multi interface and the CURLOPT_DEBUGFUNCTION and a certain order
684 of cleaning things up. I fixed it.
686 (http://curl.haxx.se/bug/view.cgi?id=2905220)
688 Daniel Stenberg (7 Dec 2009)
689 - Martin Storsjo made libcurl use the Expect: 100-continue header for posts
690 with unknown size. Previously it was only used for posts with a known size
691 larger than 1024 bytes.
693 Daniel Stenberg (1 Dec 2009)
694 - If the Expect: 100-continue header has been set by the application through
695 curl_easy_setopt with CURLOPT_HTTPHEADER, the library should set
696 data->state.expect100header accordingly - the current code (in 7.19.7 at
697 least) doesn't handle this properly. Martin Storsjo provided the fix!
699 Yang Tse (28 Nov 2009)
700 - Added Diffie-Hellman parameters to several test harness certificate files in
701 PEM format. Required by several stunnel versions used by our test harness.
703 Daniel Stenberg (28 Nov 2009)
704 - Markus Koetter provided a polished and updated version of Chad Monroe's TFTP
705 rework patch that now integrates TFTP properly into libcurl so that it can
706 be used non-blocking with the multi interface and more. BLKSIZE also works.
708 The --tftp-blksize option was added to allow setting the TFTP BLKSIZE from
711 Daniel Stenberg (26 Nov 2009)
712 - Extended and fixed the change I did on Dec 11 for the the progress
713 meter/callback during FTP command/response sequences. It turned out it was
714 really lame before and now the progress meter SHOULD get called at least
717 Daniel Stenberg (23 Nov 2009)
718 - Bjorn Augustsson reported a bug which made curl not report any problems even
719 though it failed to write a very small download to disk (done in a single
720 fwrite call). It turned out to be because fwrite() returned success, but
721 there was insufficient error-checking for the fclose() call which tricked
722 curl to believe things were fine.
724 Yang Tse (23 Nov 2009)
725 - David Byron modified Makefile.dist vc8 and vc9 targets in order to allow
726 finer granularity control when generating src and lib makefiles.
728 Yang Tse (22 Nov 2009)
729 - I modified configure to force removal of the curlbuild.h file included in
730 distribution tarballs for use by non-configure systems. As intended, this
731 would get overwriten when doing in-tree builds. But VPATH builds would end
732 having two curlbuild.h files, one in the source tree and another in the
733 build tree. With the modification I introduced 5 Nov 2009 this could become
734 an issue when running libcurl's test suite.
736 Daniel Stenberg (20 Nov 2009)
737 - Constantine Sapuntzakis identified a write after close, as the sockets were
738 closed by libcurl before the SSL lib were shutdown and they may write to its
739 socket. Detected to at least happen with OpenSSL builds.
741 - Jad Chamcham pointed out a bug with connection re-use. If a connection had
742 CURLOPT_HTTPPROXYTUNNEL enabled over a proxy, a subsequent request using the
743 same proxy with the tunnel option disabled would still wrongly re-use that
744 previous connection and the outcome would only be badness.
746 Yang Tse (18 Nov 2009)
747 - I modified the memory tracking system to make it intolerant with zero sized
748 malloc(), calloc() and realloc() function calls.
750 Daniel Stenberg (17 Nov 2009)
751 - Constantine Sapuntzakis provided another fix for the DNS cache that could
752 end up with entries that wouldn't time-out:
754 1. Set up a first web server that redirects (307) to a http://server:port
756 2. Have curl connect to the first web server using curl multi
758 After the curl_easy_cleanup call, there will be curl dns entries hanging
759 around with in_use != 0.
761 (http://curl.haxx.se/bug/view.cgi?id=2891591)
763 - Marc Kleine-Budde fixed: curl saved the LDFLAGS set during configure into
764 its pkg-config file. So -Wl stuff ended up in the .pc file, which is really
765 bad, and breaks if there are multiple -Wl in our LDFLAGS (which are in
766 PTXdist). bug #2893592 (http://curl.haxx.se/bug/view.cgi?id=2893592)
768 Kamil Dudka (15 Nov 2009)
769 - David Byron improved the configure script to use pkg-config to find OpenSSL
770 (and in particular the list of required libraries) even if a path is given
771 as argument to --with-ssl
773 Yang Tse (15 Nov 2009)
774 - I removed enable-thread / disable-thread configure option. These were only
775 placebo options. The library is always built as thread safe as possible on
778 Claes Jakobsson (14 Nov 2009)
779 - curl-config now accepts '--configure' to see what arguments was
780 passed to the configure script when building curl.
782 Daniel Stenberg (14 Nov 2009)
783 - Claes Jakobsson restored the configure functionality to detect NSS when
784 --with-nss is set but not "yes".
786 I think we can still improve that to check for pkg-config in that path etc,
787 but at least this patch brings back the same functionality we had before.
789 - Camille Moncelier added support for the file type SSL_FILETYPE_ENGINE for
790 the client certificate. It also disable the key name test as some engines
791 can select a private key/cert automatically (When there is only one key
792 and/or certificate on the hardware device used by the engine)
794 Yang Tse (14 Nov 2009)
795 - Constantine Sapuntzakis provided the fix that ensures that an SSL connection
796 won't be reused unless protection level for peer and host verification match.
798 I refactored how preprocessor symbol _THREAD_SAFE definition is done.
800 Kamil Dudka (12 Nov 2009)
801 - Kevin Baughman provided a fix preventing libcurl-NSS from crash on doubly
802 closed NSPR descriptor. The issue was hard to find, reported several times
803 before and always closed unresolved. More info at the RH bug:
804 https://bugzilla.redhat.com/534176
806 - libcurl-NSS now tries to reconnect with TLS disabled in case it detects
807 a broken TLS server. However it does not happen if SSL version is selected
808 manually. The approach was originally taken from PSM. Kaspar Brand helped me
809 to complete the patch. Original bug reports:
810 https://bugzilla.redhat.com/525496
811 https://bugzilla.redhat.com/527771
813 Yang Tse (12 Nov 2009)
814 - I modified configure script to make the getaddrinfo function check also
815 verify if the function is thread safe.
817 Yang Tse (11 Nov 2009)
818 - Marco Maggi reported that compilation failed when configured --with-gssapi
819 and GNU GSS installed due to a missing mutual exclusion of header files in
820 the Kerberos 5 code path. He also verified that my patch worked for him.
822 Daniel Stenberg (11 Nov 2009)
823 - Constantine Sapuntzakis posted bug #2891595
824 (http://curl.haxx.se/bug/view.cgi?id=2891595) which identified how an entry
825 in the DNS cache would linger too long if the request that added it was in
826 use that long. He also provided the patch that now makes libcurl capable of
827 still doing a request while the DNS hash entry may get timed out.
829 - Christian Schmitz noticed that the progress meter/callback was not properly
830 used during the FTP connection phase (after the actual TCP connect), while
831 it of course should be. I also made the speed check get called correctly so
832 that really slow servers will trigger that properly too.
834 Kamil Dudka (5 Nov 2009)
835 - Dropped misleading timeouts in libcurl-NSS and made sure the SSL socket works
836 in non-blocking mode.
838 Yang Tse (5 Nov 2009)
839 - I removed leading 'curl' path on the 'curlbuild.h' include statement in
840 curl.h, adjusting auto-makefiles include path, to enhance portability to
841 OS's without an orthogonal directory tree structure such as OS/400.
843 Daniel Stenberg (4 Nov 2009)
844 - I fixed several problems with the transfer progress meter. It showed the
845 wrong percentage for small files, most notable for <1000 bytes and could
846 easily end up showing more than 100% at the end. It also didn't show any
847 percentage, transfer size or estimated transfer times when transferring
850 Version 7.19.7 (4 November 2009)
852 Daniel Stenberg (2 Nov 2009)
853 - As reported independent by both Stan van de Burgt and Didier Brisebourg,
854 CURLINFO_SIZE_DOWNLOAD (the -w variable size_download) didn't work when
855 getting data from ldap!
857 Daniel Stenberg (31 Oct 2009)
858 - Gabriel Kuri reported a problem with CURLINFO_CONTENT_LENGTH_DOWNLOAD if the
859 download was 0 bytes, as libcurl would then return the size as unknown (-1)
860 and not 0. I wrote a fix and test case 566 to verify it.
862 Daniel Stenberg (30 Oct 2009)
863 - Liza Alenchery mentioned a problem with re-used SCP connection when a bad
864 auth is used, as it caused a crash. I failed to repeat the issue, but still
865 made a change that now forces the TCP connection used for a freed SCP
866 session to get closed and not be re-used.
868 - "Tom" posted a bug report that mentioned how libcurl did wrong when doing a
869 POST using a read callback, with Digest authentication and
870 "Transfer-Encoding: chunked" enforced. I would then cause the first request
871 to be wrongly sent and then basically hang until the server closed the
872 connection. I fixed the problem and added test case 565 to verify it.
874 Daniel Stenberg (25 Oct 2009)
875 - Dima Barsky made the curl cookie parser accept cookies even with blank or
876 unparsable expiry dates and then treat them as session cookies - previously
877 libcurl would reject cookies with a date format it couldn't parse. Research
878 shows that the major browser treat such cookies as session cookies. I
879 modified test 8 and 31 to verify this.
881 Daniel Stenberg (21 Oct 2009)
882 - Attempt to use pkg-config for finding out libssh2 installation details
885 - A patch in bug report #2883177 (http://curl.haxx.se/bug/view.cgi?id=2883177)
886 by Johan van Selst introduced the --crlfile option to curl, which makes curl
887 tell libcurl about a file with CRL (certificate revocation list) data to
890 Daniel Stenberg (18 Oct 2009)
891 - Ray Dassen provided a patch in Debian's bug tracker (bug number #551461)
892 that now makes curl_getdate(3) actually handles RFC 822 formatted dates that
893 use the "single letter military timezones".
894 http://www.rfc-ref.org/RFC-TEXTS/822/chapter5.html has the details.
896 - Fixed memory leak in the SCP/SFTP code as it never freed the knownhosts
899 - John Dennis filed bug report #2873666
900 (http://curl.haxx.se/bug/view.cgi?id=2873666) which identified a problem
901 which made libcurl loop infinitely when given incorrect credentials when
902 using HTTP GSS negotiate authentication. He also provided a small and simple
905 - Kevin Baughman found a double close() problem with libcurl-NSS, as when
906 libcurl called NSS to close the SSL "session" it also closed the actual
909 Yang Tse (17 Oct 2009)
910 - Bug report #2866724 indicated
911 (http://curl.haxx.se/bug/view.cgi?id=2866724) that curl on Windows failed
912 when writing files whose file names originally contained characters which
913 are not valid for file names on Windows. Dan Fandrich provided an initial
914 patch and another revised one to fix this issue.
916 Daniel Stenberg (1 Oct 2009)
917 - Tom Mueller correctly reported in bug report #2870221
918 (http://curl.haxx.se/bug/view.cgi?id=2870221) that libcurl returned an
919 incorrect return code from the internal trynextip() function which caused
920 him grief. This is a regression that was introduced in 7.19.1 and I find it
921 strange it hasn't hit us harder, but I won't persue into figuring out
924 - Constantine Sapuntzakis: The current implementation will always set
925 SO_SNDBUF to CURL_WRITE_SIZE even if the SO_SNDBUF starts out larger. The
926 patch doesn't do a setsockopt if SO_SNDBUF is already greater than
927 CURL_WRITE_SIZE. This should help folks who have set up their computer with
930 Daniel Stenberg (27 Sep 2009)
931 - I introduced a maximum limit for received HTTP headers. It is controlled by
932 the define CURL_MAX_HTTP_HEADER which is even exposed in the public header
933 file to allow for users to fairly easy rebuild libcurl with a modified
934 limit. The rationale for a fixed limit is that libcurl is realloc()ing a
935 buffer to be able to put a full header into it, so that it can call the
936 header callback with the entire header, but that also risk getting it into
937 trouble if a server by mistake or willingly sends a header that is more or
938 less without an end. The limit is set to 100K.
940 Daniel Stenberg (26 Sep 2009)
941 - John P. McCaskey posted a bug report that showed how libcurl did wrong when
942 saving received cookies with no given path, if the path in the request had a
943 query part. That is means a question mark (?) and characters on the right
944 side of that. I wrote test case 1105 and fixed this problem.
946 Kamil Dudka (26 Sep 2009)
947 - Implemented a protocol independent way to specify blocking direction, used by
948 transfer.c for blocking. It is currently used only by SCP and SFTP protocols.
949 This enhancement resolves an issue with 100% CPU usage during SFTP upload,
952 Daniel Stenberg (25 Sep 2009)
953 - Chris Mumford filed bug report #2861587
954 (http://curl.haxx.se/bug/view.cgi?id=2861587) identifying that libcurl used
955 the OpenSSL function X509_load_crl_file() wrongly and failed if it would
956 load a CRL file with more than one certificate within. This is now fixed.
958 Daniel Stenberg (16 Sep 2009)
959 - Sven Anders reported that we introduced a cert verfication flaw for OpenSSL-
960 powered libcurl in 7.19.6. If there was a X509v3 Subject Alternative Name
961 field in the certficate it had to match and so even if non-DNS and non-IP
962 entry was present it caused the verification to fail.
964 Daniel Fandrich (15 Sep 2009)
965 - Moved the libssh2 checks after the SSL library checks. This helps when
966 statically linking since libssh2 needs the SSL library link flags to be
967 set up already to satisfy its dependencies. This wouldn't be necessary if
968 the libssh2 configure check was changed to use pkg-config since the
969 --static flag would add the dependencies automatically.
971 Yang Tse (14 Sep 2009)
972 - Revert Joshua Kwan's patch committed 11 Sep 2009.
974 Some systems poll function sets POLLHUP in revents without setting
975 POLLIN, and sets POLLERR without setting POLLIN and POLLOUT. In some
976 libcurl code execution paths this could trigger busy wait loops with
977 high CPU usage until a timeout condition aborted the loop.
979 The reverted patch addressed the above issue for a very specific case,
980 when awaiting c-ares to resolve. A libcurl-wide fix for Curl_poll now
983 Guenter Knauf (11 Sep 2009)
984 - Joshua Kwan provided a patch to pass POLLERR / POLLHUP back to c-ares.
985 This fixes a loop problem with high CPU usage.
987 Daniel Stenberg (10 Sep 2009)
988 - Claes Jakobsson fixed a problem with cookie expiry dates at exctly the epoch
989 start second "Thu Jan 1 00:00:00 GMT 1970" as the date parser then returns 0
990 which internally then is treated as a session cookie. That particular date
991 is now made to get the value of 1.
993 Daniel Stenberg (2 Sep 2009)
994 - Daniel Johnson found a flaw in the code converting sftp-errors to libcurl
997 Daniel Stenberg (1 Sep 2009)
998 - Peter Sylvester made a debug feature for Curl_resolv() that now will force
999 libcurl to resolve 'localhost' whatever name you use in the URL *if* you set
1000 the --interface option to (exactly) "LocalHost". This will enable us to
1001 write tests for custom hosts names but still use a local host server.
1003 - configure now tries to use pkg-config for a number of sub-dependencies even
1004 when cross-compiling. The key to success is then you properly setup
1005 PKG_CONFIG_PATH before invoking configure.
1007 I also improved how NSS is detected by trying nss-config if pkg-config isn't
1008 present, and as a last resort just use the lib name and force the user to
1009 setup the LIBS/LDFLAGS/CFLAGS etc properly. The previous last resort would
1010 add a range of various libs that would almost never be quite correct.
1012 Daniel Stenberg (31 Aug 2009)
1013 - When using the multi interface with FTP and you asked for NOBODY, you did no
1014 QUOTE commands and the request used the same path as the connection had
1015 already changed to, it would decide that no commands would be necessary for
1016 the "DO" action and that was not handled properly but libcurl would instead
1019 Kamil Dudka (28 Aug 2009)
1020 - Improved error message for not matching certificate subject name in
1021 libcurl-NSS. Originally reported at:
1022 https://bugzilla.redhat.com/show_bug.cgi?id=516056#c9
1024 Patrick Monnerat (24 Aug 2009)
1025 - Introduced a SYST-based test to properly set-up name format when dealing
1026 with the OS/400 FTP server.
1028 - Fixed an ftp_readresp() bug preventing detection of failing control socket
1029 and causing FTP client to loop forever.
1031 Daniel Stenberg (24 Aug 2009)
1032 - Marc de Bruin pointed out that configure --with-gnutls=PATH didn't work
1033 properly and provided a fix. http://curl.haxx.se/bug/view.cgi?id=2843008
1035 - Eric Wong introduced support for the new option -T. (dot) that makes curl
1036 read stdin in a non-blocking fashion. This also brings back -T- (minus) to
1037 the previous blocking behavior since it could break stuff for people at
1040 Michal Marek (21 Aug 2009)
1041 - With CURLOPT_PROXY_TRANSFER_MODE, avoid sending invalid URLs like
1042 ftp://example.com;type=i if the user specified ftp://example.com without the
1045 Daniel Stenberg (21 Aug 2009)
1046 - Andre Guibert de Bruet pointed out a missing return code check for a
1047 strdup() that could lead to segfault if it returned NULL. I extended his
1048 suggest patch to now have Curl_retry_request() return a regular return code
1049 and better check that.
1051 - Lots of good work by Krister Johansen, mostly related to pipelining:
1053 Fix SIGSEGV on free'd easy_conn when pipe unexpectedly breaks
1054 Fix data corruption issue with re-connected transfers
1055 Fix use after free if we're completed but easy_conn not NULL
1057 Kamil Dudka (13 Aug 2009)
1058 - Changed NSS code to not ignore the value of ssl.verifyhost and produce more
1059 verbose error messages. Originally reported at:
1060 https://bugzilla.redhat.com/show_bug.cgi?id=516056
1062 Daniel Stenberg (12 Aug 2009)
1063 - Karl Moerder fixed the Makefile.vc* makefiles to include the new file
1064 nonblock.c so that they work fine again
1066 - I expanded test 517 with a bunch of more dates that originate from the
1067 Chrome browser test suite. It turns out most of them get parsed the same
1070 Version 7.19.6 (12 August 2009)
1072 Daniel Stenberg (12 Aug 2009)
1073 - Carsten Lange reported a bug and provided a patch for TFTP upload and the
1074 sending of the TSIZE option. I don't like fixing bugs just hours before
1075 a release, but since it was broken and the patch fixes this for him I decided
1076 to get it in anyway.
1078 Daniel Stenberg (11 Aug 2009)
1079 - Peter Sylvester made the HTTPS test server use specific certificates for
1080 each test, so that the test suite can now be used to actually test the
1081 verification of cert names etc. This made an error show up in the OpenSSL-
1082 specific code where it would attempt to match the CN field even if a
1083 subjectAltName exists that doesn't match. This is now fixed and verified
1086 - Benbuck Nason posted the bug report #2835196
1087 (http://curl.haxx.se/bug/view.cgi?id=2835196), fixing a few compiler
1088 warnings when mixing ints and bools.
1090 Daniel Fandrich (10 Aug 2009)
1091 - Fixed a memory leak in the FTP code and an off-by-one heap buffer overflow.
1093 Daniel Fandrich (9 Aug 2009)
1094 - Fixed some memory leaks in the command-line tool that caused most of the
1095 torture tests to fail.
1097 Daniel Stenberg (2 Aug 2009)
1098 - Curt Bogmine reported a problem with SNI enabled on a particular server. We
1099 should introduce an option to disable SNI, but as we're in feature freeze
1100 now I've addressed the obvious bug here (pointed out by Peter Sylvester): we
1101 shouldn't try to enable SNI when SSLv2 or SSLv3 is explicitly selected.
1102 Code for OpenSSL and GnuTLS was fixed. NSS doesn't seem to have a particular
1103 option for SNI, or are we simply not using it?
1105 Daniel Stenberg (1 Aug 2009)
1106 - Scott Cantor posted the bug report #2829955
1107 (http://curl.haxx.se/bug/view.cgi?id=2829955) mentioning the recent SSL cert
1108 verification flaw found and exploited by Moxie Marlinspike. The presentation
1109 he did at Black Hat is available here:
1110 https://www.blackhat.com/html/bh-usa-09/bh-usa-09-archives.html#Marlinspike
1112 Apparently at least one CA allowed a subjectAltName or CN that contain a
1113 zero byte, and thus clients that assumed they would never have zero bytes
1114 were exploited to OK a certificate that didn't actually match the site. Like
1115 if the name in the cert was "example.com\0theatualsite.com", libcurl would
1116 happily verify that cert for example.com.
1118 libcurl now better uses the length of the extracted name, not using the zero
1119 termination for getting the string length.
1121 This fixing only made and needed in OpenSSL interfacing code.
1123 - Tanguy Fautre pointed out that OpenSSL's function RAND_screen() (present
1124 only in some OpenSSL installs - like on Windows) isn't thread-safe and we
1125 agreed that moving it to the global_init() function is a decent way to deal
1126 with this situation.
1128 - Alexander Beedie provided the patch for a noproxy problem: If I have set
1129 CURLOPT_NOPROXY to "*", or to a host that should not use a proxy, I actually
1130 could still end up using a proxy if a proxy environment variable was set.
1132 Daniel Stenberg (27 Jul 2009)
1133 - All the quote options (CURLOPT_QUOTE, CURLOPT_POSTQUOTE and
1134 CURLOPT_PREQUOTE) now accept a preceeding asterisk before the command to
1135 send when using FTP, as a sign that libcurl shall simply ignore the response
1136 from the server instead of treating it as an error. Not treating a 400+ FTP
1137 response code as an error means that failed commands will not abort the
1138 chain of commands, nor will they cause the connection to get disconnected.
1140 Daniel Stenberg (26 Jul 2009)
1141 - Johan van Selst posted bug report #2825989
1142 (http://curl.haxx.se/bug/view.cgi?id=2825989) pointing out that
1143 OpenSSL-powered libcurl didn't support the SHA-2 digest algorithm, and
1144 provided the solution too: to use OpenSSL_add_all_algorithms() in addition
1145 to the older SSLeay_* alternative. OpenSSL_add_all_algorithms was added in
1148 Daniel Stenberg (23 Jul 2009)
1149 - Added CURLOPT_SSH_KNOWNHOSTS, CURLOPT_SSH_KEYFUNCTION, CURLOPT_SSH_KEYDATA.
1150 They introduce known_host support for SSH keys to libcurl. See docs for
1151 details. Note that this feature depends on a new enough libssh2 version, to
1152 be supported in libssh2 1.2 and later (or current git repo at this time).
1154 Michal Marek (22 Jul 2009)
1155 - David Binderman found a memory and fd leak in lib/gtls.c:load_file()
1156 (https://bugzilla.novell.com/523919). When looking at the code, I found that
1157 also the ptr pointer can leak.
1159 Kamil Dudka (20 Jul 2009)
1160 - Claes Jakobsson improved the support for client certificates handling in
1161 NSS-powered libcurl. Now the client certificates can be selected
1162 automatically by a NSS built-in hook. Additionally pre-login to all PKCS11
1163 slots is no more performed. It used to cause problems with HW tokens.
1165 - Fixed reference counting for NSS client certificates. Now the PEM reader
1166 module should be always properly unloaded on Curl_nss_cleanup(). If the
1167 unload fails though, libcurl will try to reuse the already loaded instance.
1169 Daniel Fandrich (15 Jul 2009)
1170 - Added nonblock.c to the non-automake makefiles (note that the dependencies
1171 in the Watcom makefiles aren't quite correct).
1173 Michal Marek (15 Jul 2009)
1174 - Changed the description of CURLINFO_OS_ERRNO to make it clear that the
1175 errno is not reset on success.
1177 Guenter Knauf (14 Jul 2009)
1178 - renamed generated config.h to curl_config.h to avoid any future clashes
1179 with config.h from other projects.
1181 Daniel Stenberg (9 Jul 2009)
1182 - Eric Wong introduced curlx_nonblock() that the curl tool now (re-)uses for
1183 setting a file descriptor non-blocking. Used by the functionality Eric
1184 himself brough on June 15th.
1186 Daniel Stenberg (8 Jul 2009)
1187 - Constantine Sapuntzakis posted bug report #2813123
1188 (http://curl.haxx.se/bug/view.cgi?id=2813123) and an a patch that fixes the
1191 Url A is accessed using auth. Url A redirects to Url B (on a different
1192 server0. Url B reuses a persistent connection. Url B has auth, even though
1193 it's on a different server.
1195 Note: if Url B does not reuse a persistent connection, auth is not sent.
1199 data->state.first_host is not initialized becuase Curl_http_connect is not
1200 called when a connection is reused.
1204 move initialization of data->state.first_host to Curl_http. No code before
1205 Curl_http uses data->state.first_host anyway.
1207 Guenter Knauf (4 Jul 2009)
1208 - Markus Koetter provided a patch to avoid getnameinfo() usage which broke a
1209 couple of both IPv4 and IPv6 autobuilds.
1211 Daniel Stenberg (29 Jun 2009)
1212 - Markus Koetter made CURLOPT_FTPPORT (and curl's -P/--ftpport) support a port
1213 range if given colon-separated after the host name/address part. Like
1214 "192.168.0.1:2000-10000"
1216 - Modified the separators used for CURLOPT_CERTINFO in multi-part outputs. I
1217 don't know how they got wrong in the first place, but using this output
1218 format makes it possible to quite easily separate the string into an array
1221 Daniel Fandrich (16 June 2009)
1222 - Added a few more compiler warning options for gcc.
1224 Daniel Stenberg (16 Jun 2009)
1225 - Reuven Wachtfogel made curl -o - properly produce a binary output on windows
1226 (no newline translations). Use -B/--use-ascii if you rather get the ascii
1229 Michal Marek (16 Jun 2009)
1230 - When doing non-anonymous ftp via http proxies and the password is not
1231 provided in the url, add it there (squid needs this).
1233 Daniel Stenberg (15 Jun 2009)
1234 - Eric Wong's patch:
1236 This allows curl(1) to be used as a client-side tunnel for arbitrary stream
1237 protocols by abusing chunked transfer encoding in both the HTTP request and
1238 HTTP response. This requires server support for sending a response while a
1239 request is still being read, of course.
1241 If attempting to read from stdin returns EAGAIN, then we pause our sender.
1242 This leaves curl to attempt to read from the socket while reading from stdin
1243 (and thus sending) is paused.
1245 This change was needed to allow successfully tunneling the git protocol over
1246 HTTP (--no-buffer is needed, as well).
1248 Patrick Monnerat (15 Jun 2009)
1249 - Replaced use of standard C library rand()/srand() by our own pseudo-random
1252 Yang Tse (11 Jun 2009)
1253 - I adapted testcurl script to allow building test harness programs when
1254 cross-compiling for a *-*-mingw* host.
1256 Daniel Stenberg (10 Jun 2009)
1257 - Fabian Keil ran clang on the (lib)curl code, found a bunch of warnings and
1258 contributed a range of patches to fix them.
1260 Yang Tse (10 Jun 2009)
1261 - I introduced configure script option --enable-curldebug which now allows
1262 the decoupled enabling or disabling of the curl debug memory tracking
1263 feature from the --enable-debug option which no longer controls this.
1265 curl --version will list 'Debug' feature for debug enabled builds, and
1266 will list 'TrackMemory' feature for curl debug memory tracking capable
1267 builds. These features are independent and can be controlled when running
1268 the configure script. When --enable-debug is given both features will be
1269 enabled, unless some restriction prevents memory tracking from being used.
1271 Internally, definition of preprocessor symbol DEBUGBUILD restricts code
1272 which is only compiled for debug enabled builds. And symbol CURLDEBUG is
1273 used to differentiate code which is _only_ used for memory tracking.
1275 Yang Tse (9 Jun 2009)
1276 - Daniel Steinberg pointed out that Curl_FormInit() in formdata.c was not
1277 initializing the fread callback pointer and this triggered a compiler
1278 warning, also provided a friendly suggestion on how to fix it.
1280 Daniel Stenberg (8 Jun 2009)
1281 - Claes Jakobsson provided a patch for libcurl-NSS that fixed a bad refcount
1282 issue with client certs that caused issues like segfaults.
1283 http://curl.haxx.se/mail/lib-2009-05/0316.html
1285 - Triggered by bug report #2798852 and the patch in there, I fixed configure
1286 to detect gnutls build options with pkg-config only and not libgnutls-config
1287 anymore since GnuTLS has stopped distributing that tool. If an explicit path
1288 is given to configure, we will instead guess on how to link and use that
1289 lib. I did not use the patch from the bug report.
1291 Yang Tse (8 Jun 2009)
1292 - Igor Novoseltsev adjusted Makefile.vxworks to get sources and headers
1293 included from Makefile.inc, and provided docs\INSTALL VxWorks section.
1295 - I removed buildconf.bat from release and daily snapshot archives. This
1296 file is only for CVS tree checkout builds.
1298 Daniel Stenberg (8 Jun 2009)
1299 - Eric Wong fixed --no-buffer to actually switch off output buffering. Been
1302 Bill Hoffman (6 Jun 2009)
1303 - Added some cmake docs and fixed socklen_t in the build.
1305 Yang Tse (5 Jun 2009)
1306 - John E. Malmberg provided VMS specific patch: "This fixes an existing bug
1307 in urlglob.c where it was not converting the Curl Unix exit code to a VMS
1308 DCL compatible exit code. This fix required the enhancement described next.
1309 This also adds an enhancement to main.c so that when curl is run under a
1310 Unix shell like Bash on VMS, it will return the standard Unix exit codes
1311 and messages." And another patch for docs/examples.
1313 I introduced os-specific.c and os-specific.h for use in curl tool code
1314 and adjusted John E. Malmberg's patch placement to use these new files
1315 as an effort to prevent main.c from growing ad infinitum. Code already
1316 existing in main.c which is OS specific should be moved into these files.
1318 Daniel Stenberg (4 June 2009)
1319 - Setting the Content-Length: header from your app when you do a POST or PUT
1320 is almost always a VERY BAD IDEA. Yet there are still apps out there doing
1321 this, and now recently it triggered a bug/side-effect in libcurl as when
1322 libcurl sends a POST or PUT with NTLM, it sends an empty post first when it
1323 knows it will just get a 401/407 back. If the app then replaced the
1324 Content-Length header, it caused the server to wait for input that libcurl
1325 wouldn't send. Aaron Oneal reported this problem in bug report #2799008
1326 (http://curl.haxx.se/bug/view.cgi?id=2799008) and helped us verify the fix.
1328 Yang Tse (4 Jun 2009)
1329 - Igor Novoseltsev provided patches and information, that after some
1330 adjustments to better fit curl's way of doing things, have resulted
1331 in the posibility of building libcurl for VxWorks.
1333 Daniel Fandrich (2 June 2009)
1334 - Checked in a Google Android make file. To use it, you must first
1335 create a config.h file by running configure in the Android environment,
1336 which doesn't seem to be easy to do. If no easy way can be found, a
1337 static config-android.h may need to be created and checked in to the
1338 libcurl source tree.
1340 Daniel Stenberg (1 June 2009)
1341 - Claes Jakobsson fixed the configure script to better find and use NSS
1344 Yang Tse (1 Jun 2009)
1345 - John E. Malmberg provided a VMS specific clean-up for curl.h, and pointed
1346 out that the configure script was failing to detect the timeval struct on
1347 VMS when building with _XOPEN_SOURCE_EXTENDED undefined due to definition
1348 taking place in socket.h instead of time.h. I have adjusted configure
1349 script to also include this header when checking struct timeval.
1351 Daniel Stenberg (27 May 2009)
1352 - Frank McGeough provided a small OpenSSL #include fix to make libcurl compile
1353 fine with Nokia 5th edition 1.0 SDK for Symbian.
1355 - Andre Guibert de Bruet found a call to a OpenSSL function that didn't check
1356 for a failure properly.
1358 - Mike Crowe pointed out that setting CURLOPT_USERPWD to NULL used to clear
1359 the auth credentials back in 7.19.0 and earlier while now you have to set ""
1360 to get the same effect. His patch brings back the ability to use NULL.
1362 - Claes Jakobsson fixed libcurl-NSS to build fine even without the
1363 PK11_CreateGenericObject() function.
1365 Daniel Stenberg (25 May 2009)
1366 - bug report #2796358 (http://curl.haxx.se/bug/view.cgi?id=2796358) pointed
1367 out that the cookie parser would leak memory when it parses cookies that are
1368 received with domain, path etc set multiple times in the same header. While
1369 such a cookie is questionable, they occur in the wild and libcurl no longer
1370 leaks memory for them. I added such a header to test case 8.
1372 Daniel Fandrich (22 May 2009)
1373 - Removed some obsolete digest code that caused a valgrind error in test 551.
1375 Daniel Fandrich (20 May 2009)
1376 - Added "non-existing host" test keywords to make it easy to skip those
1377 tests on machines that have broken DNS configurations (such as
1378 those configured to use OpenDNS).
1380 Daniel Stenberg (19 May 2009)
1381 - Kamil Dudka brought the patch from the Redhat bug entry
1382 https://bugzilla.redhat.com/show_bug.cgi?id=427966 which was libcurl closing
1383 a bad file descriptor when closing down the FTP data connection. Caolan
1384 McNamara seems to be the original author of it.
1386 Version 7.19.5 (18 May 2009)
1388 Daniel Stenberg (17 May 2009)
1389 - James Bursa posted a patch to the mailing list that fixed a problem with
1390 no_proxy which made it not skip the proxy if the URL entered contained a
1391 user name. I added test case 1101 to verify.
1393 Daniel Stenberg (11 May 2009)
1394 - Balint Szilakszi reported a memory leak when libcurl did gzip decompression
1395 of streams that had some parts (legitimately) missing. We now provide and use
1396 a proper cleanup function for the content encoding submodule.
1397 http://curl.haxx.se/mail/lib-2009-05/0092.html
1399 - Kamil Dudka provided a fix for libcurl-NSS reported by Michael Cronenworth
1400 at https://bugzilla.redhat.com/show_bug.cgi?id=453612#c12
1402 If an incorrect password is given while loading a private key, libcurl ends
1403 up in an infinite loop consuming memory. The bug is critical.
1405 - I fixed the problem with doing NTLM, POST and then following a 302 redirect,
1406 as reported by Ebenezer Ikonne (on curl-users) and Laurent Rabret (on
1407 curl-library). The transfer was mistakenly marked to get more data to send
1408 but since it didn't actually have that, it just hung there...
1410 Daniel Stenberg (10 May 2009)
1411 - Andre Guibert de Bruet correctly pointed out an over-alloc with one wasted
1412 byte in the digest code.
1414 Yang Tse (9 May 2009)
1415 - Removed DOS and TPF package's subdirectory Makefile.am, it was only used
1416 to include some files in the distribution tarball serving no other purpose.
1417 Files from the DOS and TPF subdirectories are now included in the EXTRA_DIST
1418 of the Makefile in the parent subdirectory.
1420 Yang Tse (8 May 2009)
1421 - Changed host name literal in several tests to one under the haxx.se domain.
1423 - Renamed vc6 workspace and project files to avoid filename clash when used
1424 for conversion to later VS versions.
1426 Daniel Stenberg (8 May 2009)
1427 - Constantine Sapuntzakis fixed bug report #2784055
1428 (http://curl.haxx.se/bug/view.cgi?id=2784055) identifying a problem to
1429 connect to SOCKS proxies when using the multi interface. It turned out to
1430 almost not work at all previously. We need to wait for the TCP connect to
1431 be properly verified before doing the SOCKS magic.
1433 There's still a flaw in the FTP code for this.
1435 Daniel Stenberg (7 May 2009)
1436 - Made the SO_SNDBUF setting for the data connection socket for ftp uploads as
1437 well. See change 28 Apr 2009.
1439 Yang Tse (7 May 2009)
1440 - Fixed an issue affecting FTP transfers, introduced with the transfer.c
1441 patch committed May 4.
1443 Daniel Stenberg (7 May 2009)
1444 - Man page *roff problems fixed thanks to input from Colin Watson. Problems
1445 reported in the Debian package.
1447 - Vijay G filed bug report #2723236
1448 (http://curl.haxx.se/bug/view.cgi?id=2723236) identifying a problem with
1449 libcurl's TFTP code and its lack of dealing with the OACK packet.
1451 Yang Tse (5 May 2009)
1452 - Fixed the --ftp-port address of test #251 to the CLIENTIP address, and
1453 reverted the change affecting test suite harness committed 4 May.
1455 Daniel Stenberg (5 May 2009)
1456 - Inspired by Michael Smith's session id fix for OpenSSL, I did the
1457 corresponding fix in the GnuTLS code: make sure to store the new session id
1458 in case the previous re-used one is rejected.
1460 Daniel Stenberg (4 May 2009)
1461 - Michael Smith posted bug report #2786255
1462 (http://curl.haxx.se/bug/view.cgi?id=2786255) with a patch, identifying how
1463 libcurl did not deal with SSL session ids properly if the server rejected a
1464 re-use of one. Starting now, it will forget the rejected one and remember
1465 the new. This change was for OpenSSL only, it is likely that other SSL lib
1466 code needs similar fixes.
1468 Yang Tse (4 May 2009)
1469 - Applied David McCreedy's "transfer.c fixes for CURL_DO_LINEEND_CONV and
1470 non-ASCII platform HTTP requests" patch addressing two HTTP PUT problems:
1471 1) On non-ASCII platforms not all of the protocol portions of the PUT are
1472 being translated to ASCII. 2) On all platforms the line endings of part of
1473 the protocol portions are mangled from CRLF to CRCRLF if data->set.crlf or
1474 data->set.prefer_ascii are set (depending on CURL_DO_LINEEND_CONV).
1476 - Applied David McCreedy's patch to fix test suite harness to allow test FTP
1477 server and client on different machines, providing FTP client address when
1478 running the FTP test server.
1480 Daniel Fandrich (3 May 2009)
1481 - Added and disabled test case 563 which shows KNOWN_BUGS #59. The bug
1482 report failed to mention that a proxy must be used to reproduce it.
1484 Yang Tse (2 May 2009)
1485 - Use a build-time configured curl_socklen_t data type instead of socklen_t.
1487 Yang Tse (1 May 2009)
1488 - Applied David McCreedy's patches "TPF-platform specific changes to various
1489 files" and "http.c fix to Curl_proxyCONNECT for non-ASCII platforms", the
1490 former with minor edits.
1492 Daniel Stenberg (30 Apr 2009)
1493 - I was going to fix issue #59 in KNOWN_BUGS
1495 If the CURLOPT_PORT option is used on an FTP URL like
1496 "ftp://example.com/file;type=A" the ";type=A" is stripped off.
1498 I added test case 562 to verify, only to find out that I couldn't repeat
1499 this bug so I hereby consider it not a bug anymore!
1501 Daniel Stenberg (29 Apr 2009)
1502 - Based on bug report #2723219 (http://curl.haxx.se/bug/view.cgi?id=2723219)
1503 I've now made TFTP "connections" not being kept for re-use within libcurl.
1504 TFTP is UDP-based so the benefit was really low (if even existing) to begin
1505 with so instead of tracking down to fix this problem we instead removed the
1506 re-use. I also enabled test case 1099 that I wrote a few days ago to verify
1507 that this change fixes the reported problem.
1509 Daniel Stenberg (28 Apr 2009)
1510 - Constantine Sapuntzakis filed bug report #2783090
1511 (http://curl.haxx.se/bug/view.cgi?id=2783090) pointing out that on windows
1512 we need to grow the SO_SNDBUF buffer somewhat to get really good upload
1513 speeds. http://support.microsoft.com/kb/823764 has the details. Friends
1514 confirmed that simply adding 32 to CURL_MAX_WRITE_SIZE is enough.
1516 - Bug report #2709004 (http://curl.haxx.se/bug/view.cgi?id=2709004) by Tim
1517 Chen pointed out how curl couldn't upload with resume when reading from a
1520 This ended up with the introduction of a new return code for the
1521 CURLOPT_SEEKFUNCTION callback that basically says that the seek failed but
1522 that libcurl may try to resolve the situation anyway. In our case this means
1523 libcurl will attempt to instead read that much data from the stream instead
1524 of seeking and that way curl can now upload with resume when data is read
1527 Daniel Stenberg (26 Apr 2009)
1528 - Bug report #2779733 (http://curl.haxx.se/bug/view.cgi?id=2779733) by Sven
1529 Wegener pointed out that CURLINFO_APPCONNECT_TIME didn't work with the multi
1530 interface and provided a patch that fixed the problem!
1532 Daniel Stenberg (24 Apr 2009)
1533 - Kamil Dudka fixed another NSS-related leak when client certs were used.
1535 - Bug report #2779245 (http://curl.haxx.se/bug/view.cgi?id=2779245) by Rainer
1536 Koenig pointed out that the man page didn't tell that the *_proxy
1537 environment variables can be specified lower case or UPPER CASE and the
1538 lower case takes precedence,
1540 Daniel Fandrich (21 Apr 2009)
1541 - Added new libcurl source files to Amiga, RiscOS and VC6 build files.
1543 Yang Tse (21 Apr 2009)
1544 - Moved potential inclusion of system's malloc.h and memory.h header files to
1545 setup_once.h. Inclusion of each header file is based on the definition of
1546 NEED_MALLOC_H and NEED_MEMORY_H respectively.
1548 Renamed libcurl's memory.h to curl_memory.h
1550 Daniel Stenberg (20 Apr 2009)
1551 - Leanic Lefever reported a crash and did some detailed research on why and
1552 how it occurs (http://curl.haxx.se/mail/lib-2009-04/0289.html). The
1553 conclusion was that if an error is detected and Curl_done() is called for
1554 the connection, ftp_done() could at times return another error code that
1555 then would take precedence and that new code confused existing logic that
1556 works for the first error code (CURLE_SEND_ERROR) only.
1558 - Gisle Vanem noticed that --libtool would produce bogus strings at times for
1559 OBJECTPOINT options. Now we've introduced a new function - my_setopt_str -
1560 within the app for setting plain string options to avoid the risk of this
1563 Daniel Stenberg (17 Apr 2009)
1564 - Pramod Sharma reported and tracked down a bug when doing FTP over a HTTP
1565 proxy. libcurl would then wrongly close the connection after each
1566 request. In his case it had the weird side-effect that it killed NTLM auth
1567 for the proxy causing an inifinite loop!
1569 I added test case 1098 to verify this fix. The test case does however not
1570 properly verify that the transfers are done persistently - as I couldn't
1571 think of a clever way to achieve it right now - but you need to read the
1572 stderr output after a test run to see that it truly did the right thing.
1574 Daniel Stenberg (13 Apr 2009)
1575 - bug report #2727981 (http://curl.haxx.se/bug/view.cgi?id=2727981) by Martin
1576 Storsjö pointed out how setting CURLOPT_NOBODY to 0 could be downright
1577 confusing as it set the method to either GET or HEAD. The example he showed
1580 curl_easy_setopt(curl, CURLOPT_PUT, 1);
1581 curl_easy_setopt(curl, CURLOPT_NOBODY, 0);
1583 The new way doesn't alter the method until the request is about to start. If
1584 CURLOPT_NOBODY is then 1 the HTTP request will be HEAD. If CURLOPT_NOBODY is
1585 0 and the request happens to have been set to HEAD, it will then instead be
1586 set to GET. I believe this will be less surprising to users, and hopefully
1587 not hit any existing users badly.
1589 - Toshio Kuratomi reported a memory leak problem with libcurl+NSS that turned
1590 out to be leaking cacerts. Kamil Dudka helped me complete the fix. The issue
1591 is found in Redhat's bug tracker:
1592 https://bugzilla.redhat.com/show_bug.cgi?id=453612
1594 There are still memory leaks present, but they seem to have other reasons.
1596 Daniel Fandrich (11 Apr 2009)
1597 - Added new libcurl source files to Symbian OS build files.
1598 - Improved Symbian support for SSL.
1600 Yang Tse (10 Apr 2009)
1601 - Daniel Johnson improved the MacOSX-Framework shell script to now perform all
1602 the steps required to build a Mac OS X four way fat ppc/i386/ppc64/x86_64
1603 libcurl.framework. Four way fat framework requires OS X 10.5 SDK or later.
1605 Yang Tse (8 Apr 2009)
1606 - Removed Sun compilers preprocessor block from curlbuild.h.dist, this also
1607 removes it from the curlbuild.h file originally distributed by the cURL
1608 project as this file is intended for systems not capable of running the
1609 configure script. For those who have been building curl out of the source
1610 code curl distribution tarball provided by curl.haxx.se the change implies
1611 nothing. Previous change in this area committed 2 Apr becomes irrelevant.
1613 Daniel Stenberg (6 Apr 2009)
1614 - I clarified in the docs that CURLOPT_SEEKFUNCTION should return 0 on success
1615 and 1 on fatal errors. Previously it only mentioned non-zero on fatal
1616 errors. This is a slight change in meaning, but it follows what we've done
1617 elsewhere before and it opens up for LOTS of more useful return codes
1618 whenever we can think of them...
1620 Yang Tse (2 Apr 2009)
1621 - Fix curl_off_t definition for builds done using Sun compilers and a
1622 non-configured libcurl. In this case curl_off_t data type was gated
1623 to the off_t data type which depends on the _FILE_OFFSET_BITS. This
1624 configuration is exactly the unwanted configuration for our curl_off_t
1625 data type which must not depend on such setting. This breaks ABI for
1626 libcurl libraries built with Sun compilers which were built without
1627 having run the configure script with _FILE_OFFSET_BITS different than
1628 64 and using the ILP32 data model.
1630 Daniel Stenberg (1 Apr 2009)
1631 - Andre Guibert de Bruet fixed a NULL pointer use in an infof() call if a
1632 strdup() call failed.
1634 Daniel Fandrich (31 Mar 2009)
1635 - Properly return an error code in curl_easy_recv (reported by Jim Freeman).
1637 Daniel Stenberg (18 Mar 2009)
1638 - Kamil Dudka brought a patch that enables 6 additional crypto algorithms when
1639 NSS is used. These ciphers were added in NSS 3.4 and require to be enabled
1642 Daniel Stenberg (13 Mar 2009)
1643 - Use libssh2_version() to present the libssh2 version in case the libssh2
1644 library is found to support it.
1646 Yang Tse (12 Mar 2009)
1647 - Added missing Curl_read() return code checking in TELNET transfers.
1649 - Pierre Brico found and fixed TELNET transfers not being aborted upon
1650 a write callback failure.
1652 Daniel Stenberg (11 Mar 2009)
1653 - Kamil Dudka made the curl tool properly call curl_global_init() before any
1654 other libcurl function.
1656 Yang Tse (11 Mar 2009)
1657 - Added missing TELNET timeout support for Windows builds. This issue was
1658 reported by Pierre Brico.
1660 Daniel Stenberg (9 Mar 2009)
1661 - Frank Hempel found out a bug and provided the fix:
1663 curl_easy_duphandle did not necessarily duplicate the CURLOPT_COOKIEFILE
1664 option. It only enabled the cookie engine in the destination handle if
1665 data->cookies is not NULL (where data is the source handle). In case of a
1666 newly initialized handle which just had the cookie support enabled by a
1667 curl_easy_setopt(handle, CURL_COOKIEFILE, "")-call, handle->cookies was
1668 still NULL because the setopt-call only appends the value to
1669 data->change.cookielist, hence duplicating this handle would not have the
1670 cookie engine switched on.
1672 We also concluded that the slist-functionality would be suitable for being
1673 put in its own module rather than simply hanging out in lib/sendf.c so I
1674 created lib/slist.[ch] for them.
1676 - Andreas Farber made the 'buildconf' script check for the presence of m4
1677 scripts to make it detect a bad checkout earlier. People with older
1678 checkouts who don't do cvs update with the -d option won't get the new dirs
1679 and then will get funny outputs that can be a bit hard to understand and
1682 Daniel Stenberg (8 Mar 2009)
1683 - Andre Guibert de Bruet found and fixed a code segment in ssluse.c where the
1684 allocation of the memory BIO was not being properly checked.
1686 - Andre Guibert de Bruet fixed the gnutls-using code: There are a few places
1687 in the gnutls code where we were checking for negative values for errors,
1688 when the man pages state that GNUTLS_E_SUCCESS is returned on success and
1689 other values indicate error conditions.
1691 - Bill Egert pointed out (http://curl.haxx.se/bug/view.cgi?id=2671602) that
1692 curl didn't use sprintf() in a way that is documented to work in POSIX but
1693 since we use our own printf() code (from libcurl) that shouldn't be a
1694 problem. Nonetheless I modified the code to not rely on such particular
1695 features and to not cause further raised eyebrowse with no good reason.
1697 Daniel Fandrich (5 Mar 2009)
1698 - Expanded the security section of the libcurl-tutorial man page to cover
1699 more issues for authors to consider when writing robust libcurl-using
1702 Yang Tse (5 Mar 2009)
1703 - Fixed NTLM authentication memory leak on SSPI enabled Windows builds. This
1704 issue was noticed by Chris Deidun.
1706 Daniel Fandrich (4 Mar 2009)
1707 - Fixed a problem with m4 quoting in the OpenSSL configure check reported
1710 Daniel Stenberg (3 Mar 2009)
1711 - David James brought a patch that make libcurl close (all) dead connections
1712 whenever you attempt to open a new connection.
1714 1. After cleaning up a dead connection, "continue" instead of
1715 returning FALSE. This ensures that we clean up all dead connections,
1716 rather than just cleaning up the first dead connection.
1717 2. Move up the cleanup for dead connections so that it occurs for
1718 all connections, rather than just the connections which have the same
1719 preferences as our current new connection.
1721 Version 7.19.4 (3 March 2009)
1723 Daniel Stenberg (3 Mar 2009)
1724 - David Kierznowski notified us about a security flaw
1725 (http://curl.haxx.se/docs/adv_20090303.html also known as CVE-2009-0037) in
1726 which previous libcurl versions (by design) can be tricked to access an
1727 arbitrary local/different file instead of a remote one when
1728 CURLOPT_FOLLOWLOCATION is enabled. This flaw is now fixed in this release
1729 together this the addition of two new setopt options for controlling this
1732 o CURLOPT_REDIR_PROTOCOLS controls what protocols libcurl is allowed to
1733 follow to when CURLOPT_FOLLOWLOCATION is enabled. By default, this option
1734 excludes the FILE and SCP protocols and thus you nee to explicitly allow
1735 them in your app if you really want that behavior.
1737 o CURLOPT_PROTOCOLS controls what protocol(s) libcurl is allowed to fetch
1738 using the primary URL option. This is useful if you want to allow a user or
1739 other outsiders control what URL to pass to libcurl and yet not allow all
1740 protocols libcurl may have been built to support.
1742 Daniel Stenberg (27 Feb 2009)
1743 - Senthil Raja Velu reported a problem when CURLOPT_INTERFACE and
1744 CURLOPT_LOCALPORT were used together (the local port bind failed), and
1745 Markus Koetter provided the fix!
1747 Daniel Stenberg (25 Feb 2009)
1748 - As Daniel Fandrich figured out, we must do the GnuTLS initing in the
1749 curl_global_init() function to properly maintain the performing functions
1750 thread-safe. We've previously (28 April 2007) moved the init to a later time
1751 just to avoid it to fail very early when libgcrypt dislikes the situation,
1752 but that move was bad and the fix should rather be in libgcrypt or
1755 Daniel Stenberg (24 Feb 2009)
1756 - Brian J. Murrell found out that Negotiate proxy authentication didn't work.
1757 It happened because the code used the struct for server-based auth all the
1758 time for both proxy and server auth which of course was wrong.
1760 Daniel Stenberg (23 Feb 2009)
1761 - After a bug reported by James Cheng I've made curl_easy_getinfo() for
1762 CURLINFO_CONTENT_LENGTH_DOWNLOAD and CURLINFO_CONTENT_LENGTH_UPLOAD return
1763 -1 if the sizes aren't know. Previously these returned 0, make it impossible
1764 to detect the difference between actually zero and unknown.
1766 Yang Tse (23 Feb 2009)
1767 - Daniel Johnson provided a shell script that will perform all the steps needed
1768 to build a Mac OS X fat ppc/i386 or ppc64/x86_64 libcurl.framework
1770 Daniel Stenberg (23 Feb 2009)
1771 - I renamed everything in the windows builds files that used the name 'curllib'
1772 to the proper 'libcurl' as clearly this caused confusion.
1774 Yang Tse (20 Feb 2009)
1775 - Do not halt compilation when using VS2008 to build a Windows 2000 target.
1777 Daniel Stenberg (20 Feb 2009)
1778 - Linus Nielsen Feltzing reported and helped me repeat and fix a problem with
1779 FTP with the multi interface: when a transfer fails, like when aborted by a
1780 write callback, the control connection was wrongly closed and thus not
1783 This change is also an attempt to cleanup the code somewhat in this area, as
1784 now the FTP code attempts to keep (better) track on pending responses
1785 necessary to get read in ftp_done().
1787 Daniel Stenberg (19 Feb 2009)
1788 - Patrik Thunstrom reported a problem and helped me repeat it. It turned out
1789 libcurl did a superfluous 1000ms wait when doing SFTP downloads!
1791 We read data with libssh2 while doing the "DO" operation for SFTP and then
1792 when we were about to start getting data for the actual file part, the
1793 "TRANSFER" part, we waited for socket action (in 1000ms) before doing a
1794 libssh2-read. But in this case libssh2 had already read and buffered the
1795 data so we ended up always just waiting 1000ms before we get working on the
1798 Patrick Monnerat (18 Feb 2009)
1799 - FTP downloads (i.e.: RETR) ending with code 550 now return error
1800 CURLE_REMOTE_FILE_NOT_FOUND instead of CURLE_FTP_COULDNT_RETR_FILE.
1802 Daniel Stenberg (17 Feb 2009)
1803 - Kamil Dudka made NSS-powered builds compile and run again!
1805 - A second follow-up change by Andre Guibert de Bruet to fix a related memory
1806 leak like that fixed on the 14th. When zlib returns failure, we need to
1807 cleanup properly before returning error.
1809 - CURLOPT_FTP_CREATE_MISSING_DIRS can now be set to 2 in addition to 1 for
1810 plain FTP connections, and it will then allow MKD to fail once and retry the
1811 CWD afterwards. This is especially useful if you're doing many simultanoes
1812 connections against the same server and they all have this option enabled,
1813 as then CWD may first fail but then another connection does MKD before this
1814 connection and thus MKD fails but trying CWD works! The numbers can
1815 (should?) now be set with the convenience enums now called
1816 CURLFTP_CREATE_DIR and CURLFTP_CREATE_DIR_RETRY.
1818 Tests has proven that if you're making an application that uploads a set of
1819 files to an ftp server, you will get a noticable gain in speed if you're
1820 using multiple connections and this option will be then be very useful.
1822 Daniel Stenberg (14 Feb 2009)
1823 - Andre Guibert de Bruet found and fixed a memory leak in the content encoding
1824 code, which could happen on libz errors.
1826 Daniel Fandrich (12 Feb 2009)
1827 - Added support for Digest and NTLM authentication using GnuTLS.
1829 Daniel Stenberg (11 Feb 2009)
1830 - CURLINFO_CONDITION_UNMET was added to allow an application to get to know if
1831 the condition in the previous request was unmet. This is typically a time
1832 condition set with CURLOPT_TIMECONDITION and was previously not possible to
1833 reliably figure out. From bug report #2565128
1834 (http://curl.haxx.se/bug/view.cgi?id=2565128) filed by Jocelyn Jaubert.
1836 Daniel Fandrich (4 Feb 2009)
1837 - Don't add the standard /usr/lib or /usr/include paths to LDFLAGS and CPPFLAGS
1838 (respectively) when --with-ssl=/usr is used (patch based on FreeBSD).
1840 - Added an explicit buffer limit check in msdosify() (patch based on FreeBSD).
1841 This couldn't ever overflow in curl, but might if the code were used
1842 elsewhere or under different conditions.
1844 Daniel Stenberg (3 Feb 2009)
1845 - Hidemoto Nakada provided a small fix that makes it possible to get the
1846 CURLINFO_CONTENT_LENGTH_DOWNLOAD size from file:// "transfers" with
1847 CURLOPT_NOBODY set true.
1849 Daniel Stenberg (2 Feb 2009)
1850 - Patrick Scott found a rather large memory leak when using the multi
1851 interface and setting CURLMOPT_MAXCONNECTS to something less than the number
1852 of handles you add to the multi handle. All the connections that didn't fit
1853 in the cache would not be properly disconnected nor freed!
1855 - Craig A West brought us: libcurl now defaults to do CONNECT with HTTP
1856 version 1.1 instead of 1.0 like before. This change also introduces the new
1857 proxy type for libcurl called 'CURLPROXY_HTTP_1_0' that then allows apps to
1858 switch (back) to CONNECT 1.0 requests. The curl tool also got a --proxy1.0
1859 option that works exactly like --proxy but sets CURLPROXY_HTTP_1_0.
1861 I updated all test cases cases that use CONNECT and I tried to do some using
1862 --proxy1.0 and some updated to do CONNECT 1.1 to get both versions run.
1864 Daniel Stenberg (31 Jan 2009)
1865 - When building with c-ares 1.6.1 (not yet released) or later and IPv6 support
1866 enabled, we can now take advantage of its brand new AF_UNSPEC support in
1867 ares_gethostbyname(). This makes test case 241 finally run fine for me with
1868 this setup since it now parses the "::1 ip6-localhost" line fine in my
1871 Daniel Stenberg (30 Jan 2009)
1872 - Scott Cantor filed bug report #2550061
1873 (http://curl.haxx.se/bug/view.cgi?id=2550061) mentioning that I failed to
1874 properly make sure that the VC9 makefiles got included in the latest
1875 release. I've now fixed the release script and verified it so next release
1876 will hopefully include them properly!
1878 Daniel Fandrich (30 Jan 2009)
1879 - Fixed --disable-proxy for FTP and SOCKS. Thanks to Daniel Egger for
1882 Yang Tse (29 Jan 2009)
1883 - Introduced curl_sspi.c and curl_sspi.h for the implementation of functions
1884 Curl_sspi_global_init() and Curl_sspi_global_cleanup() which previously were
1885 named Curl_ntlm_global_init() and Curl_ntlm_global_cleanup() in http_ntlm.c
1886 Also adjusted socks_sspi.c to remove the link-time dependency on the Windows
1887 SSPI library using it now in the same way as it was done in http_ntlm.c.
1889 Daniel Stenberg (28 Jan 2009)
1890 - Markus Moeller introduced two new options to libcurl:
1891 CURLOPT_SOCKS5_GSSAPI_SERVICE and CURLOPT_SOCKS5_GSSAPI_NEC to allow libcurl
1892 to do GSS-style authentication with SOCKS5 proxies. The curl tool got the
1893 options called --socks5-gssapi-service and --socks5-gssapi-nec to enable
1896 Daniel Stenberg (26 Jan 2009)
1897 - Chad Monroe provided the new CURLOPT_TFTP_BLKSIZE option that allows an app
1898 to set desired block size to use for TFTP transfers instead of the default
1901 - The "-no_ticket" option was introduced in Openssl0.9.8j. It's a flag to
1902 disable "rfc4507bis session ticket support". rfc4507bis was later turned
1903 into the proper RFC5077 it seems: http://tools.ietf.org/html/rfc5077
1905 The enabled extension concerns the session management. I wonder how often
1906 libcurl stops a connection and then resumes a TLS session. also, sending the
1907 session data is some overhead. .I suggest that you just use your proposed
1908 patch (which explicitly disables TICKET).
1910 If someone writes an application with libcurl and openssl who wants to
1911 enable the feature, one can do this in the SSL callback.
1913 Sharad Gupta brought this to my attention. Peter Sylvester helped me decide
1914 on the proper action.
1916 - Alexey Borzov filed bug report #2535504
1917 (http://curl.haxx.se/bug/view.cgi?id=2535504) pointing out that realms with
1918 quoted quotation marks in HTTP Digest headers didn't work. I've now added
1919 test case 1095 that verifies my fix.
1921 - Craig A West brought CURLOPT_NOPROXY and the corresponding --noproxy option.
1922 They basically offer the same thing the NO_PROXY environment variable only
1923 offered previously: list a set of host names that shall not use the proxy
1924 even if one is specified.
1926 Daniel Fandrich (20 Jan 2009)
1927 - Call setlocale() for libtest tests to test the effects of locale-induced
1928 libc changes on libcurl.
1930 - Fixed a couple more locale-dependent toupper conversions, mainly for
1931 clarity. This does fix one problem that causes ;type=i FTP URLs
1932 to fail in the Turkish locale when CURLOPT_PROXY_TRANSFER_MODE is
1933 used (test case 561)
1935 - Added tests 561 and 1091 through 1094 to test various combinations
1936 of ;type= and ;mode= URLs that could potentially fail in the Turkish
1939 Daniel Stenberg (20 Jan 2009)
1940 - Lisa Xu pointed out that the ssh.obj file was missing from the
1941 lib/Makefile.vc6 file (and thus from the vc8 and vc9 ones too).
1943 Version 7.19.3 (19 January 2009)
1945 Daniel Stenberg (16 Jan 2009)
1946 - Andrew de los Reyes fixed curlbuild.h for "generic" gcc builds on PPC, both
1949 Daniel Stenberg (15 Jan 2009)
1950 - Tim Ansell fixed a compiler warning in lib/cookie.c
1952 Daniel Stenberg (14 Jan 2009)
1953 - Grant Erickson fixed timeouts for TFTP such that specifying a
1954 connect-timeout, a max-time or both options work correctly and as expected
1955 by passing the correct boolean value to Curl_timeleft via the
1956 'duringconnect' parameter.
1958 With this small change, curl TFTP now behaves as expected (and likely as
1959 originally-designed):
1961 1) For non-existent or unreachable dotted IP addresses:
1963 a) With no options, follows the default curl 300s timeout...
1964 b) With --connect-timeout only, follows that value...
1965 c) With --max-time only, follows that value...
1966 d) With both --connect-timeout and --max-time, follows the smaller value...
1968 and times out with a "curl: (7) Couldn't connect to server" error.
1970 2) For transfers to/from a valid host:
1972 a) With no options, follows default curl 300s timeout for the
1973 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1974 timeout for the remainder of the transfer...
1976 b) With --connect-time only, follows that value for the
1977 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1978 timeout for the remainder of the transfer...
1980 c) With --max-time only, follows that value for the first
1981 XRQ/DATA/ACK transaction and for the remainder of the
1984 d) With both --connect-timeout and --max-time, follows the former
1985 for the first XRQ/DATA/ACK transaction and the latter for the
1986 remainder of the transfer...
1988 and times out with a "curl: (28) Timeout was reached" error as
1991 Daniel Stenberg (13 Jan 2009)
1992 - Michael Wallner fixed a NULL pointer deref when calling
1993 curl_easy_setup(curl, CURLOPT_COOKIELIST, "SESS") on a CURL handle with no
1996 - Stefan Teleman brought a patch to fix the default curlbuild.h file for the
1999 Daniel Stenberg (12 Jan 2009)
2000 - Based on bug report #2498665 (http://curl.haxx.se/bug/view.cgi?id=2498665)
2001 by Daniel Black, I've now added magic to the configure script that makes it
2002 use pkg-config to detect gnutls details as well if the existing method
2003 (using libgnutls-config) fails. While doing this, I cleaned up and unified
2004 the pkg-config usage when detecting openssl and nss as well.
2006 Daniel Stenberg (11 Jan 2009)
2007 - Karl Moerder brought the patch that creates vc9 Makefiles, and I made
2008 'maketgz' now use the actual makefile targets to do the VC8 and VC9
2011 Daniel Stenberg (10 Jan 2009)
2012 - Emil Romanus fixed:
2014 When using the multi interface over HTTP and the server returns a Location
2015 header, the running easy handle will get stuck in the CURLM_STATE_PERFORM
2016 state, leaving the external event loop stuck waiting for data from the
2017 ingoing socket (when using the curl_multi_socket_action stuff). While this
2018 bug was pretty hard to find, it seems to require only a one-line fix. The
2019 break statement on line 1374 in multi.c caused the function to skip the call
2022 How to reproduce this bug? Well, that's another question. evhiperfifo.c in
2023 the examples directory chokes on this bug only _sometimes_, probably
2024 depending on how fast the URLs are added. One way of testing the bug out is
2025 writing to hiper.fifo from more than one source at the same time.
2027 Daniel Fandrich (7 Jan 2009)
2028 - Unified much of the SessionHandle initialization done in Curl_open() and
2029 curl_easy_reset() by creating Curl_init_userdefined(). This had the side
2030 effect of fixing curl_easy_reset() so it now also resets
2031 CURLOPT_FTP_FILEMETHOD and CURLOPT_SSL_SESSIONID_CACHE
2033 Daniel Stenberg (7 Jan 2009)
2034 - Rob Crittenden did once again provide an NSS update:
2036 I have to jump through a few hoops now with the NSS library initialization
2037 since another part of an application may have already initialized NSS by the
2038 time Curl gets invoked. This patch is more careful to only shutdown the NSS
2039 library if Curl did the initialization.
2041 It also adds in a bit of code to set the default ciphers if the app that
2042 call NSS_Init* did not call NSS_SetDomesticPolicy() or set specific
2043 ciphers. One might argue that this lets other application developers get
2044 lazy and/or they aren't using the NSS API correctly, and you'd be right.
2045 But still, this will avoid terribly difficult-to-trace crashes and is
2048 Daniel Stenberg (1 Jan 2009)
2049 - 'reconf' is removed since we rather have users use 'buildconf'
2051 Daniel Stenberg (31 Dec 2008)
2052 - Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
2053 out that 'reconf' didn't properly point out the m4 subdirectory when running
2056 Daniel Stenberg (29 Dec 2008)
2057 - Phil Lisiecki filed bug report #2413067
2058 (http://curl.haxx.se/bug/view.cgi?id=2413067) that identified a problem that
2059 would cause libcurl to mark a DNS cache entry "in use" eternally if the
2060 subsequence TCP connect failed. It would thus never get pruned and refreshed
2061 as it should've been.
2063 Phil provided his own patch to this problem that while it seemed to work
2064 wasn't complete and thus I wrote my own fix to the problem.
2066 Daniel Stenberg (28 Dec 2008)
2067 - Peter Korsgaard fixed building libcurl with "configure --with-ssl
2070 - Anthony Bryan fixed more language and spelling flaws in man pages.
2072 Daniel Stenberg (22 Dec 2008)
2073 - Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
2074 on file indexes beyond 2 or 4GB.
2076 - Anthony Bryan provided a set of patches that cleaned up manual language,
2077 corrected spellings and more.
2079 Daniel Stenberg (20 Dec 2008)
2080 - Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
2081 pipelining, as libcurl could then easily get confused and A) work on the
2082 handle that was not "first in queue" on a pipeline, or even B) tell the app
2083 to REMOVE a socket while it was in use by a second handle in a pipeline. Both
2084 errors caused hanging or stalling applications.
2086 Daniel Stenberg (19 Dec 2008)
2087 - curl_multi_timeout() could return a timeout value of 0 even though nothing
2088 was actually ready to get done, as the internal time resolution is higher
2089 than the returned millisecond timer. Therefore it could cause applications
2090 running on fast processors to do short bursts of busy-loops.
2091 curl_multi_timeout() will now only return 0 if the timeout is actually
2094 - Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
2095 now has an improved ability to do right when the multi interface (both
2096 "regular" and multi_socket) is used for SCP and SFTP transfers. This should
2097 result in (much) less busy-loop situations and thus less CPU usage with no
2100 Daniel Stenberg (17 Dec 2008)
2101 - SCP and SFTP with the multi interface had the same flaw: the 'DONE'
2102 operation didn't complete properly if the EAGAIN equivalent was returned but
2103 libcurl would simply continue with a half-completed close operation
2104 performed. This ruined persistent connection re-use and cause some
2105 SSH-protocol errors in general. The correction is unfortunately adding a
2106 blocking function - doing it entirely non-blocking should be considered for
2109 Gisle Vanem (16 Dec 2008)
2110 - Added the possibility to use the Watt-32 tcp/ip stack under Windows.
2111 The change simply involved adding a USE_WATT32 section in the
2112 config-win32.h files (under ./lib and ./src). This section disables
2113 the use of any Winsock headers.
2115 Daniel Stenberg (16 Dec 2008)
2116 - libssh2_sftp_last_error() was wrongly used at some places in libcurl which
2117 made libcurl sometimes not properly abort problematic SFTP transfers.
2119 Daniel Stenberg (12 Dec 2008)
2120 - More work with Igor Novoseltsev to first fix the remaining stuff for
2121 removing easy handles from multi handles when the easy handle is/was within
2122 a HTTP pipeline. His bug report #2351653
2123 (http://curl.haxx.se/bug/view.cgi?id=2351653) was also related and was
2124 eventually fixed by a patch by Igor himself.
2126 Yang Tse (12 Dec 2008)
2127 - Patrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
2128 OS/400 compilations with IPv6 enabled.
2130 Daniel Stenberg (12 Dec 2008)
2131 - Mark Karpeles filed bug report #2416182 titled "crash in ConnectionExists
2132 when using duphandle+curl_mutli"
2133 (http://curl.haxx.se/bug/view.cgi?id=2416182) which showed that
2134 curl_easy_duphandle() wrongly also copied the pointer to the connection
2135 cache, which was plain wrong and caused a segfault if the handle would be
2136 used in a different multi handle than the handle it was duplicated from.
2138 Daniel Stenberg (11 Dec 2008)
2139 - Keshav Krity found out that libcurl failed to deal with dotted IPv6
2140 addresses if they were very long (>39 letters) due to a too strict address
2141 validity parser. It now accepts addresses up to 45 bytes long.
2143 Daniel Stenberg (11 Dec 2008)
2144 - Internet Explorer had a broken HTTP digest authentication before v7 and
2145 there are servers "out there" that relies on the client doing this broken
2146 Digest authentication. Apache even comes with an option to work with such
2149 The difference is only for URLs that contain a query-part (a '?'-letter and
2150 text to the right of it).
2152 libcurl now supports this quirk, and you enable it by setting the
2153 CURLAUTH_DIGEST_IE bit in the bitmask you pass to the CURLOPT_HTTPAUTH or
2154 CURLOPT_PROXYAUTH options. They are thus individually controlled to server
2157 (note that there's no way to activate this with the curl tool yet)
2159 Daniel Fandrich (9 Dec 2008)
2160 - Added test cases 1089 and 1090 to test --write-out after a redirect to
2161 test a report that the size didn't work, but these test cases pass.
2163 - Documented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.
2165 Daniel Stenberg (9 Dec 2008)
2166 - Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
2167 particular state for the control connection like it did before for implicit
2168 FTPS (libcurl assumed such control connections to be encrypted while some
2169 FTPS servers such as FileZilla assumes such connections to be clear
2170 mode). Use the CURLOPT_USE_SSL option to set your desired level.
2172 Daniel Stenberg (8 Dec 2008)
2173 - Fred Machado posted about a weird FTP problem on the curl-users list and when
2174 researching it, it turned out he got a 550 response back from a SIZE command
2175 and then I fell over the text in RFC3659 that says:
2177 The presence of the 550 error response to a SIZE command MUST NOT be taken
2178 by the client as an indication that the file cannot be transferred in the
2179 current MODE and TYPE.
2181 In other words: the change I did on September 30th 2008 and that has been
2182 included in the last two releases were a regression and a bad idea. We MUST
2183 NOT take a 550 response from SIZE as a hint that the file doesn't exist.
2185 - Christian Krause filed bug #2221237
2186 (http://curl.haxx.se/bug/view.cgi?id=2221237) that identified an infinite
2187 loop during GSS authentication given some specific conditions. With his
2188 patience and great feedback I managed to narrow down the problem and
2189 eventually fix it although I can't test any of this myself!
2191 Daniel Fandrich (3 Dec 2008)
2192 - Fixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
2193 support (e.g. Minix)
2195 Daniel Stenberg (3 Dec 2008)
2196 - Igor Novoseltsev filed bug #2351645
2197 (http://curl.haxx.se/bug/view.cgi?id=2351645) that identified a problem with
2198 the multi interface that occured if you removed an easy handle while in
2199 progress and the handle was used in a HTTP pipeline.
2201 - Pawel Kierski pointed out a mistake in the cookie code that could lead to a
2202 bad fclose() after a fatal error had occured.
2203 (http://curl.haxx.se/bug/view.cgi?id=2382219)
2205 Daniel Fandrich (25 Nov 2008)
2206 - If a HTTP request is Basic and num is already >=1000, the HTTP test
2207 server adds 1 to num to get the data section to return. This allows
2208 testing authentication negotiations using the Basic authentication
2211 - Added tests 1087 and 1088 to test Basic authentication on a redirect
2212 with and without --location-trusted
2214 Daniel Stenberg (24 Nov 2008)
2215 - Based on a patch by Vlad Grachov, libcurl now uses a new libssh2 0.19
2216 function when built to support SCP and SFTP that helps the library to know
2217 in which direction a particular libssh2 operation would return EAGAIN so
2218 that libcurl knows what socket conditions to wait for before trying the
2219 function call again. Previously (and still when using libssh2 0.18 or
2220 earlier), libcurl will busy-loop in this situation when the easy interface
2223 Daniel Fandrich (20 Nov 2008)
2224 - Automatically detect OpenBSD's CA cert bundle.
2226 Daniel Stenberg (19 Nov 2008)
2227 - I removed the default use of "Pragma: no-cache" from libcurl when a proxy is
2228 used. It has been used since forever but it was never a good idea to use
2229 unless explicitly asked for.
2231 - Josef Wolf's extension that allows a $TESTDIR/gdbinit$testnum file that when
2232 you use runtests.pl -g, will be sourced by gdb to allow additional fancy or
2233 whatever you see fit
2235 - Christian Krause reported and fixed a memory leak that would occur with HTTP
2236 GSS/kerberos authentication (http://curl.haxx.se/bug/view.cgi?id=2284386)
2238 - Andreas Wurf and Markus Koetter helped me analyze a problem that Andreas got
2239 when uploading files to a single FTP server using multiple easy handle
2240 handles with the multi interface. Occasionally a handle would stall in
2243 The problem turned out to be a side-effect of the ConnectionExists()
2244 function's eagerness to re-use a handle for HTTP pipelining so it would
2245 select it even if already being in use, due to an inadequate check for its
2246 chances of being used for pipelnining.
2248 Daniel Fandrich (17 Nov 2008)
2249 - Added more compiler warning options for gcc 4.3
2251 Yang Tse (17 Nov 2008)
2252 - Fix a remaining problem in the inet_pton() runtime configure check. And
2253 fix internal Curl_inet_pton() failures to reject certain malformed literals.
2255 - Make configure script check if ioctl with the SIOCGIFADDR command can be
2256 used, and define HAVE_IOCTL_SIOCGIFADDR if appropriate.
2258 Daniel Stenberg (16 Nov 2008)
2259 - Christian Krause fixed a build failure when building with gss support
2260 enabled and FTP disabled.
2262 - Added check for NULL returns from strdup() in src/main.c and lib/formdata.c
2263 - reported by Jim Meyering also prevent buffer overflow on MSDOS when you do
2264 for example -O on a url with a file name part longer than PATH_MAX letters
2266 - lib/nss.c fixes based on the report by Jim Meyering: I went over and added
2267 checks for return codes for all calls to malloc and strdup that were
2268 missing. I also changed a few malloc(13) to use arrays on the stack and a
2269 few malloc(PATH_MAX) to instead use aprintf() to lower memory use.
2271 - I fixed a memory leak in Curl_nss_connect() when CURLOPT_ISSUERCERT is
2274 Daniel Fandrich (14 Nov 2008)
2275 - Added .xml as one of the few common file extensions known by the multipart
2278 - Added some #ifdefs around header files and change the EAGAIN test to
2279 fix compilation on Cell (reported by Jeff Curley).
2281 Yang Tse (14 Nov 2008)
2282 - Fixed several configure script issues affecting checks for inet_ntoa_r(),
2283 inet_ntop(), inet_pton(), getifaddrs(), fcntl() and getaddrinfo().
2285 Yang Tse (13 Nov 2008)
2286 - Refactored configure script detection of functions used to set sockets into
2287 non-blocking mode, and decouple function detection from function capability.