5 \___|\___/|_| \_\_____|
9 Daniel Stenberg (16 Apr 2010)
10 - The recent overhaul of the SSL recv function made the GnuTLS specific code
11 treat a zero returned from gnutls_record_recv() as an error, and this caused
12 our HTTPS test cases to fail. We leave it to upper layer code to detect if
13 an EOF is a problem or not.
15 - I reverted the resolver fix from yesterday and instead removed all uses of
16 AI_CANONNAME all over libcurl and made the only user of that info (krb5.c)
17 use the host name from the URL instead. No reverse resolving is a good
20 - Paul Howarth made configure properly detect GSS "on ancient Linux distros"
21 by editing in which order we use headers to detect GSS.
23 Daniel Stenberg (15 Apr 2010)
24 - Rainer Canavan filed bug report #2987196 that identified libcurl doing
25 unnecesary reverse name lookups in many cases when built to use IPv4 and
26 getaddrinfo(). The logic for ipv6 is now used for ipv4 too.
28 (http://curl.haxx.se/bug/view.cgi?id=2963679)
30 Version 7.20.1 (14 April 2010)
32 Daniel Stenberg (9 Apr 2010)
33 - Prefixing the FTP quote commands with an asterisk really only worked for the
34 postquote actions. This is now fixed and test case 227 has been extended to
37 Kamil Dudka (4 Apr 2010)
38 - Eliminated a race condition in Curl_resolv_timeout().
40 - Refactorized interface of Curl_ssl_recv()/Curl_ssl_send().
42 - libcurl-NSS now provides more accurate messages and error codes in case of
43 client certificate problem. Either during connection, or transfer phase.
45 Daniel Stenberg (1 Apr 2010)
46 - Matt Wixson found and fixed a bug in the SCP/SFTP area where the code
47 treated a 0 return code from libssh2 to be the same as EAGAIN while in
48 reality it isn't. The problem caused a hang in SFTP transfers from a
51 Daniel Stenberg (28 Mar 2010)
52 - Ben Greear: If you pass a URL to pop3 that does not contain a message ID as
53 part of the URL, it would previously ask for 'INBOX' which just causes the
54 pop3 server to return an error.
56 Now libcurl treats en empty message ID as a request for LIST (list of pop3
57 message IDs). User's code could then parse this and download individual
60 Daniel Stenberg (27 Mar 2010)
61 - Ben Greear brought a patch that from now on allows all protocols to specify
62 name and user within the URL, in the same manner HTTP and FTP have been
63 allowed to in the past - although far from all of the libcurl supported
64 protocls actually have that feature in their URL definition spec.
66 Daniel Stenberg (26 Mar 2010)
67 - Ben Greear brought code that makes the rate limiting code for the easy
68 interface a bit smoother as it introduces sub-second sleeps during it and it
69 also takes the buffer sizes into account.
71 Daniel Stenberg (24 Mar 2010)
72 - Bob Richmond: There's an annoying situation where libcurl will read new HTTP
73 response data from a socket, then check if it's a timeout if one is set. If
74 the last packet received constitutes the end of the response body, libcurl
75 still treats it as a timeout condition and reports a message like:
77 "Operation timed out after 3000 milliseconds with 876 out of 876 bytes
80 It should only a timeout if the timer lapsed and we DIDN'T receive the end
81 of the response body yet.
83 - Christopher Conroy fixed a problem with RTSP and GET_PARAMETER reported
84 to us by Massimo Callegari. There's a new test case 572 that verifies this
87 - The 'ares' subtree has been removed from the source repository. It was
88 always a separate project that sort of piggybacked on the curl project since
89 the dawn of times and now the time has come for it to go stand on its own
90 legs and continue living its own life. All details on c-ares and its new
91 source code repository is found at http://c-ares.haxx.se/
93 Daniel Stenberg (23 Mar 2010)
94 - Kenny To filed the bug report #2963679 with patch to fix a problem he
95 experienced with doing multi interface HTTP POST over a proxy using
96 PROXYTUNNEL. He found a case where it would connect fine but bits.tcpconnect
97 was not set correct so libcurl didn't work properly.
99 (http://curl.haxx.se/bug/view.cgi?id=2963679)
101 - Akos Pasztory filed debian bug report #572276
102 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=572276 mentioning a problem
103 with a resource that returns chunked-encoded _and_ with a Content-Length
104 and libcurl failed to properly ignore the latter information.
106 - Hauke Duden provided an example program that made the multi interface crash.
107 His example simply used the multi interface and did first one FTP transfer
108 and after completion it used a second easy handle and did another FTP
109 transfer on the same FTP server.
111 This triggered a bug in the "delayed easy handle kill" system that curl
112 uses: when an FTP connection is left alive it must keep an easy handle
113 around internally - only for the purpose of having an easy handle when it
114 later disconnects it. The code assumed that when the easy handle was removed
115 and an internal reference was made, that version could be killed later on
116 when a new easy handle came using the same connection. This was wrong as
117 Hauke's example showed that the removed handle wasn't killed for real until
118 later. This caused a double close attempt => segfault.
120 Daniel Stenberg (22 Mar 2010)
121 - Thomas Lopatic fixed the alarm()-based DNS timeout:
123 Looking at the code of Curl_resolv_timeout() in hostip.c, I think that in
124 case of a timeout, the signal handler for SIGALRM never gets removed. I
125 think that in my case it gets executed at some point later on when execution
126 has long left Curl_resolv_timeout() or even the cURL library.
128 The code that is jumped to with siglongjmp() simply sets the error message
129 to "name lookup timed out" and then returns with CURLRESOLV_ERROR. I guess
130 that instead of simply returning without cleaning up, the code should have a
131 goto that jumps to the spot right after the call to Curl_resolv().
133 Kamil Dudka (22 Mar 2010)
134 - Douglas Steinwand contributed a patch fixing insufficient initialization in
135 Curl_clone_ssl_config()
137 Daniel Stenberg (21 Mar 2010)
138 - Ben Greear improved TFTP: the error code returning and the treatment
139 of TSIZE == 0 when uploading.
141 - We've switched from CVS to git. See http://curl.haxx.se/source.html
143 Kamil Dudka (19 Mar 2010)
144 - Improved Curl_read() to not ignore the error returned from Curl_ssl_recv().
146 Daniel Stenberg (15 Mar 2010)
147 - Constantine Sapuntzakis brought a patch:
149 The problem mentioned on Dec 10 2009
150 (http://curl.haxx.se/bug/view.cgi?id=2905220) was only partially fixed.
151 Partially because an easy handle can be associated with many connections in
152 the cache (e.g. if there is a redirect during the lifetime of the easy
153 handle). The previous patch only cleaned up the first one. The new fix now
154 removes the easy handle from all connections, not just the first one.
156 Daniel Stenberg (6 Mar 2010)
157 - Ben Greear brought a patch that fixed the rate limiting logic for TFTP when
158 the easy interface was used.
160 Daniel Stenberg (5 Mar 2010)
161 - Daniel Johnson provided fixes for building curl with the clang compiler.
163 Yang Tse (5 Mar 2010)
164 - Constantine Sapuntzakis detected and fixed a double free in builds done
165 with threaded resolver enabled (Windows default configuration) that would
166 get triggered when a curl handle is closed while doing DNS resolution.
168 Daniel Stenberg (2 Mar 2010)
169 - [Daniel Johnson] I've been trying to build libcurl with clang on Darwin and
170 ran into some issues with the GSSAPI tests in configure.ac. The tests first
171 try to determine the include dirs and libs and set CPPFLAGS and LIBS
172 accordingly. It then checks for the headers and finally sets LIBS a second
173 time, causing the libs to be included twice. The first setting of LIBS seems
174 redundant and should be left out, since the first part is otherwise just
175 about finding headers.
177 My second issue is that 'krb5-config --libs gssapi' on Darwin is less than
178 useless and returns junk that, while it happens to work with gcc, causes
179 clang to choke. For example, --libs returns $CFLAGS along with the libs,
180 which is really retarded. Simply setting 'LIBS="$LIBS -lgssapi_krb5
181 -lresolv"' on Darwin is sufficient.
183 - Based on patch provided by Jacob Moshenko, the transfer logic now properly
184 makes sure that when using sub-second timeouts, there's no final bad 1000ms
185 wait. Previously, a sub-second timeout would often make the elapsed time end
186 up the time rounded up to the nearest second (e.g. 1s for 200ms timeout)
188 - Andrei Benea filed bug report #2956698 and pointed out that the
189 CURLOPT_CERTINFO feature leaked memory due to a missing OpenSSL function
190 call. He provided the patch to fix it too.
192 http://curl.haxx.se/bug/view.cgi?id=2956698
194 - Markus Duft pointed out in bug #2961796 that even though Interix has a
195 poll() function it doesn't quite work the way we want it so we must disable
196 it, and he also provided a patch for it.
198 http://curl.haxx.se/bug/view.cgi?id=2961796
200 - Made the pingpong timeout code properly deal with the response timeout AND
201 the global timeout if set. Also, as was reported in the bug report #2956437
202 by Ryan Chan, the time stamp to use as basis for the per command timeout was
203 not set properly in the DONE phase for FTP (and not for SMTP) so I fixed
204 that just now. This was a regression compared to 7.19.7 due to the
205 conversion of FTP code over to the generic pingpong concepts.
207 http://curl.haxx.se/bug/view.cgi?id=2956437
209 Daniel Stenberg (1 Mar 2010)
210 - Ben Greear provided an update for TFTP that fixes upload.
212 - Wesley Miaw reported bug #2958179 which identified a case of looping during
213 OpenSSL based SSL handshaking even though the multi interface was used and
214 there was no good reason for it.
216 http://curl.haxx.se/bug/view.cgi?id=2958179
218 Daniel Stenberg (26 Feb 2010)
219 - Pat Ray in bug #2958474 pointed out an off-by-one case when receiving a
220 chunked-encoding trailer.
222 http://curl.haxx.se/bug/view.cgi?id=2958474
224 Daniel Fandrich (25 Feb 2010)
225 - Fixed a couple of out of memory leaks and a segfault in the SMTP & IMAP code.
227 Yang Tse (25 Feb 2010)
228 - I fixed bug report #2958074 indicating
229 (http://curl.haxx.se/bug/view.cgi?id=2958074) that curl on Windows with
230 option --trace-time did not use local time when timestamping trace lines.
231 This could also happen on other systems depending on time souurce.
233 Patrick Monnerat (22 Feb 2010)
234 - Proper handling of STARTTLS on SMTP, taking CURLUSESSL_TRY into account.
235 - SMTP falls back to RFC821 HELO when EHLO fails (and SSL is not required).
236 - Use of true local host name (i.e.: via gethostname()) when available, as
237 default argument to SMTP HELO/EHLO.
238 - Test case 804 for HELO fallback.
240 Daniel Stenberg (20 Feb 2010)
241 - Fixed the SMTP compliance by making sure RCPT TO addresses are specified
242 properly in angle brackets. Recipients provided with CURLOPT_MAIL_RCPT now
243 get angle bracket wrapping automatically by libcurl unless the recipient
244 starts with an angle bracket as then the app is assumed to deal with that
247 - I made the SMTP code expect a 250 response back from the server after the
248 full DATA has been sent, and I modified the test SMTP server to also send
249 that response. As usual, the DONE operation that is made after a completed
250 transfer is still not doable in a non-blocking way so this waiting for 250
251 is unfortunately made blockingly.
253 Yang Tse (14 Feb 2010)
254 - Overhauled test suite getpart() function. Fixing potential out of bounds
255 stack and memory overwrites triggered with huge test case definitions.
257 Daniel Stenberg (13 Feb 2010)
258 - Martin Hager reported and fixed a problem with a missing quote in libcurl.m4
260 (http://curl.haxx.se/bug/view.cgi?id=2951319)
262 - Tom Donovan fixed the CURL_FORMAT_* defines when building with cmake.
264 (http://curl.haxx.se/bug/view.cgi?id=2951269)
266 Daniel Stenberg (12 Feb 2010)
267 - Jack Zhang reported a problem with SMTP: we wrongly used multiple addresses
268 in the same RCPT TO line, when they should be sent in separate single
269 commands. I updated test case 802 to verify this.
271 - I also fixed a bad use of my_setopt_str() of CURLOPT_MAIL_RCPT in the curl
272 tool which made it try to output it as string for the --libcurl feature
273 which could lead to crashes.
275 Yang Tse (11 Feb 2010)
276 - Steven M. Schweda fixed VMS builder bad behavior when used in a batch job,
277 removed obsolete batch_compile.com and defines.com and updated VMS readme.
279 Version 7.20.0 (9 February 2010)
281 Daniel Stenberg (9 Feb 2010)
282 - When downloading compressed content over HTTP and the app asked libcurl to
283 automatically uncompress it with the CURLOPT_ENCODING option, libcurl could
284 wrongly provide the callback with more data than the maximum documented
285 amount. An application could thus get tricked into badness if the maximum
286 limit was trusted to be enforced by libcurl itself (as it is documented).
288 This is further detailed and explained in the libcurl security advisory
291 http://curl.haxx.se/docs/adv_20100209.html
293 Daniel Fandrich (3 Feb 2010)
294 - Changed the Watcom makefiles to make them easier to keep in sync with
295 Makefile.inc since that can't be included directly.
297 Yang Tse (2 Feb 2010)
298 - Symbol CURL_FORMAT_OFF_T now obsoleted, will be removed in a future release,
299 symbol will not be available when building with CURL_NO_OLDIES defined. Use
300 of CURL_FORMAT_CURL_OFF_T is preferred since 7.19.0
302 Daniel Stenberg (1 Feb 2010)
303 - Using the multi_socket API, it turns out at times it seemed to "forget"
304 connections (which caused a hang). It turned out to be an existing (7.19.7)
305 bug in libcurl (that's been around for a long time) and it happened like
308 The app calls curl_multi_add_handle() to add a new easy handle, libcurl will
309 then set it to timeout in 1 millisecond so libcurl will tell the app about
312 The app's timeout fires off that there's a timeout, the app calls libcurl as
313 we so often document it:
316 res = curl_multi_socket_action(... TIMEOUT ...);
317 } while(CURLM_CALL_MULTI_PERFORM == res);
319 And this is the problem number one:
321 When curl_multi_socket_action() is called with no specific handle, but only
322 a timeout-action, it will *only* perform actions within libcurl that are
323 marked to run at this time. In this case, the request would go from INIT to
324 CONNECT and return CURLM_CALL_MULTI_PERFORM. When the app then calls libcurl
325 again, there's no timer set for this handle so it remains in the CONNECT
326 state. The CONNECT state is a transitional state in libcurl so it reports no
327 sockets there, and thus libcurl never tells the app anything more about that
328 easy handle/connection.
330 libcurl _does_ set a 1ms timeout for the handle at the end of
331 multi_runsingle() if it returns CURLM_CALL_MULTI_PERFORM, but since the loop
332 is instant the new job is not ready to run at that point (and there's no
333 code that makes libcurl call the app to update the timout for this new
334 timeout). It will simply rely on that some other timeout will trigger later
335 on or that something else will update the timeout callback. This makes the
336 bug fairly hard to repeat.
338 The fix made to adress this issue:
340 We introduce a loop in lib/multi.c around all calls to multi_runsingle() and
341 simply check for CURLM_CALL_MULTI_PERFORM internally. This has the added
342 benefit that this goes in line with my long-term wishes to get rid of the
343 CURLM_CALL_MULTI_PERFORM all together from the public API.
345 The downside of this fix, is that the counter we return in 'running_handles'
346 in several of our public functions then gets a slightly new and possibly
347 confusing behavior during times:
349 If an app adds a handle that fails to connect (very quickly) it may just
350 as well never appear as a 'running_handle' with this fix. Previously it
351 would first bump the counter only to get it decreased again at next call.
352 Even I have used that change in handle counter to signal "end of a
353 transfer". The only *good* way to find the end of a individual transfer
354 is calling curl_multi_info_read() to see if it returns one.
356 Of course, if the app previously did the looping before it checked the
357 counter, it really shouldn't be any new effect.
359 Yang Tse (26 Jan 2010)
360 - Constantine Sapuntzakis' and Joshua Kwan's work done in the last four months
361 relative to the asynchronous DNS lookups, along with with some integration
362 adjustments I have done are finally committed to CVS.
364 Currently these enhancements will benefit builds done using c-ares on any
365 platform as well as Windows builds using the default threaded resolver.
367 This release does not make generally available POSIX threaded DNS lookups
368 yet. There is no configure option to enable this feature yet. It is possible
369 to experimantally try this feature running configure with compiler flags that
370 make simultaneous definition of preprocessor symbols USE_THREADS_POSIX and
371 HAVE_PTHREAD_H, as well as whatever reentrancy compiler flags and linker ones
372 are required to link and properly use pthread_* functions on each platform.
374 Daniel Stenberg (26 Jan 2010)
375 - Mike Crowe made libcurl return CURLE_COULDNT_RESOLVE_PROXY when it is the
376 proxy that cannot be resolved when using c-ares. This matches the behaviour
377 when not using c-ares.
379 Björn Stenberg (23 Jan 2010)
380 - Added a new flag: -J/--remote-header-name. This option tells the
381 -O/--remote-name option to use the server-specified Content-Disposition
382 filename instead of extracting a filename from the URL.
384 Daniel Stenberg (21 Jan 2010)
385 - Chris Conroy brought support for RTSP transfers, and with it comes 8(!) new
386 libcurl options for controlling what to get and how to receive posssibly
387 interleaved RTP data.
389 Daniel Stenberg (20 Jan 2010)
390 - As was pointed out on the http-state mailing list, the order of cookies in a
391 HTTP Cookie: header _needs_ to be sorted on the path length in the cases
392 where two cookies using the same name are set more than once using
393 (overlapping) paths. Realizing this, identically named cookies must be
394 sorted correctly. But detecting only identically named cookies and take care
395 of them individually is harder than just to blindly and unconditionally sort
396 all cookies based on their path lengths. All major browsers also already do
397 this, so this makes our behavior one step closer to them in the cookie area.
399 Test case 8 was the only one that broke due to this change and I updated it
402 Daniel Stenberg (19 Jan 2010)
403 - David McCreedy brought a fix and a new test case (129) to make libcurl work
404 again when downloading files over FTP using ASCII and it turns out that the
405 final size of the file is not the same as the initial size the server
406 reported. This is very common since servers don't take the newline
407 conversions into account.
409 Kamil Dudka (14 Jan 2010)
410 - Suppressed side effect of OpenSSL configure checks, which prevented NSS from
411 being properly detected under certain circumstances. It had been caused by
412 strange behavior of pkg-config when handling PKG_CONFIG_LIBDIR. pkg-config
413 distinguishes among empty and non-existent environment variable in that case.
415 Daniel Stenberg (12 Jan 2010)
416 - Gil Weber reported a peculiar flaw with the multi interface when doing SFTP
417 transfers: curl_multi_fdset() would return -1 and not set and file
418 descriptors several times during a transfer of a single file. It turned out
419 to be due to two different flaws now fixed. Gil's excellent recipe helped me
422 Daniel Stenberg (11 Jan 2010)
423 - Made sure that the progress callback is repeatedly called at a regular
424 interval even during very slow connects.
426 - The tests/runtests.pl script now checks to see if the test case that runs is
427 present in the tests/data/Makefile.am and outputs a notice message on the
428 screen if not. Each test file has to be included in that Makefile.am to get
429 included in release archives and forgetting to add files there is a common
430 mistake. This is an attempt to make it harder to forget.
432 Daniel Stenberg (9 Jan 2010)
433 - Johan van Selst found and fixed a OpenSSL session ref count leak:
435 ossl_connect_step3() increments an SSL session handle reference counter on
436 each call. When sessions are re-used this reference counter may be
437 incremented many times, but it will be decremented only once when done (by
438 Curl_ossl_session_free()); and the internal OpenSSL data will not be freed
439 if this reference count remains positive. When a session is re-used the
440 reference counter should be corrected by explicitly calling
441 SSL_SESSION_free() after each consecutive SSL_get1_session() to avoid
442 introducing a memory leak.
444 (http://curl.haxx.se/bug/view.cgi?id=2926284)
446 Daniel Stenberg (7 Jan 2010)
447 - Make sure the progress callback is called repeatedly even during very slow
448 name resolves when c-ares is used for resolving.
450 Claes Jakobsson (6 Jan 2010)
451 - Julien Chaffraix fixed so that the fragment part in an URL is not sent
452 to the server anymore.
454 Kamil Dudka (3 Jan 2010)
455 - Julien Chaffraix eliminated a duplicated initialization in singlesocket().
457 Daniel Stenberg (2 Jan 2010)
458 - Make curl support --ssl and --ssl-reqd instead of the previous FTP-specific
459 versions --ftp-ssl and --ftp-ssl-reqd as these options are now used to
460 control SSL/TLS for IMAP, POP3 and SMTP as well in addition to FTP. The old
461 option names are still working but the new ones are the ones listed and
464 Daniel Stenberg (1 Jan 2010)
465 - Ingmar Runge enhanced libcurl's FTP engine to support the PRET command. This
466 command is a special "hack" used by the drftpd server, but even though it is
467 a custom extension I've deemed it fine to add to libcurl since this server
468 seems to survive and people keep using it and want libcurl to support
469 it. The new libcurl option is named CURLOPT_FTP_USE_PRET, and it is also
470 usable from the curl tool with --ftp-pret. Using this option on a server
471 that doesn't support this command will make libcurl fail.
473 I added test cases 1107 and 1108 to verify the functionality.
475 The PRET command is documented at
476 http://www.drftpd.org/index.php/Distributed_PASV
478 Yang Tse (30 Dec 2009)
479 - Steven M. Schweda improved VMS build system, and Craig A. Berry helped
480 with the patch and testing.
482 Daniel Stenberg (26 Dec 2009)
483 - Renato Botelho and Peter Pentchev brought a patch that makes the libcurl
484 headers work correctly even on FreeBSD systems before v8.
486 (http://curl.haxx.se/bug/view.cgi?id=2916915)
488 Daniel Stenberg (17 Dec 2009)
489 - David Byron fixed Curl_ossl_cleanup to actually call ENGINE_cleanup when
492 - Follow-up fix for the proxy fix I did for Jon Nelson's bug. It turned out I
493 was a bit too quick and broke test case 1101 with that change. The order of
494 some of the setups is sensitive. I now changed it slightly again to make
495 sure we do them in this order:
497 1 - parse URL and figure out what protocol is used in the URL
498 2 - prepend protocol:// to URL if missing
499 3 - parse name+password off URL, which needs to know what protocol is used
500 (since only some allows for name+password in the URL)
501 4 - figure out if a proxy should be used set by an option
502 5 - if no proxy option, check proxy environment variables
503 6 - run the protocol-specific setup function, which needs to have the proxy
506 Daniel Stenberg (15 Dec 2009)
507 - Jon Nelson found a regression that turned out to be a flaw in how libcurl
508 detects and uses proxies based on the environment variables. If the proxy
509 was given as an explicit option it worked, but due to the setup order
510 mistake proxies would not be used fine for a few protocols when picked up
511 from '[protocol]_proxy'. Obviously this broke after 7.19.4. I now also added
512 test case 1106 that verifies this functionality.
514 (http://curl.haxx.se/bug/view.cgi?id=2913886)
516 Daniel Stenberg (12 Dec 2009)
517 - IMAP, POP3 and SMTP support and their TLS versions (including IMAPS, POP3S
518 and SMTPS) are now supported. The current state may not yet be solid, but
519 the foundation is in place and the test suite has some initial support for
520 these protocols. Work will now persue to make them nice libcurl citizens
523 The work with supporting these new protocols was sponsored by
524 networking4all.com - thanks!
526 Daniel Stenberg (10 Dec 2009)
527 - Siegfried Gyuricsko found out that the curl manual said --retry would retry
528 on FTP errors in the transient 5xx range. Transient FTP errors are in the
529 4xx range. The code itself only tried on 5xx errors that occured _at login_.
530 Now the retry code retries on all FTP transfer failures that ended with a
533 (http://curl.haxx.se/bug/view.cgi?id=2911279)
535 - Constantine Sapuntzakis figured out a case which would lead to libcurl
536 accessing alredy freed memory and thus crash when using HTTPS (with
537 OpenSSL), multi interface and the CURLOPT_DEBUGFUNCTION and a certain order
538 of cleaning things up. I fixed it.
540 (http://curl.haxx.se/bug/view.cgi?id=2905220)
542 Daniel Stenberg (7 Dec 2009)
543 - Martin Storsjo made libcurl use the Expect: 100-continue header for posts
544 with unknown size. Previously it was only used for posts with a known size
545 larger than 1024 bytes.
547 Daniel Stenberg (1 Dec 2009)
548 - If the Expect: 100-continue header has been set by the application through
549 curl_easy_setopt with CURLOPT_HTTPHEADER, the library should set
550 data->state.expect100header accordingly - the current code (in 7.19.7 at
551 least) doesn't handle this properly. Martin Storsjo provided the fix!
553 Yang Tse (28 Nov 2009)
554 - Added Diffie-Hellman parameters to several test harness certificate files in
555 PEM format. Required by several stunnel versions used by our test harness.
557 Daniel Stenberg (28 Nov 2009)
558 - Markus Koetter provided a polished and updated version of Chad Monroe's TFTP
559 rework patch that now integrates TFTP properly into libcurl so that it can
560 be used non-blocking with the multi interface and more. BLKSIZE also works.
562 The --tftp-blksize option was added to allow setting the TFTP BLKSIZE from
565 Daniel Stenberg (26 Nov 2009)
566 - Extended and fixed the change I did on Dec 11 for the the progress
567 meter/callback during FTP command/response sequences. It turned out it was
568 really lame before and now the progress meter SHOULD get called at least
571 Daniel Stenberg (23 Nov 2009)
572 - Bjorn Augustsson reported a bug which made curl not report any problems even
573 though it failed to write a very small download to disk (done in a single
574 fwrite call). It turned out to be because fwrite() returned success, but
575 there was insufficient error-checking for the fclose() call which tricked
576 curl to believe things were fine.
578 Yang Tse (23 Nov 2009)
579 - David Byron modified Makefile.dist vc8 and vc9 targets in order to allow
580 finer granularity control when generating src and lib makefiles.
582 Yang Tse (22 Nov 2009)
583 - I modified configure to force removal of the curlbuild.h file included in
584 distribution tarballs for use by non-configure systems. As intended, this
585 would get overwriten when doing in-tree builds. But VPATH builds would end
586 having two curlbuild.h files, one in the source tree and another in the
587 build tree. With the modification I introduced 5 Nov 2009 this could become
588 an issue when running libcurl's test suite.
590 Daniel Stenberg (20 Nov 2009)
591 - Constantine Sapuntzakis identified a write after close, as the sockets were
592 closed by libcurl before the SSL lib were shutdown and they may write to its
593 socket. Detected to at least happen with OpenSSL builds.
595 - Jad Chamcham pointed out a bug with connection re-use. If a connection had
596 CURLOPT_HTTPPROXYTUNNEL enabled over a proxy, a subsequent request using the
597 same proxy with the tunnel option disabled would still wrongly re-use that
598 previous connection and the outcome would only be badness.
600 Yang Tse (18 Nov 2009)
601 - I modified the memory tracking system to make it intolerant with zero sized
602 malloc(), calloc() and realloc() function calls.
604 Daniel Stenberg (17 Nov 2009)
605 - Constantine Sapuntzakis provided another fix for the DNS cache that could
606 end up with entries that wouldn't time-out:
608 1. Set up a first web server that redirects (307) to a http://server:port
610 2. Have curl connect to the first web server using curl multi
612 After the curl_easy_cleanup call, there will be curl dns entries hanging
613 around with in_use != 0.
615 (http://curl.haxx.se/bug/view.cgi?id=2891591)
617 - Marc Kleine-Budde fixed: curl saved the LDFLAGS set during configure into
618 its pkg-config file. So -Wl stuff ended up in the .pc file, which is really
619 bad, and breaks if there are multiple -Wl in our LDFLAGS (which are in
620 PTXdist). bug #2893592 (http://curl.haxx.se/bug/view.cgi?id=2893592)
622 Kamil Dudka (15 Nov 2009)
623 - David Byron improved the configure script to use pkg-config to find OpenSSL
624 (and in particular the list of required libraries) even if a path is given
625 as argument to --with-ssl
627 Yang Tse (15 Nov 2009)
628 - I removed enable-thread / disable-thread configure option. These were only
629 placebo options. The library is always built as thread safe as possible on
632 Claes Jakobsson (14 Nov 2009)
633 - curl-config now accepts '--configure' to see what arguments was
634 passed to the configure script when building curl.
636 Daniel Stenberg (14 Nov 2009)
637 - Claes Jakobsson restored the configure functionality to detect NSS when
638 --with-nss is set but not "yes".
640 I think we can still improve that to check for pkg-config in that path etc,
641 but at least this patch brings back the same functionality we had before.
643 - Camille Moncelier added support for the file type SSL_FILETYPE_ENGINE for
644 the client certificate. It also disable the key name test as some engines
645 can select a private key/cert automatically (When there is only one key
646 and/or certificate on the hardware device used by the engine)
648 Yang Tse (14 Nov 2009)
649 - Constantine Sapuntzakis provided the fix that ensures that an SSL connection
650 won't be reused unless protection level for peer and host verification match.
652 I refactored how preprocessor symbol _THREAD_SAFE definition is done.
654 Kamil Dudka (12 Nov 2009)
655 - Kevin Baughman provided a fix preventing libcurl-NSS from crash on doubly
656 closed NSPR descriptor. The issue was hard to find, reported several times
657 before and always closed unresolved. More info at the RH bug:
658 https://bugzilla.redhat.com/534176
660 - libcurl-NSS now tries to reconnect with TLS disabled in case it detects
661 a broken TLS server. However it does not happen if SSL version is selected
662 manually. The approach was originally taken from PSM. Kaspar Brand helped me
663 to complete the patch. Original bug reports:
664 https://bugzilla.redhat.com/525496
665 https://bugzilla.redhat.com/527771
667 Yang Tse (12 Nov 2009)
668 - I modified configure script to make the getaddrinfo function check also
669 verify if the function is thread safe.
671 Yang Tse (11 Nov 2009)
672 - Marco Maggi reported that compilation failed when configured --with-gssapi
673 and GNU GSS installed due to a missing mutual exclusion of header files in
674 the Kerberos 5 code path. He also verified that my patch worked for him.
676 Daniel Stenberg (11 Nov 2009)
677 - Constantine Sapuntzakis posted bug #2891595
678 (http://curl.haxx.se/bug/view.cgi?id=2891595) which identified how an entry
679 in the DNS cache would linger too long if the request that added it was in
680 use that long. He also provided the patch that now makes libcurl capable of
681 still doing a request while the DNS hash entry may get timed out.
683 - Christian Schmitz noticed that the progress meter/callback was not properly
684 used during the FTP connection phase (after the actual TCP connect), while
685 it of course should be. I also made the speed check get called correctly so
686 that really slow servers will trigger that properly too.
688 Kamil Dudka (5 Nov 2009)
689 - Dropped misleading timeouts in libcurl-NSS and made sure the SSL socket works
690 in non-blocking mode.
692 Yang Tse (5 Nov 2009)
693 - I removed leading 'curl' path on the 'curlbuild.h' include statement in
694 curl.h, adjusting auto-makefiles include path, to enhance portability to
695 OS's without an orthogonal directory tree structure such as OS/400.
697 Daniel Stenberg (4 Nov 2009)
698 - I fixed several problems with the transfer progress meter. It showed the
699 wrong percentage for small files, most notable for <1000 bytes and could
700 easily end up showing more than 100% at the end. It also didn't show any
701 percentage, transfer size or estimated transfer times when transferring
704 Version 7.19.7 (4 November 2009)
706 Daniel Stenberg (2 Nov 2009)
707 - As reported independent by both Stan van de Burgt and Didier Brisebourg,
708 CURLINFO_SIZE_DOWNLOAD (the -w variable size_download) didn't work when
709 getting data from ldap!
711 Daniel Stenberg (31 Oct 2009)
712 - Gabriel Kuri reported a problem with CURLINFO_CONTENT_LENGTH_DOWNLOAD if the
713 download was 0 bytes, as libcurl would then return the size as unknown (-1)
714 and not 0. I wrote a fix and test case 566 to verify it.
716 Daniel Stenberg (30 Oct 2009)
717 - Liza Alenchery mentioned a problem with re-used SCP connection when a bad
718 auth is used, as it caused a crash. I failed to repeat the issue, but still
719 made a change that now forces the TCP connection used for a freed SCP
720 session to get closed and not be re-used.
722 - "Tom" posted a bug report that mentioned how libcurl did wrong when doing a
723 POST using a read callback, with Digest authentication and
724 "Transfer-Encoding: chunked" enforced. I would then cause the first request
725 to be wrongly sent and then basically hang until the server closed the
726 connection. I fixed the problem and added test case 565 to verify it.
728 Daniel Stenberg (25 Oct 2009)
729 - Dima Barsky made the curl cookie parser accept cookies even with blank or
730 unparsable expiry dates and then treat them as session cookies - previously
731 libcurl would reject cookies with a date format it couldn't parse. Research
732 shows that the major browser treat such cookies as session cookies. I
733 modified test 8 and 31 to verify this.
735 Daniel Stenberg (21 Oct 2009)
736 - Attempt to use pkg-config for finding out libssh2 installation details
739 - A patch in bug report #2883177 (http://curl.haxx.se/bug/view.cgi?id=2883177)
740 by Johan van Selst introduced the --crlfile option to curl, which makes curl
741 tell libcurl about a file with CRL (certificate revocation list) data to
744 Daniel Stenberg (18 Oct 2009)
745 - Ray Dassen provided a patch in Debian's bug tracker (bug number #551461)
746 that now makes curl_getdate(3) actually handles RFC 822 formatted dates that
747 use the "single letter military timezones".
748 http://www.rfc-ref.org/RFC-TEXTS/822/chapter5.html has the details.
750 - Fixed memory leak in the SCP/SFTP code as it never freed the knownhosts
753 - John Dennis filed bug report #2873666
754 (http://curl.haxx.se/bug/view.cgi?id=2873666) which identified a problem
755 which made libcurl loop infinitely when given incorrect credentials when
756 using HTTP GSS negotiate authentication. He also provided a small and simple
759 - Kevin Baughman found a double close() problem with libcurl-NSS, as when
760 libcurl called NSS to close the SSL "session" it also closed the actual
763 Yang Tse (17 Oct 2009)
764 - Bug report #2866724 indicated
765 (http://curl.haxx.se/bug/view.cgi?id=2866724) that curl on Windows failed
766 when writing files whose file names originally contained characters which
767 are not valid for file names on Windows. Dan Fandrich provided an initial
768 patch and another revised one to fix this issue.
770 Daniel Stenberg (1 Oct 2009)
771 - Tom Mueller correctly reported in bug report #2870221
772 (http://curl.haxx.se/bug/view.cgi?id=2870221) that libcurl returned an
773 incorrect return code from the internal trynextip() function which caused
774 him grief. This is a regression that was introduced in 7.19.1 and I find it
775 strange it hasn't hit us harder, but I won't persue into figuring out
778 - Constantine Sapuntzakis: The current implementation will always set
779 SO_SNDBUF to CURL_WRITE_SIZE even if the SO_SNDBUF starts out larger. The
780 patch doesn't do a setsockopt if SO_SNDBUF is already greater than
781 CURL_WRITE_SIZE. This should help folks who have set up their computer with
784 Daniel Stenberg (27 Sep 2009)
785 - I introduced a maximum limit for received HTTP headers. It is controlled by
786 the define CURL_MAX_HTTP_HEADER which is even exposed in the public header
787 file to allow for users to fairly easy rebuild libcurl with a modified
788 limit. The rationale for a fixed limit is that libcurl is realloc()ing a
789 buffer to be able to put a full header into it, so that it can call the
790 header callback with the entire header, but that also risk getting it into
791 trouble if a server by mistake or willingly sends a header that is more or
792 less without an end. The limit is set to 100K.
794 Daniel Stenberg (26 Sep 2009)
795 - John P. McCaskey posted a bug report that showed how libcurl did wrong when
796 saving received cookies with no given path, if the path in the request had a
797 query part. That is means a question mark (?) and characters on the right
798 side of that. I wrote test case 1105 and fixed this problem.
800 Kamil Dudka (26 Sep 2009)
801 - Implemented a protocol independent way to specify blocking direction, used by
802 transfer.c for blocking. It is currently used only by SCP and SFTP protocols.
803 This enhancement resolves an issue with 100% CPU usage during SFTP upload,
806 Daniel Stenberg (25 Sep 2009)
807 - Chris Mumford filed bug report #2861587
808 (http://curl.haxx.se/bug/view.cgi?id=2861587) identifying that libcurl used
809 the OpenSSL function X509_load_crl_file() wrongly and failed if it would
810 load a CRL file with more than one certificate within. This is now fixed.
812 Daniel Stenberg (16 Sep 2009)
813 - Sven Anders reported that we introduced a cert verfication flaw for OpenSSL-
814 powered libcurl in 7.19.6. If there was a X509v3 Subject Alternative Name
815 field in the certficate it had to match and so even if non-DNS and non-IP
816 entry was present it caused the verification to fail.
818 Daniel Fandrich (15 Sep 2009)
819 - Moved the libssh2 checks after the SSL library checks. This helps when
820 statically linking since libssh2 needs the SSL library link flags to be
821 set up already to satisfy its dependencies. This wouldn't be necessary if
822 the libssh2 configure check was changed to use pkg-config since the
823 --static flag would add the dependencies automatically.
825 Yang Tse (14 Sep 2009)
826 - Revert Joshua Kwan's patch committed 11 Sep 2009.
828 Some systems poll function sets POLLHUP in revents without setting
829 POLLIN, and sets POLLERR without setting POLLIN and POLLOUT. In some
830 libcurl code execution paths this could trigger busy wait loops with
831 high CPU usage until a timeout condition aborted the loop.
833 The reverted patch addressed the above issue for a very specific case,
834 when awaiting c-ares to resolve. A libcurl-wide fix for Curl_poll now
837 Guenter Knauf (11 Sep 2009)
838 - Joshua Kwan provided a patch to pass POLLERR / POLLHUP back to c-ares.
839 This fixes a loop problem with high CPU usage.
841 Daniel Stenberg (10 Sep 2009)
842 - Claes Jakobsson fixed a problem with cookie expiry dates at exctly the epoch
843 start second "Thu Jan 1 00:00:00 GMT 1970" as the date parser then returns 0
844 which internally then is treated as a session cookie. That particular date
845 is now made to get the value of 1.
847 Daniel Stenberg (2 Sep 2009)
848 - Daniel Johnson found a flaw in the code converting sftp-errors to libcurl
851 Daniel Stenberg (1 Sep 2009)
852 - Peter Sylvester made a debug feature for Curl_resolv() that now will force
853 libcurl to resolve 'localhost' whatever name you use in the URL *if* you set
854 the --interface option to (exactly) "LocalHost". This will enable us to
855 write tests for custom hosts names but still use a local host server.
857 - configure now tries to use pkg-config for a number of sub-dependencies even
858 when cross-compiling. The key to success is then you properly setup
859 PKG_CONFIG_PATH before invoking configure.
861 I also improved how NSS is detected by trying nss-config if pkg-config isn't
862 present, and as a last resort just use the lib name and force the user to
863 setup the LIBS/LDFLAGS/CFLAGS etc properly. The previous last resort would
864 add a range of various libs that would almost never be quite correct.
866 Daniel Stenberg (31 Aug 2009)
867 - When using the multi interface with FTP and you asked for NOBODY, you did no
868 QUOTE commands and the request used the same path as the connection had
869 already changed to, it would decide that no commands would be necessary for
870 the "DO" action and that was not handled properly but libcurl would instead
873 Kamil Dudka (28 Aug 2009)
874 - Improved error message for not matching certificate subject name in
875 libcurl-NSS. Originally reported at:
876 https://bugzilla.redhat.com/show_bug.cgi?id=516056#c9
878 Patrick Monnerat (24 Aug 2009)
879 - Introduced a SYST-based test to properly set-up name format when dealing
880 with the OS/400 FTP server.
882 - Fixed an ftp_readresp() bug preventing detection of failing control socket
883 and causing FTP client to loop forever.
885 Daniel Stenberg (24 Aug 2009)
886 - Marc de Bruin pointed out that configure --with-gnutls=PATH didn't work
887 properly and provided a fix. http://curl.haxx.se/bug/view.cgi?id=2843008
889 - Eric Wong introduced support for the new option -T. (dot) that makes curl
890 read stdin in a non-blocking fashion. This also brings back -T- (minus) to
891 the previous blocking behavior since it could break stuff for people at
894 Michal Marek (21 Aug 2009)
895 - With CURLOPT_PROXY_TRANSFER_MODE, avoid sending invalid URLs like
896 ftp://example.com;type=i if the user specified ftp://example.com without the
899 Daniel Stenberg (21 Aug 2009)
900 - Andre Guibert de Bruet pointed out a missing return code check for a
901 strdup() that could lead to segfault if it returned NULL. I extended his
902 suggest patch to now have Curl_retry_request() return a regular return code
903 and better check that.
905 - Lots of good work by Krister Johansen, mostly related to pipelining:
907 Fix SIGSEGV on free'd easy_conn when pipe unexpectedly breaks
908 Fix data corruption issue with re-connected transfers
909 Fix use after free if we're completed but easy_conn not NULL
911 Kamil Dudka (13 Aug 2009)
912 - Changed NSS code to not ignore the value of ssl.verifyhost and produce more
913 verbose error messages. Originally reported at:
914 https://bugzilla.redhat.com/show_bug.cgi?id=516056
916 Daniel Stenberg (12 Aug 2009)
917 - Karl Moerder fixed the Makefile.vc* makefiles to include the new file
918 nonblock.c so that they work fine again
920 - I expanded test 517 with a bunch of more dates that originate from the
921 Chrome browser test suite. It turns out most of them get parsed the same
924 Version 7.19.6 (12 August 2009)
926 Daniel Stenberg (12 Aug 2009)
927 - Carsten Lange reported a bug and provided a patch for TFTP upload and the
928 sending of the TSIZE option. I don't like fixing bugs just hours before
929 a release, but since it was broken and the patch fixes this for him I decided
932 Daniel Stenberg (11 Aug 2009)
933 - Peter Sylvester made the HTTPS test server use specific certificates for
934 each test, so that the test suite can now be used to actually test the
935 verification of cert names etc. This made an error show up in the OpenSSL-
936 specific code where it would attempt to match the CN field even if a
937 subjectAltName exists that doesn't match. This is now fixed and verified
940 - Benbuck Nason posted the bug report #2835196
941 (http://curl.haxx.se/bug/view.cgi?id=2835196), fixing a few compiler
942 warnings when mixing ints and bools.
944 Daniel Fandrich (10 Aug 2009)
945 - Fixed a memory leak in the FTP code and an off-by-one heap buffer overflow.
947 Daniel Fandrich (9 Aug 2009)
948 - Fixed some memory leaks in the command-line tool that caused most of the
949 torture tests to fail.
951 Daniel Stenberg (2 Aug 2009)
952 - Curt Bogmine reported a problem with SNI enabled on a particular server. We
953 should introduce an option to disable SNI, but as we're in feature freeze
954 now I've addressed the obvious bug here (pointed out by Peter Sylvester): we
955 shouldn't try to enable SNI when SSLv2 or SSLv3 is explicitly selected.
956 Code for OpenSSL and GnuTLS was fixed. NSS doesn't seem to have a particular
957 option for SNI, or are we simply not using it?
959 Daniel Stenberg (1 Aug 2009)
960 - Scott Cantor posted the bug report #2829955
961 (http://curl.haxx.se/bug/view.cgi?id=2829955) mentioning the recent SSL cert
962 verification flaw found and exploited by Moxie Marlinspike. The presentation
963 he did at Black Hat is available here:
964 https://www.blackhat.com/html/bh-usa-09/bh-usa-09-archives.html#Marlinspike
966 Apparently at least one CA allowed a subjectAltName or CN that contain a
967 zero byte, and thus clients that assumed they would never have zero bytes
968 were exploited to OK a certificate that didn't actually match the site. Like
969 if the name in the cert was "example.com\0theatualsite.com", libcurl would
970 happily verify that cert for example.com.
972 libcurl now better uses the length of the extracted name, not using the zero
973 termination for getting the string length.
975 This fixing only made and needed in OpenSSL interfacing code.
977 - Tanguy Fautre pointed out that OpenSSL's function RAND_screen() (present
978 only in some OpenSSL installs - like on Windows) isn't thread-safe and we
979 agreed that moving it to the global_init() function is a decent way to deal
982 - Alexander Beedie provided the patch for a noproxy problem: If I have set
983 CURLOPT_NOPROXY to "*", or to a host that should not use a proxy, I actually
984 could still end up using a proxy if a proxy environment variable was set.
986 Daniel Stenberg (27 Jul 2009)
987 - All the quote options (CURLOPT_QUOTE, CURLOPT_POSTQUOTE and
988 CURLOPT_PREQUOTE) now accept a preceeding asterisk before the command to
989 send when using FTP, as a sign that libcurl shall simply ignore the response
990 from the server instead of treating it as an error. Not treating a 400+ FTP
991 response code as an error means that failed commands will not abort the
992 chain of commands, nor will they cause the connection to get disconnected.
994 Daniel Stenberg (26 Jul 2009)
995 - Johan van Selst posted bug report #2825989
996 (http://curl.haxx.se/bug/view.cgi?id=2825989) pointing out that
997 OpenSSL-powered libcurl didn't support the SHA-2 digest algorithm, and
998 provided the solution too: to use OpenSSL_add_all_algorithms() in addition
999 to the older SSLeay_* alternative. OpenSSL_add_all_algorithms was added in
1002 Daniel Stenberg (23 Jul 2009)
1003 - Added CURLOPT_SSH_KNOWNHOSTS, CURLOPT_SSH_KEYFUNCTION, CURLOPT_SSH_KEYDATA.
1004 They introduce known_host support for SSH keys to libcurl. See docs for
1005 details. Note that this feature depends on a new enough libssh2 version, to
1006 be supported in libssh2 1.2 and later (or current git repo at this time).
1008 Michal Marek (22 Jul 2009)
1009 - David Binderman found a memory and fd leak in lib/gtls.c:load_file()
1010 (https://bugzilla.novell.com/523919). When looking at the code, I found that
1011 also the ptr pointer can leak.
1013 Kamil Dudka (20 Jul 2009)
1014 - Claes Jakobsson improved the support for client certificates handling in
1015 NSS-powered libcurl. Now the client certificates can be selected
1016 automatically by a NSS built-in hook. Additionally pre-login to all PKCS11
1017 slots is no more performed. It used to cause problems with HW tokens.
1019 - Fixed reference counting for NSS client certificates. Now the PEM reader
1020 module should be always properly unloaded on Curl_nss_cleanup(). If the
1021 unload fails though, libcurl will try to reuse the already loaded instance.
1023 Daniel Fandrich (15 Jul 2009)
1024 - Added nonblock.c to the non-automake makefiles (note that the dependencies
1025 in the Watcom makefiles aren't quite correct).
1027 Michal Marek (15 Jul 2009)
1028 - Changed the description of CURLINFO_OS_ERRNO to make it clear that the
1029 errno is not reset on success.
1031 Guenter Knauf (14 Jul 2009)
1032 - renamed generated config.h to curl_config.h to avoid any future clashes
1033 with config.h from other projects.
1035 Daniel Stenberg (9 Jul 2009)
1036 - Eric Wong introduced curlx_nonblock() that the curl tool now (re-)uses for
1037 setting a file descriptor non-blocking. Used by the functionality Eric
1038 himself brough on June 15th.
1040 Daniel Stenberg (8 Jul 2009)
1041 - Constantine Sapuntzakis posted bug report #2813123
1042 (http://curl.haxx.se/bug/view.cgi?id=2813123) and an a patch that fixes the
1045 Url A is accessed using auth. Url A redirects to Url B (on a different
1046 server0. Url B reuses a persistent connection. Url B has auth, even though
1047 it's on a different server.
1049 Note: if Url B does not reuse a persistent connection, auth is not sent.
1053 data->state.first_host is not initialized becuase Curl_http_connect is not
1054 called when a connection is reused.
1058 move initialization of data->state.first_host to Curl_http. No code before
1059 Curl_http uses data->state.first_host anyway.
1061 Guenter Knauf (4 Jul 2009)
1062 - Markus Koetter provided a patch to avoid getnameinfo() usage which broke a
1063 couple of both IPv4 and IPv6 autobuilds.
1065 Daniel Stenberg (29 Jun 2009)
1066 - Markus Koetter made CURLOPT_FTPPORT (and curl's -P/--ftpport) support a port
1067 range if given colon-separated after the host name/address part. Like
1068 "192.168.0.1:2000-10000"
1070 - Modified the separators used for CURLOPT_CERTINFO in multi-part outputs. I
1071 don't know how they got wrong in the first place, but using this output
1072 format makes it possible to quite easily separate the string into an array
1075 Daniel Fandrich (16 June 2009)
1076 - Added a few more compiler warning options for gcc.
1078 Daniel Stenberg (16 Jun 2009)
1079 - Reuven Wachtfogel made curl -o - properly produce a binary output on windows
1080 (no newline translations). Use -B/--use-ascii if you rather get the ascii
1083 Michal Marek (16 Jun 2009)
1084 - When doing non-anonymous ftp via http proxies and the password is not
1085 provided in the url, add it there (squid needs this).
1087 Daniel Stenberg (15 Jun 2009)
1088 - Eric Wong's patch:
1090 This allows curl(1) to be used as a client-side tunnel for arbitrary stream
1091 protocols by abusing chunked transfer encoding in both the HTTP request and
1092 HTTP response. This requires server support for sending a response while a
1093 request is still being read, of course.
1095 If attempting to read from stdin returns EAGAIN, then we pause our sender.
1096 This leaves curl to attempt to read from the socket while reading from stdin
1097 (and thus sending) is paused.
1099 This change was needed to allow successfully tunneling the git protocol over
1100 HTTP (--no-buffer is needed, as well).
1102 Patrick Monnerat (15 Jun 2009)
1103 - Replaced use of standard C library rand()/srand() by our own pseudo-random
1106 Yang Tse (11 Jun 2009)
1107 - I adapted testcurl script to allow building test harness programs when
1108 cross-compiling for a *-*-mingw* host.
1110 Daniel Stenberg (10 Jun 2009)
1111 - Fabian Keil ran clang on the (lib)curl code, found a bunch of warnings and
1112 contributed a range of patches to fix them.
1114 Yang Tse (10 Jun 2009)
1115 - I introduced configure script option --enable-curldebug which now allows
1116 the decoupled enabling or disabling of the curl debug memory tracking
1117 feature from the --enable-debug option which no longer controls this.
1119 curl --version will list 'Debug' feature for debug enabled builds, and
1120 will list 'TrackMemory' feature for curl debug memory tracking capable
1121 builds. These features are independent and can be controlled when running
1122 the configure script. When --enable-debug is given both features will be
1123 enabled, unless some restriction prevents memory tracking from being used.
1125 Internally, definition of preprocessor symbol DEBUGBUILD restricts code
1126 which is only compiled for debug enabled builds. And symbol CURLDEBUG is
1127 used to differentiate code which is _only_ used for memory tracking.
1129 Yang Tse (9 Jun 2009)
1130 - Daniel Steinberg pointed out that Curl_FormInit() in formdata.c was not
1131 initializing the fread callback pointer and this triggered a compiler
1132 warning, also provided a friendly suggestion on how to fix it.
1134 Daniel Stenberg (8 Jun 2009)
1135 - Claes Jakobsson provided a patch for libcurl-NSS that fixed a bad refcount
1136 issue with client certs that caused issues like segfaults.
1137 http://curl.haxx.se/mail/lib-2009-05/0316.html
1139 - Triggered by bug report #2798852 and the patch in there, I fixed configure
1140 to detect gnutls build options with pkg-config only and not libgnutls-config
1141 anymore since GnuTLS has stopped distributing that tool. If an explicit path
1142 is given to configure, we will instead guess on how to link and use that
1143 lib. I did not use the patch from the bug report.
1145 Yang Tse (8 Jun 2009)
1146 - Igor Novoseltsev adjusted Makefile.vxworks to get sources and headers
1147 included from Makefile.inc, and provided docs\INSTALL VxWorks section.
1149 - I removed buildconf.bat from release and daily snapshot archives. This
1150 file is only for CVS tree checkout builds.
1152 Daniel Stenberg (8 Jun 2009)
1153 - Eric Wong fixed --no-buffer to actually switch off output buffering. Been
1156 Bill Hoffman (6 Jun 2009)
1157 - Added some cmake docs and fixed socklen_t in the build.
1159 Yang Tse (5 Jun 2009)
1160 - John E. Malmberg provided VMS specific patch: "This fixes an existing bug
1161 in urlglob.c where it was not converting the Curl Unix exit code to a VMS
1162 DCL compatible exit code. This fix required the enhancement described next.
1163 This also adds an enhancement to main.c so that when curl is run under a
1164 Unix shell like Bash on VMS, it will return the standard Unix exit codes
1165 and messages." And another patch for docs/examples.
1167 I introduced os-specific.c and os-specific.h for use in curl tool code
1168 and adjusted John E. Malmberg's patch placement to use these new files
1169 as an effort to prevent main.c from growing ad infinitum. Code already
1170 existing in main.c which is OS specific should be moved into these files.
1172 Daniel Stenberg (4 June 2009)
1173 - Setting the Content-Length: header from your app when you do a POST or PUT
1174 is almost always a VERY BAD IDEA. Yet there are still apps out there doing
1175 this, and now recently it triggered a bug/side-effect in libcurl as when
1176 libcurl sends a POST or PUT with NTLM, it sends an empty post first when it
1177 knows it will just get a 401/407 back. If the app then replaced the
1178 Content-Length header, it caused the server to wait for input that libcurl
1179 wouldn't send. Aaron Oneal reported this problem in bug report #2799008
1180 (http://curl.haxx.se/bug/view.cgi?id=2799008) and helped us verify the fix.
1182 Yang Tse (4 Jun 2009)
1183 - Igor Novoseltsev provided patches and information, that after some
1184 adjustments to better fit curl's way of doing things, have resulted
1185 in the posibility of building libcurl for VxWorks.
1187 Daniel Fandrich (2 June 2009)
1188 - Checked in a Google Android make file. To use it, you must first
1189 create a config.h file by running configure in the Android environment,
1190 which doesn't seem to be easy to do. If no easy way can be found, a
1191 static config-android.h may need to be created and checked in to the
1192 libcurl source tree.
1194 Daniel Stenberg (1 June 2009)
1195 - Claes Jakobsson fixed the configure script to better find and use NSS
1198 Yang Tse (1 Jun 2009)
1199 - John E. Malmberg provided a VMS specific clean-up for curl.h, and pointed
1200 out that the configure script was failing to detect the timeval struct on
1201 VMS when building with _XOPEN_SOURCE_EXTENDED undefined due to definition
1202 taking place in socket.h instead of time.h. I have adjusted configure
1203 script to also include this header when checking struct timeval.
1205 Daniel Stenberg (27 May 2009)
1206 - Frank McGeough provided a small OpenSSL #include fix to make libcurl compile
1207 fine with Nokia 5th edition 1.0 SDK for Symbian.
1209 - Andre Guibert de Bruet found a call to a OpenSSL function that didn't check
1210 for a failure properly.
1212 - Mike Crowe pointed out that setting CURLOPT_USERPWD to NULL used to clear
1213 the auth credentials back in 7.19.0 and earlier while now you have to set ""
1214 to get the same effect. His patch brings back the ability to use NULL.
1216 - Claes Jakobsson fixed libcurl-NSS to build fine even without the
1217 PK11_CreateGenericObject() function.
1219 Daniel Stenberg (25 May 2009)
1220 - bug report #2796358 (http://curl.haxx.se/bug/view.cgi?id=2796358) pointed
1221 out that the cookie parser would leak memory when it parses cookies that are
1222 received with domain, path etc set multiple times in the same header. While
1223 such a cookie is questionable, they occur in the wild and libcurl no longer
1224 leaks memory for them. I added such a header to test case 8.
1226 Daniel Fandrich (22 May 2009)
1227 - Removed some obsolete digest code that caused a valgrind error in test 551.
1229 Daniel Fandrich (20 May 2009)
1230 - Added "non-existing host" test keywords to make it easy to skip those
1231 tests on machines that have broken DNS configurations (such as
1232 those configured to use OpenDNS).
1234 Daniel Stenberg (19 May 2009)
1235 - Kamil Dudka brought the patch from the Redhat bug entry
1236 https://bugzilla.redhat.com/show_bug.cgi?id=427966 which was libcurl closing
1237 a bad file descriptor when closing down the FTP data connection. Caolan
1238 McNamara seems to be the original author of it.
1240 Version 7.19.5 (18 May 2009)
1242 Daniel Stenberg (17 May 2009)
1243 - James Bursa posted a patch to the mailing list that fixed a problem with
1244 no_proxy which made it not skip the proxy if the URL entered contained a
1245 user name. I added test case 1101 to verify.
1247 Daniel Stenberg (11 May 2009)
1248 - Balint Szilakszi reported a memory leak when libcurl did gzip decompression
1249 of streams that had some parts (legitimately) missing. We now provide and use
1250 a proper cleanup function for the content encoding submodule.
1251 http://curl.haxx.se/mail/lib-2009-05/0092.html
1253 - Kamil Dudka provided a fix for libcurl-NSS reported by Michael Cronenworth
1254 at https://bugzilla.redhat.com/show_bug.cgi?id=453612#c12
1256 If an incorrect password is given while loading a private key, libcurl ends
1257 up in an infinite loop consuming memory. The bug is critical.
1259 - I fixed the problem with doing NTLM, POST and then following a 302 redirect,
1260 as reported by Ebenezer Ikonne (on curl-users) and Laurent Rabret (on
1261 curl-library). The transfer was mistakenly marked to get more data to send
1262 but since it didn't actually have that, it just hung there...
1264 Daniel Stenberg (10 May 2009)
1265 - Andre Guibert de Bruet correctly pointed out an over-alloc with one wasted
1266 byte in the digest code.
1268 Yang Tse (9 May 2009)
1269 - Removed DOS and TPF package's subdirectory Makefile.am, it was only used
1270 to include some files in the distribution tarball serving no other purpose.
1271 Files from the DOS and TPF subdirectories are now included in the EXTRA_DIST
1272 of the Makefile in the parent subdirectory.
1274 Yang Tse (8 May 2009)
1275 - Changed host name literal in several tests to one under the haxx.se domain.
1277 - Renamed vc6 workspace and project files to avoid filename clash when used
1278 for conversion to later VS versions.
1280 Daniel Stenberg (8 May 2009)
1281 - Constantine Sapuntzakis fixed bug report #2784055
1282 (http://curl.haxx.se/bug/view.cgi?id=2784055) identifying a problem to
1283 connect to SOCKS proxies when using the multi interface. It turned out to
1284 almost not work at all previously. We need to wait for the TCP connect to
1285 be properly verified before doing the SOCKS magic.
1287 There's still a flaw in the FTP code for this.
1289 Daniel Stenberg (7 May 2009)
1290 - Made the SO_SNDBUF setting for the data connection socket for ftp uploads as
1291 well. See change 28 Apr 2009.
1293 Yang Tse (7 May 2009)
1294 - Fixed an issue affecting FTP transfers, introduced with the transfer.c
1295 patch committed May 4.
1297 Daniel Stenberg (7 May 2009)
1298 - Man page *roff problems fixed thanks to input from Colin Watson. Problems
1299 reported in the Debian package.
1301 - Vijay G filed bug report #2723236
1302 (http://curl.haxx.se/bug/view.cgi?id=2723236) identifying a problem with
1303 libcurl's TFTP code and its lack of dealing with the OACK packet.
1305 Yang Tse (5 May 2009)
1306 - Fixed the --ftp-port address of test #251 to the CLIENTIP address, and
1307 reverted the change affecting test suite harness committed 4 May.
1309 Daniel Stenberg (5 May 2009)
1310 - Inspired by Michael Smith's session id fix for OpenSSL, I did the
1311 corresponding fix in the GnuTLS code: make sure to store the new session id
1312 in case the previous re-used one is rejected.
1314 Daniel Stenberg (4 May 2009)
1315 - Michael Smith posted bug report #2786255
1316 (http://curl.haxx.se/bug/view.cgi?id=2786255) with a patch, identifying how
1317 libcurl did not deal with SSL session ids properly if the server rejected a
1318 re-use of one. Starting now, it will forget the rejected one and remember
1319 the new. This change was for OpenSSL only, it is likely that other SSL lib
1320 code needs similar fixes.
1322 Yang Tse (4 May 2009)
1323 - Applied David McCreedy's "transfer.c fixes for CURL_DO_LINEEND_CONV and
1324 non-ASCII platform HTTP requests" patch addressing two HTTP PUT problems:
1325 1) On non-ASCII platforms not all of the protocol portions of the PUT are
1326 being translated to ASCII. 2) On all platforms the line endings of part of
1327 the protocol portions are mangled from CRLF to CRCRLF if data->set.crlf or
1328 data->set.prefer_ascii are set (depending on CURL_DO_LINEEND_CONV).
1330 - Applied David McCreedy's patch to fix test suite harness to allow test FTP
1331 server and client on different machines, providing FTP client address when
1332 running the FTP test server.
1334 Daniel Fandrich (3 May 2009)
1335 - Added and disabled test case 563 which shows KNOWN_BUGS #59. The bug
1336 report failed to mention that a proxy must be used to reproduce it.
1338 Yang Tse (2 May 2009)
1339 - Use a build-time configured curl_socklen_t data type instead of socklen_t.
1341 Yang Tse (1 May 2009)
1342 - Applied David McCreedy's patches "TPF-platform specific changes to various
1343 files" and "http.c fix to Curl_proxyCONNECT for non-ASCII platforms", the
1344 former with minor edits.
1346 Daniel Stenberg (30 Apr 2009)
1347 - I was going to fix issue #59 in KNOWN_BUGS
1349 If the CURLOPT_PORT option is used on an FTP URL like
1350 "ftp://example.com/file;type=A" the ";type=A" is stripped off.
1352 I added test case 562 to verify, only to find out that I couldn't repeat
1353 this bug so I hereby consider it not a bug anymore!
1355 Daniel Stenberg (29 Apr 2009)
1356 - Based on bug report #2723219 (http://curl.haxx.se/bug/view.cgi?id=2723219)
1357 I've now made TFTP "connections" not being kept for re-use within libcurl.
1358 TFTP is UDP-based so the benefit was really low (if even existing) to begin
1359 with so instead of tracking down to fix this problem we instead removed the
1360 re-use. I also enabled test case 1099 that I wrote a few days ago to verify
1361 that this change fixes the reported problem.
1363 Daniel Stenberg (28 Apr 2009)
1364 - Constantine Sapuntzakis filed bug report #2783090
1365 (http://curl.haxx.se/bug/view.cgi?id=2783090) pointing out that on windows
1366 we need to grow the SO_SNDBUF buffer somewhat to get really good upload
1367 speeds. http://support.microsoft.com/kb/823764 has the details. Friends
1368 confirmed that simply adding 32 to CURL_MAX_WRITE_SIZE is enough.
1370 - Bug report #2709004 (http://curl.haxx.se/bug/view.cgi?id=2709004) by Tim
1371 Chen pointed out how curl couldn't upload with resume when reading from a
1374 This ended up with the introduction of a new return code for the
1375 CURLOPT_SEEKFUNCTION callback that basically says that the seek failed but
1376 that libcurl may try to resolve the situation anyway. In our case this means
1377 libcurl will attempt to instead read that much data from the stream instead
1378 of seeking and that way curl can now upload with resume when data is read
1381 Daniel Stenberg (26 Apr 2009)
1382 - Bug report #2779733 (http://curl.haxx.se/bug/view.cgi?id=2779733) by Sven
1383 Wegener pointed out that CURLINFO_APPCONNECT_TIME didn't work with the multi
1384 interface and provided a patch that fixed the problem!
1386 Daniel Stenberg (24 Apr 2009)
1387 - Kamil Dudka fixed another NSS-related leak when client certs were used.
1389 - Bug report #2779245 (http://curl.haxx.se/bug/view.cgi?id=2779245) by Rainer
1390 Koenig pointed out that the man page didn't tell that the *_proxy
1391 environment variables can be specified lower case or UPPER CASE and the
1392 lower case takes precedence,
1394 Daniel Fandrich (21 Apr 2009)
1395 - Added new libcurl source files to Amiga, RiscOS and VC6 build files.
1397 Yang Tse (21 Apr 2009)
1398 - Moved potential inclusion of system's malloc.h and memory.h header files to
1399 setup_once.h. Inclusion of each header file is based on the definition of
1400 NEED_MALLOC_H and NEED_MEMORY_H respectively.
1402 Renamed libcurl's memory.h to curl_memory.h
1404 Daniel Stenberg (20 Apr 2009)
1405 - Leanic Lefever reported a crash and did some detailed research on why and
1406 how it occurs (http://curl.haxx.se/mail/lib-2009-04/0289.html). The
1407 conclusion was that if an error is detected and Curl_done() is called for
1408 the connection, ftp_done() could at times return another error code that
1409 then would take precedence and that new code confused existing logic that
1410 works for the first error code (CURLE_SEND_ERROR) only.
1412 - Gisle Vanem noticed that --libtool would produce bogus strings at times for
1413 OBJECTPOINT options. Now we've introduced a new function - my_setopt_str -
1414 within the app for setting plain string options to avoid the risk of this
1417 Daniel Stenberg (17 Apr 2009)
1418 - Pramod Sharma reported and tracked down a bug when doing FTP over a HTTP
1419 proxy. libcurl would then wrongly close the connection after each
1420 request. In his case it had the weird side-effect that it killed NTLM auth
1421 for the proxy causing an inifinite loop!
1423 I added test case 1098 to verify this fix. The test case does however not
1424 properly verify that the transfers are done persistently - as I couldn't
1425 think of a clever way to achieve it right now - but you need to read the
1426 stderr output after a test run to see that it truly did the right thing.
1428 Daniel Stenberg (13 Apr 2009)
1429 - bug report #2727981 (http://curl.haxx.se/bug/view.cgi?id=2727981) by Martin
1430 Storsjö pointed out how setting CURLOPT_NOBODY to 0 could be downright
1431 confusing as it set the method to either GET or HEAD. The example he showed
1434 curl_easy_setopt(curl, CURLOPT_PUT, 1);
1435 curl_easy_setopt(curl, CURLOPT_NOBODY, 0);
1437 The new way doesn't alter the method until the request is about to start. If
1438 CURLOPT_NOBODY is then 1 the HTTP request will be HEAD. If CURLOPT_NOBODY is
1439 0 and the request happens to have been set to HEAD, it will then instead be
1440 set to GET. I believe this will be less surprising to users, and hopefully
1441 not hit any existing users badly.
1443 - Toshio Kuratomi reported a memory leak problem with libcurl+NSS that turned
1444 out to be leaking cacerts. Kamil Dudka helped me complete the fix. The issue
1445 is found in Redhat's bug tracker:
1446 https://bugzilla.redhat.com/show_bug.cgi?id=453612
1448 There are still memory leaks present, but they seem to have other reasons.
1450 Daniel Fandrich (11 Apr 2009)
1451 - Added new libcurl source files to Symbian OS build files.
1452 - Improved Symbian support for SSL.
1454 Yang Tse (10 Apr 2009)
1455 - Daniel Johnson improved the MacOSX-Framework shell script to now perform all
1456 the steps required to build a Mac OS X four way fat ppc/i386/ppc64/x86_64
1457 libcurl.framework. Four way fat framework requires OS X 10.5 SDK or later.
1459 Yang Tse (8 Apr 2009)
1460 - Removed Sun compilers preprocessor block from curlbuild.h.dist, this also
1461 removes it from the curlbuild.h file originally distributed by the cURL
1462 project as this file is intended for systems not capable of running the
1463 configure script. For those who have been building curl out of the source
1464 code curl distribution tarball provided by curl.haxx.se the change implies
1465 nothing. Previous change in this area committed 2 Apr becomes irrelevant.
1467 Daniel Stenberg (6 Apr 2009)
1468 - I clarified in the docs that CURLOPT_SEEKFUNCTION should return 0 on success
1469 and 1 on fatal errors. Previously it only mentioned non-zero on fatal
1470 errors. This is a slight change in meaning, but it follows what we've done
1471 elsewhere before and it opens up for LOTS of more useful return codes
1472 whenever we can think of them...
1474 Yang Tse (2 Apr 2009)
1475 - Fix curl_off_t definition for builds done using Sun compilers and a
1476 non-configured libcurl. In this case curl_off_t data type was gated
1477 to the off_t data type which depends on the _FILE_OFFSET_BITS. This
1478 configuration is exactly the unwanted configuration for our curl_off_t
1479 data type which must not depend on such setting. This breaks ABI for
1480 libcurl libraries built with Sun compilers which were built without
1481 having run the configure script with _FILE_OFFSET_BITS different than
1482 64 and using the ILP32 data model.
1484 Daniel Stenberg (1 Apr 2009)
1485 - Andre Guibert de Bruet fixed a NULL pointer use in an infof() call if a
1486 strdup() call failed.
1488 Daniel Fandrich (31 Mar 2009)
1489 - Properly return an error code in curl_easy_recv (reported by Jim Freeman).
1491 Daniel Stenberg (18 Mar 2009)
1492 - Kamil Dudka brought a patch that enables 6 additional crypto algorithms when
1493 NSS is used. These ciphers were added in NSS 3.4 and require to be enabled
1496 Daniel Stenberg (13 Mar 2009)
1497 - Use libssh2_version() to present the libssh2 version in case the libssh2
1498 library is found to support it.
1500 Yang Tse (12 Mar 2009)
1501 - Added missing Curl_read() return code checking in TELNET transfers.
1503 - Pierre Brico found and fixed TELNET transfers not being aborted upon
1504 a write callback failure.
1506 Daniel Stenberg (11 Mar 2009)
1507 - Kamil Dudka made the curl tool properly call curl_global_init() before any
1508 other libcurl function.
1510 Yang Tse (11 Mar 2009)
1511 - Added missing TELNET timeout support for Windows builds. This issue was
1512 reported by Pierre Brico.
1514 Daniel Stenberg (9 Mar 2009)
1515 - Frank Hempel found out a bug and provided the fix:
1517 curl_easy_duphandle did not necessarily duplicate the CURLOPT_COOKIEFILE
1518 option. It only enabled the cookie engine in the destination handle if
1519 data->cookies is not NULL (where data is the source handle). In case of a
1520 newly initialized handle which just had the cookie support enabled by a
1521 curl_easy_setopt(handle, CURL_COOKIEFILE, "")-call, handle->cookies was
1522 still NULL because the setopt-call only appends the value to
1523 data->change.cookielist, hence duplicating this handle would not have the
1524 cookie engine switched on.
1526 We also concluded that the slist-functionality would be suitable for being
1527 put in its own module rather than simply hanging out in lib/sendf.c so I
1528 created lib/slist.[ch] for them.
1530 - Andreas Farber made the 'buildconf' script check for the presence of m4
1531 scripts to make it detect a bad checkout earlier. People with older
1532 checkouts who don't do cvs update with the -d option won't get the new dirs
1533 and then will get funny outputs that can be a bit hard to understand and
1536 Daniel Stenberg (8 Mar 2009)
1537 - Andre Guibert de Bruet found and fixed a code segment in ssluse.c where the
1538 allocation of the memory BIO was not being properly checked.
1540 - Andre Guibert de Bruet fixed the gnutls-using code: There are a few places
1541 in the gnutls code where we were checking for negative values for errors,
1542 when the man pages state that GNUTLS_E_SUCCESS is returned on success and
1543 other values indicate error conditions.
1545 - Bill Egert pointed out (http://curl.haxx.se/bug/view.cgi?id=2671602) that
1546 curl didn't use sprintf() in a way that is documented to work in POSIX but
1547 since we use our own printf() code (from libcurl) that shouldn't be a
1548 problem. Nonetheless I modified the code to not rely on such particular
1549 features and to not cause further raised eyebrowse with no good reason.
1551 Daniel Fandrich (5 Mar 2009)
1552 - Expanded the security section of the libcurl-tutorial man page to cover
1553 more issues for authors to consider when writing robust libcurl-using
1556 Yang Tse (5 Mar 2009)
1557 - Fixed NTLM authentication memory leak on SSPI enabled Windows builds. This
1558 issue was noticed by Chris Deidun.
1560 Daniel Fandrich (4 Mar 2009)
1561 - Fixed a problem with m4 quoting in the OpenSSL configure check reported
1564 Daniel Stenberg (3 Mar 2009)
1565 - David James brought a patch that make libcurl close (all) dead connections
1566 whenever you attempt to open a new connection.
1568 1. After cleaning up a dead connection, "continue" instead of
1569 returning FALSE. This ensures that we clean up all dead connections,
1570 rather than just cleaning up the first dead connection.
1571 2. Move up the cleanup for dead connections so that it occurs for
1572 all connections, rather than just the connections which have the same
1573 preferences as our current new connection.
1575 Version 7.19.4 (3 March 2009)
1577 Daniel Stenberg (3 Mar 2009)
1578 - David Kierznowski notified us about a security flaw
1579 (http://curl.haxx.se/docs/adv_20090303.html also known as CVE-2009-0037) in
1580 which previous libcurl versions (by design) can be tricked to access an
1581 arbitrary local/different file instead of a remote one when
1582 CURLOPT_FOLLOWLOCATION is enabled. This flaw is now fixed in this release
1583 together this the addition of two new setopt options for controlling this
1586 o CURLOPT_REDIR_PROTOCOLS controls what protocols libcurl is allowed to
1587 follow to when CURLOPT_FOLLOWLOCATION is enabled. By default, this option
1588 excludes the FILE and SCP protocols and thus you nee to explicitly allow
1589 them in your app if you really want that behavior.
1591 o CURLOPT_PROTOCOLS controls what protocol(s) libcurl is allowed to fetch
1592 using the primary URL option. This is useful if you want to allow a user or
1593 other outsiders control what URL to pass to libcurl and yet not allow all
1594 protocols libcurl may have been built to support.
1596 Daniel Stenberg (27 Feb 2009)
1597 - Senthil Raja Velu reported a problem when CURLOPT_INTERFACE and
1598 CURLOPT_LOCALPORT were used together (the local port bind failed), and
1599 Markus Koetter provided the fix!
1601 Daniel Stenberg (25 Feb 2009)
1602 - As Daniel Fandrich figured out, we must do the GnuTLS initing in the
1603 curl_global_init() function to properly maintain the performing functions
1604 thread-safe. We've previously (28 April 2007) moved the init to a later time
1605 just to avoid it to fail very early when libgcrypt dislikes the situation,
1606 but that move was bad and the fix should rather be in libgcrypt or
1609 Daniel Stenberg (24 Feb 2009)
1610 - Brian J. Murrell found out that Negotiate proxy authentication didn't work.
1611 It happened because the code used the struct for server-based auth all the
1612 time for both proxy and server auth which of course was wrong.
1614 Daniel Stenberg (23 Feb 2009)
1615 - After a bug reported by James Cheng I've made curl_easy_getinfo() for
1616 CURLINFO_CONTENT_LENGTH_DOWNLOAD and CURLINFO_CONTENT_LENGTH_UPLOAD return
1617 -1 if the sizes aren't know. Previously these returned 0, make it impossible
1618 to detect the difference between actually zero and unknown.
1620 Yang Tse (23 Feb 2009)
1621 - Daniel Johnson provided a shell script that will perform all the steps needed
1622 to build a Mac OS X fat ppc/i386 or ppc64/x86_64 libcurl.framework
1624 Daniel Stenberg (23 Feb 2009)
1625 - I renamed everything in the windows builds files that used the name 'curllib'
1626 to the proper 'libcurl' as clearly this caused confusion.
1628 Yang Tse (20 Feb 2009)
1629 - Do not halt compilation when using VS2008 to build a Windows 2000 target.
1631 Daniel Stenberg (20 Feb 2009)
1632 - Linus Nielsen Feltzing reported and helped me repeat and fix a problem with
1633 FTP with the multi interface: when a transfer fails, like when aborted by a
1634 write callback, the control connection was wrongly closed and thus not
1637 This change is also an attempt to cleanup the code somewhat in this area, as
1638 now the FTP code attempts to keep (better) track on pending responses
1639 necessary to get read in ftp_done().
1641 Daniel Stenberg (19 Feb 2009)
1642 - Patrik Thunstrom reported a problem and helped me repeat it. It turned out
1643 libcurl did a superfluous 1000ms wait when doing SFTP downloads!
1645 We read data with libssh2 while doing the "DO" operation for SFTP and then
1646 when we were about to start getting data for the actual file part, the
1647 "TRANSFER" part, we waited for socket action (in 1000ms) before doing a
1648 libssh2-read. But in this case libssh2 had already read and buffered the
1649 data so we ended up always just waiting 1000ms before we get working on the
1652 Patrick Monnerat (18 Feb 2009)
1653 - FTP downloads (i.e.: RETR) ending with code 550 now return error
1654 CURLE_REMOTE_FILE_NOT_FOUND instead of CURLE_FTP_COULDNT_RETR_FILE.
1656 Daniel Stenberg (17 Feb 2009)
1657 - Kamil Dudka made NSS-powered builds compile and run again!
1659 - A second follow-up change by Andre Guibert de Bruet to fix a related memory
1660 leak like that fixed on the 14th. When zlib returns failure, we need to
1661 cleanup properly before returning error.
1663 - CURLOPT_FTP_CREATE_MISSING_DIRS can now be set to 2 in addition to 1 for
1664 plain FTP connections, and it will then allow MKD to fail once and retry the
1665 CWD afterwards. This is especially useful if you're doing many simultanoes
1666 connections against the same server and they all have this option enabled,
1667 as then CWD may first fail but then another connection does MKD before this
1668 connection and thus MKD fails but trying CWD works! The numbers can
1669 (should?) now be set with the convenience enums now called
1670 CURLFTP_CREATE_DIR and CURLFTP_CREATE_DIR_RETRY.
1672 Tests has proven that if you're making an application that uploads a set of
1673 files to an ftp server, you will get a noticable gain in speed if you're
1674 using multiple connections and this option will be then be very useful.
1676 Daniel Stenberg (14 Feb 2009)
1677 - Andre Guibert de Bruet found and fixed a memory leak in the content encoding
1678 code, which could happen on libz errors.
1680 Daniel Fandrich (12 Feb 2009)
1681 - Added support for Digest and NTLM authentication using GnuTLS.
1683 Daniel Stenberg (11 Feb 2009)
1684 - CURLINFO_CONDITION_UNMET was added to allow an application to get to know if
1685 the condition in the previous request was unmet. This is typically a time
1686 condition set with CURLOPT_TIMECONDITION and was previously not possible to
1687 reliably figure out. From bug report #2565128
1688 (http://curl.haxx.se/bug/view.cgi?id=2565128) filed by Jocelyn Jaubert.
1690 Daniel Fandrich (4 Feb 2009)
1691 - Don't add the standard /usr/lib or /usr/include paths to LDFLAGS and CPPFLAGS
1692 (respectively) when --with-ssl=/usr is used (patch based on FreeBSD).
1694 - Added an explicit buffer limit check in msdosify() (patch based on FreeBSD).
1695 This couldn't ever overflow in curl, but might if the code were used
1696 elsewhere or under different conditions.
1698 Daniel Stenberg (3 Feb 2009)
1699 - Hidemoto Nakada provided a small fix that makes it possible to get the
1700 CURLINFO_CONTENT_LENGTH_DOWNLOAD size from file:// "transfers" with
1701 CURLOPT_NOBODY set true.
1703 Daniel Stenberg (2 Feb 2009)
1704 - Patrick Scott found a rather large memory leak when using the multi
1705 interface and setting CURLMOPT_MAXCONNECTS to something less than the number
1706 of handles you add to the multi handle. All the connections that didn't fit
1707 in the cache would not be properly disconnected nor freed!
1709 - Craig A West brought us: libcurl now defaults to do CONNECT with HTTP
1710 version 1.1 instead of 1.0 like before. This change also introduces the new
1711 proxy type for libcurl called 'CURLPROXY_HTTP_1_0' that then allows apps to
1712 switch (back) to CONNECT 1.0 requests. The curl tool also got a --proxy1.0
1713 option that works exactly like --proxy but sets CURLPROXY_HTTP_1_0.
1715 I updated all test cases cases that use CONNECT and I tried to do some using
1716 --proxy1.0 and some updated to do CONNECT 1.1 to get both versions run.
1718 Daniel Stenberg (31 Jan 2009)
1719 - When building with c-ares 1.6.1 (not yet released) or later and IPv6 support
1720 enabled, we can now take advantage of its brand new AF_UNSPEC support in
1721 ares_gethostbyname(). This makes test case 241 finally run fine for me with
1722 this setup since it now parses the "::1 ip6-localhost" line fine in my
1725 Daniel Stenberg (30 Jan 2009)
1726 - Scott Cantor filed bug report #2550061
1727 (http://curl.haxx.se/bug/view.cgi?id=2550061) mentioning that I failed to
1728 properly make sure that the VC9 makefiles got included in the latest
1729 release. I've now fixed the release script and verified it so next release
1730 will hopefully include them properly!
1732 Daniel Fandrich (30 Jan 2009)
1733 - Fixed --disable-proxy for FTP and SOCKS. Thanks to Daniel Egger for
1736 Yang Tse (29 Jan 2009)
1737 - Introduced curl_sspi.c and curl_sspi.h for the implementation of functions
1738 Curl_sspi_global_init() and Curl_sspi_global_cleanup() which previously were
1739 named Curl_ntlm_global_init() and Curl_ntlm_global_cleanup() in http_ntlm.c
1740 Also adjusted socks_sspi.c to remove the link-time dependency on the Windows
1741 SSPI library using it now in the same way as it was done in http_ntlm.c.
1743 Daniel Stenberg (28 Jan 2009)
1744 - Markus Moeller introduced two new options to libcurl:
1745 CURLOPT_SOCKS5_GSSAPI_SERVICE and CURLOPT_SOCKS5_GSSAPI_NEC to allow libcurl
1746 to do GSS-style authentication with SOCKS5 proxies. The curl tool got the
1747 options called --socks5-gssapi-service and --socks5-gssapi-nec to enable
1750 Daniel Stenberg (26 Jan 2009)
1751 - Chad Monroe provided the new CURLOPT_TFTP_BLKSIZE option that allows an app
1752 to set desired block size to use for TFTP transfers instead of the default
1755 - The "-no_ticket" option was introduced in Openssl0.9.8j. It's a flag to
1756 disable "rfc4507bis session ticket support". rfc4507bis was later turned
1757 into the proper RFC5077 it seems: http://tools.ietf.org/html/rfc5077
1759 The enabled extension concerns the session management. I wonder how often
1760 libcurl stops a connection and then resumes a TLS session. also, sending the
1761 session data is some overhead. .I suggest that you just use your proposed
1762 patch (which explicitly disables TICKET).
1764 If someone writes an application with libcurl and openssl who wants to
1765 enable the feature, one can do this in the SSL callback.
1767 Sharad Gupta brought this to my attention. Peter Sylvester helped me decide
1768 on the proper action.
1770 - Alexey Borzov filed bug report #2535504
1771 (http://curl.haxx.se/bug/view.cgi?id=2535504) pointing out that realms with
1772 quoted quotation marks in HTTP Digest headers didn't work. I've now added
1773 test case 1095 that verifies my fix.
1775 - Craig A West brought CURLOPT_NOPROXY and the corresponding --noproxy option.
1776 They basically offer the same thing the NO_PROXY environment variable only
1777 offered previously: list a set of host names that shall not use the proxy
1778 even if one is specified.
1780 Daniel Fandrich (20 Jan 2009)
1781 - Call setlocale() for libtest tests to test the effects of locale-induced
1782 libc changes on libcurl.
1784 - Fixed a couple more locale-dependent toupper conversions, mainly for
1785 clarity. This does fix one problem that causes ;type=i FTP URLs
1786 to fail in the Turkish locale when CURLOPT_PROXY_TRANSFER_MODE is
1787 used (test case 561)
1789 - Added tests 561 and 1091 through 1094 to test various combinations
1790 of ;type= and ;mode= URLs that could potentially fail in the Turkish
1793 Daniel Stenberg (20 Jan 2009)
1794 - Lisa Xu pointed out that the ssh.obj file was missing from the
1795 lib/Makefile.vc6 file (and thus from the vc8 and vc9 ones too).
1797 Version 7.19.3 (19 January 2009)
1799 Daniel Stenberg (16 Jan 2009)
1800 - Andrew de los Reyes fixed curlbuild.h for "generic" gcc builds on PPC, both
1803 Daniel Stenberg (15 Jan 2009)
1804 - Tim Ansell fixed a compiler warning in lib/cookie.c
1806 Daniel Stenberg (14 Jan 2009)
1807 - Grant Erickson fixed timeouts for TFTP such that specifying a
1808 connect-timeout, a max-time or both options work correctly and as expected
1809 by passing the correct boolean value to Curl_timeleft via the
1810 'duringconnect' parameter.
1812 With this small change, curl TFTP now behaves as expected (and likely as
1813 originally-designed):
1815 1) For non-existent or unreachable dotted IP addresses:
1817 a) With no options, follows the default curl 300s timeout...
1818 b) With --connect-timeout only, follows that value...
1819 c) With --max-time only, follows that value...
1820 d) With both --connect-timeout and --max-time, follows the smaller value...
1822 and times out with a "curl: (7) Couldn't connect to server" error.
1824 2) For transfers to/from a valid host:
1826 a) With no options, follows default curl 300s timeout for the
1827 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1828 timeout for the remainder of the transfer...
1830 b) With --connect-time only, follows that value for the
1831 first XRQ/DATA/ACK transaction and the default TFTP 3600s
1832 timeout for the remainder of the transfer...
1834 c) With --max-time only, follows that value for the first
1835 XRQ/DATA/ACK transaction and for the remainder of the
1838 d) With both --connect-timeout and --max-time, follows the former
1839 for the first XRQ/DATA/ACK transaction and the latter for the
1840 remainder of the transfer...
1842 and times out with a "curl: (28) Timeout was reached" error as
1845 Daniel Stenberg (13 Jan 2009)
1846 - Michael Wallner fixed a NULL pointer deref when calling
1847 curl_easy_setup(curl, CURLOPT_COOKIELIST, "SESS") on a CURL handle with no
1850 - Stefan Teleman brought a patch to fix the default curlbuild.h file for the
1853 Daniel Stenberg (12 Jan 2009)
1854 - Based on bug report #2498665 (http://curl.haxx.se/bug/view.cgi?id=2498665)
1855 by Daniel Black, I've now added magic to the configure script that makes it
1856 use pkg-config to detect gnutls details as well if the existing method
1857 (using libgnutls-config) fails. While doing this, I cleaned up and unified
1858 the pkg-config usage when detecting openssl and nss as well.
1860 Daniel Stenberg (11 Jan 2009)
1861 - Karl Moerder brought the patch that creates vc9 Makefiles, and I made
1862 'maketgz' now use the actual makefile targets to do the VC8 and VC9
1865 Daniel Stenberg (10 Jan 2009)
1866 - Emil Romanus fixed:
1868 When using the multi interface over HTTP and the server returns a Location
1869 header, the running easy handle will get stuck in the CURLM_STATE_PERFORM
1870 state, leaving the external event loop stuck waiting for data from the
1871 ingoing socket (when using the curl_multi_socket_action stuff). While this
1872 bug was pretty hard to find, it seems to require only a one-line fix. The
1873 break statement on line 1374 in multi.c caused the function to skip the call
1876 How to reproduce this bug? Well, that's another question. evhiperfifo.c in
1877 the examples directory chokes on this bug only _sometimes_, probably
1878 depending on how fast the URLs are added. One way of testing the bug out is
1879 writing to hiper.fifo from more than one source at the same time.
1881 Daniel Fandrich (7 Jan 2009)
1882 - Unified much of the SessionHandle initialization done in Curl_open() and
1883 curl_easy_reset() by creating Curl_init_userdefined(). This had the side
1884 effect of fixing curl_easy_reset() so it now also resets
1885 CURLOPT_FTP_FILEMETHOD and CURLOPT_SSL_SESSIONID_CACHE
1887 Daniel Stenberg (7 Jan 2009)
1888 - Rob Crittenden did once again provide an NSS update:
1890 I have to jump through a few hoops now with the NSS library initialization
1891 since another part of an application may have already initialized NSS by the
1892 time Curl gets invoked. This patch is more careful to only shutdown the NSS
1893 library if Curl did the initialization.
1895 It also adds in a bit of code to set the default ciphers if the app that
1896 call NSS_Init* did not call NSS_SetDomesticPolicy() or set specific
1897 ciphers. One might argue that this lets other application developers get
1898 lazy and/or they aren't using the NSS API correctly, and you'd be right.
1899 But still, this will avoid terribly difficult-to-trace crashes and is
1902 Daniel Stenberg (1 Jan 2009)
1903 - 'reconf' is removed since we rather have users use 'buildconf'
1905 Daniel Stenberg (31 Dec 2008)
1906 - Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
1907 out that 'reconf' didn't properly point out the m4 subdirectory when running
1910 Daniel Stenberg (29 Dec 2008)
1911 - Phil Lisiecki filed bug report #2413067
1912 (http://curl.haxx.se/bug/view.cgi?id=2413067) that identified a problem that
1913 would cause libcurl to mark a DNS cache entry "in use" eternally if the
1914 subsequence TCP connect failed. It would thus never get pruned and refreshed
1915 as it should've been.
1917 Phil provided his own patch to this problem that while it seemed to work
1918 wasn't complete and thus I wrote my own fix to the problem.
1920 Daniel Stenberg (28 Dec 2008)
1921 - Peter Korsgaard fixed building libcurl with "configure --with-ssl
1924 - Anthony Bryan fixed more language and spelling flaws in man pages.
1926 Daniel Stenberg (22 Dec 2008)
1927 - Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
1928 on file indexes beyond 2 or 4GB.
1930 - Anthony Bryan provided a set of patches that cleaned up manual language,
1931 corrected spellings and more.
1933 Daniel Stenberg (20 Dec 2008)
1934 - Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
1935 pipelining, as libcurl could then easily get confused and A) work on the
1936 handle that was not "first in queue" on a pipeline, or even B) tell the app
1937 to REMOVE a socket while it was in use by a second handle in a pipeline. Both
1938 errors caused hanging or stalling applications.
1940 Daniel Stenberg (19 Dec 2008)
1941 - curl_multi_timeout() could return a timeout value of 0 even though nothing
1942 was actually ready to get done, as the internal time resolution is higher
1943 than the returned millisecond timer. Therefore it could cause applications
1944 running on fast processors to do short bursts of busy-loops.
1945 curl_multi_timeout() will now only return 0 if the timeout is actually
1948 - Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
1949 now has an improved ability to do right when the multi interface (both
1950 "regular" and multi_socket) is used for SCP and SFTP transfers. This should
1951 result in (much) less busy-loop situations and thus less CPU usage with no
1954 Daniel Stenberg (17 Dec 2008)
1955 - SCP and SFTP with the multi interface had the same flaw: the 'DONE'
1956 operation didn't complete properly if the EAGAIN equivalent was returned but
1957 libcurl would simply continue with a half-completed close operation
1958 performed. This ruined persistent connection re-use and cause some
1959 SSH-protocol errors in general. The correction is unfortunately adding a
1960 blocking function - doing it entirely non-blocking should be considered for
1963 Gisle Vanem (16 Dec 2008)
1964 - Added the possibility to use the Watt-32 tcp/ip stack under Windows.
1965 The change simply involved adding a USE_WATT32 section in the
1966 config-win32.h files (under ./lib and ./src). This section disables
1967 the use of any Winsock headers.
1969 Daniel Stenberg (16 Dec 2008)
1970 - libssh2_sftp_last_error() was wrongly used at some places in libcurl which
1971 made libcurl sometimes not properly abort problematic SFTP transfers.
1973 Daniel Stenberg (12 Dec 2008)
1974 - More work with Igor Novoseltsev to first fix the remaining stuff for
1975 removing easy handles from multi handles when the easy handle is/was within
1976 a HTTP pipeline. His bug report #2351653
1977 (http://curl.haxx.se/bug/view.cgi?id=2351653) was also related and was
1978 eventually fixed by a patch by Igor himself.
1980 Yang Tse (12 Dec 2008)
1981 - Patrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
1982 OS/400 compilations with IPv6 enabled.
1984 Daniel Stenberg (12 Dec 2008)
1985 - Mark Karpeles filed bug report #2416182 titled "crash in ConnectionExists
1986 when using duphandle+curl_mutli"
1987 (http://curl.haxx.se/bug/view.cgi?id=2416182) which showed that
1988 curl_easy_duphandle() wrongly also copied the pointer to the connection
1989 cache, which was plain wrong and caused a segfault if the handle would be
1990 used in a different multi handle than the handle it was duplicated from.
1992 Daniel Stenberg (11 Dec 2008)
1993 - Keshav Krity found out that libcurl failed to deal with dotted IPv6
1994 addresses if they were very long (>39 letters) due to a too strict address
1995 validity parser. It now accepts addresses up to 45 bytes long.
1997 Daniel Stenberg (11 Dec 2008)
1998 - Internet Explorer had a broken HTTP digest authentication before v7 and
1999 there are servers "out there" that relies on the client doing this broken
2000 Digest authentication. Apache even comes with an option to work with such
2003 The difference is only for URLs that contain a query-part (a '?'-letter and
2004 text to the right of it).
2006 libcurl now supports this quirk, and you enable it by setting the
2007 CURLAUTH_DIGEST_IE bit in the bitmask you pass to the CURLOPT_HTTPAUTH or
2008 CURLOPT_PROXYAUTH options. They are thus individually controlled to server
2011 (note that there's no way to activate this with the curl tool yet)
2013 Daniel Fandrich (9 Dec 2008)
2014 - Added test cases 1089 and 1090 to test --write-out after a redirect to
2015 test a report that the size didn't work, but these test cases pass.
2017 - Documented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.
2019 Daniel Stenberg (9 Dec 2008)
2020 - Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
2021 particular state for the control connection like it did before for implicit
2022 FTPS (libcurl assumed such control connections to be encrypted while some
2023 FTPS servers such as FileZilla assumes such connections to be clear
2024 mode). Use the CURLOPT_USE_SSL option to set your desired level.
2026 Daniel Stenberg (8 Dec 2008)
2027 - Fred Machado posted about a weird FTP problem on the curl-users list and when
2028 researching it, it turned out he got a 550 response back from a SIZE command
2029 and then I fell over the text in RFC3659 that says:
2031 The presence of the 550 error response to a SIZE command MUST NOT be taken
2032 by the client as an indication that the file cannot be transferred in the
2033 current MODE and TYPE.
2035 In other words: the change I did on September 30th 2008 and that has been
2036 included in the last two releases were a regression and a bad idea. We MUST
2037 NOT take a 550 response from SIZE as a hint that the file doesn't exist.
2039 - Christian Krause filed bug #2221237
2040 (http://curl.haxx.se/bug/view.cgi?id=2221237) that identified an infinite
2041 loop during GSS authentication given some specific conditions. With his
2042 patience and great feedback I managed to narrow down the problem and
2043 eventually fix it although I can't test any of this myself!
2045 Daniel Fandrich (3 Dec 2008)
2046 - Fixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
2047 support (e.g. Minix)
2049 Daniel Stenberg (3 Dec 2008)
2050 - Igor Novoseltsev filed bug #2351645
2051 (http://curl.haxx.se/bug/view.cgi?id=2351645) that identified a problem with
2052 the multi interface that occured if you removed an easy handle while in
2053 progress and the handle was used in a HTTP pipeline.
2055 - Pawel Kierski pointed out a mistake in the cookie code that could lead to a
2056 bad fclose() after a fatal error had occured.
2057 (http://curl.haxx.se/bug/view.cgi?id=2382219)
2059 Daniel Fandrich (25 Nov 2008)
2060 - If a HTTP request is Basic and num is already >=1000, the HTTP test
2061 server adds 1 to num to get the data section to return. This allows
2062 testing authentication negotiations using the Basic authentication
2065 - Added tests 1087 and 1088 to test Basic authentication on a redirect
2066 with and without --location-trusted
2068 Daniel Stenberg (24 Nov 2008)
2069 - Based on a patch by Vlad Grachov, libcurl now uses a new libssh2 0.19
2070 function when built to support SCP and SFTP that helps the library to know
2071 in which direction a particular libssh2 operation would return EAGAIN so
2072 that libcurl knows what socket conditions to wait for before trying the
2073 function call again. Previously (and still when using libssh2 0.18 or
2074 earlier), libcurl will busy-loop in this situation when the easy interface
2077 Daniel Fandrich (20 Nov 2008)
2078 - Automatically detect OpenBSD's CA cert bundle.
2080 Daniel Stenberg (19 Nov 2008)
2081 - I removed the default use of "Pragma: no-cache" from libcurl when a proxy is
2082 used. It has been used since forever but it was never a good idea to use
2083 unless explicitly asked for.
2085 - Josef Wolf's extension that allows a $TESTDIR/gdbinit$testnum file that when
2086 you use runtests.pl -g, will be sourced by gdb to allow additional fancy or
2087 whatever you see fit
2089 - Christian Krause reported and fixed a memory leak that would occur with HTTP
2090 GSS/kerberos authentication (http://curl.haxx.se/bug/view.cgi?id=2284386)
2092 - Andreas Wurf and Markus Koetter helped me analyze a problem that Andreas got
2093 when uploading files to a single FTP server using multiple easy handle
2094 handles with the multi interface. Occasionally a handle would stall in
2097 The problem turned out to be a side-effect of the ConnectionExists()
2098 function's eagerness to re-use a handle for HTTP pipelining so it would
2099 select it even if already being in use, due to an inadequate check for its
2100 chances of being used for pipelnining.
2102 Daniel Fandrich (17 Nov 2008)
2103 - Added more compiler warning options for gcc 4.3
2105 Yang Tse (17 Nov 2008)
2106 - Fix a remaining problem in the inet_pton() runtime configure check. And
2107 fix internal Curl_inet_pton() failures to reject certain malformed literals.
2109 - Make configure script check if ioctl with the SIOCGIFADDR command can be
2110 used, and define HAVE_IOCTL_SIOCGIFADDR if appropriate.
2112 Daniel Stenberg (16 Nov 2008)
2113 - Christian Krause fixed a build failure when building with gss support
2114 enabled and FTP disabled.
2116 - Added check for NULL returns from strdup() in src/main.c and lib/formdata.c
2117 - reported by Jim Meyering also prevent buffer overflow on MSDOS when you do
2118 for example -O on a url with a file name part longer than PATH_MAX letters
2120 - lib/nss.c fixes based on the report by Jim Meyering: I went over and added
2121 checks for return codes for all calls to malloc and strdup that were
2122 missing. I also changed a few malloc(13) to use arrays on the stack and a
2123 few malloc(PATH_MAX) to instead use aprintf() to lower memory use.
2125 - I fixed a memory leak in Curl_nss_connect() when CURLOPT_ISSUERCERT is
2128 Daniel Fandrich (14 Nov 2008)
2129 - Added .xml as one of the few common file extensions known by the multipart
2132 - Added some #ifdefs around header files and change the EAGAIN test to
2133 fix compilation on Cell (reported by Jeff Curley).
2135 Yang Tse (14 Nov 2008)
2136 - Fixed several configure script issues affecting checks for inet_ntoa_r(),
2137 inet_ntop(), inet_pton(), getifaddrs(), fcntl() and getaddrinfo().
2139 Yang Tse (13 Nov 2008)
2140 - Refactored configure script detection of functions used to set sockets into
2141 non-blocking mode, and decouple function detection from function capability.