9
Version 7.16.2 (11 April 2007)
11
Yang Tse (10 April 2007)
12
- Ravi Pratap provided some fixes for HTTP pipelining
14
- configure script will ignore --enable-sspi option for non-native Windows.
16
Daniel S (9 April 2007)
17
- Nick Zitzmann did ssh.c cleanups
19
Daniel S (3 April 2007)
20
- Rob Jones fixed better #ifdef'ing for a bunch of #include lines.
22
Daniel S (2 April 2007)
23
- Nick Zitzmann made the CURLOPT_POSTQUOTE option work for SFTP as well. The
24
accepted commands are as follows:
27
Changes the group ID of the file or directory at (path) to (gid). (gid)
31
Changes the permissions of the file or directory at (path) to
32
(perms). (perms) must be a number in the format used by the chmod Unix
36
Changes the user ID of the file or directory at (path) to (uid). (uid)
40
Creates a symbolic link at (dest) that points to the file located at
44
Creates a new directory at (path).
46
rename (source) (dest)
47
Moves the file or directory at (source) to (dest).
50
Deletes the file located at (path).
53
Deletes the directory located at (path). This command will raise an error
54
if the directory is not empty.
56
symlink (source) (dest)
59
Daniel S (1 April 2007)
60
- Robert Iakobashvili made curl_multi_remove_handle() a lot faster when many
61
easy handles are added to a multi handle, by avoiding the looping over all
62
the handles to find which one to remove.
64
- Matt Kraai provided a patch that makes curl build on QNX 6 fine again.
66
Daniel S (31 March 2007)
67
- Fixed several minor issues detected by the coverity.com scanner.
69
- "Pixel" fixed a problem that appeared when you used -f with user+password
73
- Don't tear down the ftp connection if the maximum filesize was exceeded
74
and added tests 290 and 291 to check.
76
- Added ftps upload and SSL required tests 401 and 402.
78
- Send an EOF message before closing an SCP channel, as recommended by
79
RFC4254. Enable libssh2 tracing when ssh debugging is turned on.
81
Yang Tse (27 March 2007)
82
- Internal function Curl_select() renamed to Curl_socket_ready()
84
New Internal wrapper function Curl_select() around select (2), it
85
uses poll() when a fine poll() is available, so now libcurl can be
86
built without select() support at all if a fine poll() is available.
88
Daniel S (25 March 2007)
89
- Daniel Johnson fixed multi code to traverse the easy handle list properly.
90
A left-over bug from the February 21 fix.
93
- Added --pubkey option to curl and made --key also work for SCP/SFTP,
94
plus made --pass work on an SSH private key as well.
96
- Changed the test harness to attempt to gracefully shut down servers
97
before resorting to the kill -9 hammer.
99
- Added test harness infrastructure to support scp/sftp tests, using
100
OpenSSH as the server.
102
- Fixed a memory leak when specifying a proxy with a file: URL.
104
Yang Tse (20 March 2007)
105
- Fixed: When a signal was caught awaiting for an event using Curl_select()
106
or Curl_poll() with a non-zero timeout both functions would restart the
107
specified timeout. This could even lead to the extreme case that if a
108
signal arrived with a frecuency lower to the specified timeout neither
109
function would ever exit.
111
Added experimental symbol definition check CURL_ACKNOWLEDGE_EINTR in
112
Curl_select() and Curl_poll(). When compiled with CURL_ACKNOWLEDGE_EINTR
113
defined both functions will return as soon as a signal is caught. Use it
114
at your own risk, all calls to these functions in the library should be
115
revisited and checked before fully supporting this feature.
117
Yang Tse (19 March 2007)
118
- Bryan Henderson fixed the progress function so that it can get called more
119
frequently allowing same calling frecuency for the client progress callback.
121
Dan F (15 March 2007)
122
- Various memory leaks plugged and NULL pointer fixes made in the ssh code.
124
Daniel (15 March 2007)
125
- Nick made the curl tool accept globbing ranges that only is one number, i.e
126
you can now use [1-1] without curl complaining.
128
Daniel (10 March 2007)
131
The problem is the following: when we're calling Curl_done and it decides to
132
keep the connection opened ('left intact'), then the caller is not notified
133
that the connection was done via the NULLifying of the pointer, so some easy
134
handle is keeping the pointer to this connection.
136
Later ConnectionExists can select such connection for reuse even if we're
137
not pipelining: pipeLen is zero, so the (pipeLen > 0 && !canPipeline) is
138
false and we can reuse this connection for another easy handle. But thus the
139
connection will be shared between two easy handles if the handle that wants
140
to take the ownership is not the same as was not notified of the connection
141
was done in Curl_done. And when some of these easy handles will get their
142
connection really freed the another one will still keep the pointer.
144
My fix was rather trivial: I just added the NULLification to the 'else'
145
branch in the Curl_done. My tests with Git and ElectricFence showed no
146
problems both for HTTP pulling and cloning. Repository size is about 250 Mb,
147
so it was a considerable amount of Curl's work.
150
- Updated the test harness to add a new "crypto" feature check and updated the
151
appropriate test case to use it. For now, this is treated the same as the
152
"SSL" feature because curl doesn't list it separately.
154
Daniel (9 March 2007)
155
- Robert Iakobashvili fixed CURLOPT_INTERFACE for IPv6.
157
- Robert A. Monat improved the maketgz and VC6/8 generating to set the correct
160
- Justin Fletcher fixed a file descriptor leak in the curl tool when trying to
161
upload a file it couldn't open. Bug #1676581
162
(http://curl.haxx.se/bug/view.cgi?id=1676581)
165
- Updated the test harness to check for protocol support before running each
166
test, fixing KNOWN_BUGS #11.
169
- Reintroduced (after a 3 year hiatus) an FTPS test case (400) into the test
170
harness. It is very limited as it supports only ftps:// URLs with
171
--ftp-ssl-control specified, which implicitly encrypts the control
172
channel but not the data channels. That allows stunnel to be used with
173
an unmodified ftp server in exactly the same way that the test https
177
- Honour --ftp-ssl-control on ftps:// URLs to allow encrypted control and
178
unencrypted data connections.
181
- Fixed a couple of improper pointer uses detected by valgrind in test
184
Daniel (2 March 2007)
185
- Robert A. Monat and Shmulik Regev helped out to fix the new */Makefile.vc8
186
makefiles that are included in the source release archives, generated from
187
the Makefile.vc6 files by the maketgz script. I also modified the root
188
Makefile to have a VC variable that defaults to vc6 but can be overridden to
189
allow it to be used for vc8 as well. Like this:
193
Daniel (27 February 2007)
194
- Hang Kin Lau found and fixed: When I use libcurl to connect to an https
195
server through a proxy and have the remote https server port set using the
196
CURLOPT_PORT option, protocol gets reset to http from https after the first
199
User defined URL was modified internally by libcurl and subsequent reuse of
200
the easy handle may lead to connection using a different protocol (if not
203
I found that libcurl hardcoded the protocol to "http" when it tries to
204
regenerate the URL if CURLOPT_PORT is set. I tried to fix the problem as
205
follows and it's working fine so far
207
Daniel (25 February 2007)
208
- Adam D. Moss made the HTTP CONNECT procedure less blocking when used from
209
the multi interface. Note that it still does a part of the connection in a
212
Daniel (23 February 2007)
213
- Added warning outputs if the command line uses more than one of the options
214
-v, --trace and --trace-ascii, since it could really confuse the user.
215
Clarified this fact in the man page.
217
Daniel (21 February 2007)
218
- Ravi Pratap provided work on libcurl making pipelining more robust and
220
o Don't mix GET and POST requests in a pipeline
221
o Fix the order in which requests are dispatched from the pipeline
222
o Fixed several curl bugs with pipelining when the server is returning
224
* Added states to chunked parsing for final CRLF
225
* Rewind buffer after parsing chunk with data remaining
226
* Moved chunked header initializing to a spot just before receiving
229
Daniel (20 February 2007)
230
- Linus Nielsen Feltzing changed the CURLOPT_FTP_SSL_CCC option to handle
231
active and passive CCC shutdown and added the --ftp-ssl-ccc-mode command
234
Daniel (19 February 2007)
235
- Ian Turner fixed the libcurl.m4 macro's support for --with-libcurl.
237
- Shmulik Regev found a memory leak in re-used HTTPS connections, at least
238
when the multi interface was used.
240
- Robson Braga Araujo made passive FTP transfers work with SOCKS (both 4 and
243
Daniel (18 February 2007)
244
- Jeff Pohlmeyer identified two problems: first a rather obscure problem with
245
the multi interface and connection re-use that could make a
246
curl_multi_remove_handle() ruin a pointer in another handle.
248
The second problem was less of an actual problem but more of minor quirk:
249
the re-using of connections wasn't properly checking if the connection was
252
Daniel (16 February 2007)
253
- Duncan Mac-Vicar Prett and Michal Marek reported problems with resetting
254
CURLOPT_RANGE back to no range on an easy handle when using FTP.
256
Dan F (14 February 2007)
257
- Fixed curl-config --libs so it doesn't list unnecessary libraries (and
258
therefore introduce unnecessary dependencies) when it's not needed.
259
Also, don't bother adding a library path of /usr/lib
261
Daniel (13 February 2007)
262
- The default password for anonymous FTP connections is now changed to be
265
- Robert A. Monat made libcurl build fine with VC2005 - it doesn't have
266
gmtime_r() like the older VC versions. He also made use of some machine-
267
specific defines to differentiate the "OS" define.
269
Daniel (12 February 2007)
270
- Rob Crittenden added support for NSS (Network Security Service) for the
271
SSL/TLS layer. http://www.mozilla.org/projects/security/pki/nss/
273
This is the fourth supported library for TLS/SSL that libcurl supports!
275
- Shmulik Regev fixed so that the final CRLF of HTTP response headers are sent
276
to the debug callback.
278
- Shmulik Regev added CURLOPT_HTTP_CONTENT_DECODING and
279
CURLOPT_HTTP_TRANSFER_DECODING that if set to zero will disable libcurl's
280
internal decoding of content or transfer encoded content. This may be
281
preferable in cases where you use libcurl for proxy purposes or similar. The
282
command line tool got a --raw option to disable both at once.
284
- release tarballs made with maketgz will from now on have a LIBCURL_TIMESTAMP
285
define set to hold the exact date and time of when the tarball was built, as
286
a human readable string using the UTC time zone.
288
- Jeff Pohlmeyer fixed a flaw in curl_multi_add_handle() when adding a handle
289
that has an easy handle present in the "closure" list pending closure.
291
Daniel (6 February 2007)
292
- Regular file downloads wiht SFTP and SCP are now done using the non-blocking
293
API of libssh2, if the libssh2 headers seem to support them. This will make
294
SCP and SFTP much more responsive and better libcurl citizens when used with
295
the multi interface etc.
297
Daniel (5 February 2007)
298
- Michael Wallner added support for CURLOPT_TIMEOUT_MS and
299
CURLOPT_CONNECTTIMEOUT_MS that, as their names suggest, do the timeouts with
300
millisecond resolution. The only restriction to that is the alarm()
301
(sometimes) used to abort name resolves as that uses full seconds. I fixed
302
the FTP response timeout part of the patch.
304
Internally we now count and keep the timeouts in milliseconds but it also
305
means we multiply set timeouts with 1000. The effect of this is that no
306
timeout can be set to more than 2^31 milliseconds (on 32 bit systems), which
307
equals 24.86 days. We probably couldn't before either since the code did
308
*1000 on the timeout values on several places already.
310
Daniel (3 February 2007)
311
- Yang Tse fixed the cookie expiry date in several test cases that started to
312
fail since they used "1 feb 2007"...
314
- Manfred Schwarb reported that socks5 support was broken and help us pinpoint
315
the problem. The code now tries harder to use httproxy and proxy where
316
apppropriate, as not all proxies are HTTP...
318
Version 7.16.1 (29 January 2007)
320
Daniel (29 January 2007)
321
- Michael Wallner reported that when doing a CONNECT with a custom User-Agent
322
header, you got _two_ User-Agent headers in the CONNECT request...! Added
323
test case 287 to verify the fix.
325
Daniel (28 January 2007)
326
- curl_easy_reset() now resets the CA bundle path correctly.
328
- David McCreedy fixed the Curl command line tool for HTTP on non-ASCII
331
Daniel (25 January 2007)
332
- Added the --libcurl [file] option to curl. Append this option to any
333
ordinary curl command line, and you will get a libcurl-using source code
334
written to the file that does the equivalent operation of what your command
337
Dan F (24 January 2007)
338
- Fixed a dangling pointer problem that prevented the http_proxy environment
339
variable from being properly used in many cases (and caused test case 63
342
Daniel (23 January 2007)
343
- David McCreedy did NTLM changes mainly for non-ASCII platforms:
346
There's a compilation error in http_ntlm.c if USE_NTLM2SESSION is NOT
347
defined. I noticed this while testing various configurations. Line 867 of
348
the current http_ntlm.c is a closing bracket for an if/else pair that only
349
gets compiled in if USE_NTLM2SESSION is defined. But this closing bracket
350
wasn't in an #ifdef so the code fails to compile unless USE_NTLM2SESSION was
351
defined. Lines 198 and 140 of my patch wraps that closing bracket in an
352
#ifdef USE_NTLM2SESSION.
355
I noticed several picky compiler warnings when DEBUG_ME is defined. I've
356
fixed them with casting. By the way, DEBUG_ME was a huge help in
357
understanding this code.
360
Hopefully the last non-ASCII conversion patch for libcurl in a while. I
361
changed the "NTLMSSP" literal to hex since this signature must always be in
364
Conversion code was strategically added where necessary. And the
365
Curl_base64_encode calls were changed so the binary "blobs" http_ntlm.c
366
creates are NOT translated on non-ASCII platforms.
368
Dan F (22 January 2007)
369
- Converted (most of) the test data files into genuine XML. A handful still
370
are not, due mainly to the lack of support for XML character entities
371
(e.g. & => & ). This will make it easier to validate test files using
372
tools like xmllint, as well as to edit and view them using XML tools.
374
Daniel (16 January 2007)
375
- Armel Asselin improved libcurl to behave a lot better when an easy handle
376
doing an FTP transfer is removed from a multi handle before completion. The
377
fix also fixed the "alive counter" to be correct on "premature removal" for
380
Dan F (16 January 2007)
381
- Fixed a small memory leak in tftp uploads discovered by curl's memory leak
382
detector. Also changed tftp downloads to URL-unescape the downloaded
385
Daniel (14 January 2007)
386
- David McCreedy provided libcurl changes for doing HTTP communication on
387
non-ASCII platforms. It does add some complexity, most notably with more
388
#ifdefs, but I want to see this supported added and I can't see how we can
389
add it without the extra stuff added.
391
- Setting CURLOPT_COOKIELIST to "ALL" when no cookies at all was present,
392
libcurl would crash when trying to read a NULL pointer.
394
Daniel (12 January 2007)
395
- Toby Peterson found a nasty bug that prevented (lib)curl from properly
396
downloading (most) things that were larger than 4GB on 32 bit systems. Matt
397
Witherspoon helped as narrow down the problem.
399
Daniel (5 January 2007)
400
- Linus Nielsen Feltzing introduced the --ftp-ssl-ccc command line option to
401
curl that uses the new CURLOPT_FTP_SSL_CCC option in libcurl. If enabled, it
402
will make libcurl shutdown SSL/TLS after the authentication is done on a
405
Daniel (4 January 2007)
406
- David McCreedy made changes to allow base64 encoding/decoding to work on
409
Daniel (3 January 2007)
410
- Matt Witherspoon fixed the flaw which made libcurl 7.16.0 always store
411
downloaded data in two buffers, just to be able to deal with a special HTTP
412
pipelining case. That is now only activated for pipelined transfers. In
413
Matt's case, it showed as a considerable performance difference,
415
Daniel (2 January 2007)
416
- Victor Snezhko helped us fix bug report #1603712
417
(http://curl.haxx.se/bug/view.cgi?id=1603712) (known bug #36) --limit-rate
418
(CURLOPT_MAX_SEND_SPEED_LARGE and CURLOPT_MAX_RECV_SPEED_LARGE) are broken
419
on Windows (since 7.16.0, but that's when they were introduced as previous
420
to that the limiting logic was made in the application only and not in the
421
library). It was actually also broken on select()-based systems (as apposed
422
to poll()) but we haven't had any such reports. We now use select(), Sleep()
423
or delay() properly to sleep a while without waiting for anything input or
424
output when the rate limiting is activated with the easy interface.
426
- Modified libcurl.pc.in to use Libs.private for the libs libcurl itself needs
427
to get built static. It has been mentioned before and was again brought to
428
our attention by Nathanael Nerode who filed debian bug report #405226
429
(http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=405226).
431
Daniel (29 December 2006)
432
- Make curl_easy_duphandle() set the magic number in the new handle.
434
Daniel (22 December 2006)
435
- Robert Foreman provided a prime example snippet showing how libcurl would
436
get confused and not acknowledge the 'no_proxy' variable properly once it
437
had used the proxy and you re-used the same easy handle. I made sure the
438
proxy name is properly stored in the connect struct rather than the
439
sessionhandle/easy struct.
441
- David McCreedy fixed a bad call to getsockname() that wrongly used a size_t
442
variable to point to when it should be a socklen_t.
444
- When setting a proxy with environment variables and (for example) running
445
'curl [URL]' with a URL without a protocol prefix, curl would not send a
446
correct request as it failed to add the protocol prefix.
448
Daniel (21 December 2006)
449
- Robson Braga Araujo reported bug #1618359
450
(http://curl.haxx.se/bug/view.cgi?id=1618359) and subsequently provided a
451
patch for it: when downloading 2 zero byte files in a row, curl 7.16.0
452
enters an infinite loop, while curl 7.16.1-20061218 does one additional
455
Fix: During the "Major overhaul introducing http pipelining support and
456
shared connection cache within the multi handle." change, headerbytecount
457
was moved to live in the Curl_transfer_keeper structure. But that structure
458
is reset in the Transfer method, losing the information that we had about
459
the header size. This patch moves it back to the connectdata struct.
461
Daniel (16 December 2006)
462
- Brendan Jurd provided a fix that now prevents libcurl from getting a SIGPIPE
463
during certain conditions when GnuTLS is used.
465
Daniel (11 December 2006)
466
- Alexey Simak found out that when doing FTP with the multi interface and
467
something went wrong like it got a bad response code back from the server,
468
libcurl would leak memory. Added test case 538 to verify the fix.
470
I also noted that the connection would get cached in that case, which
471
doesn't make sense since it cannot be re-use when the authentication has
472
failed. I fixed that issue too at the same time, and also that the path
473
would be "remembered" in vain for cases where the connection was about to
476
Daniel (6 December 2006)
477
- Sebastien Willemijns reported bug #1603712
478
(http://curl.haxx.se/bug/view.cgi?id=1603712) which is about connections
479
getting cut off prematurely when --limit-rate is used. While I found no such
480
problems in my tests nor in my reading of the code, I found that the
481
--limit-rate code was severly flawed (since it was moved into the lib, since
482
7.15.5) when used with the easy interface and it didn't work as documented
483
so I reworked it somewhat and now it works for my tests.
485
Daniel (5 December 2006)
486
- Stefan Krause pointed out a compiler warning with a picky MSCV compiler when
487
passing a curl_off_t argument to the Curl_read_rewind() function which takes
488
an size_t argument. Curl_read_rewind() also had debug code left in it and it
489
was put in a different source file with no good reason when only used from
492
- Sh Diao reported that CURLOPT_CLOSEPOLICY doesn't work, and indeed, there is
493
no code present in the library that receives the option. Since it was not
494
possible to use, we know that no current users exist and thus we simply
495
removed it from the docs and made the code always use the default path of
498
- Jared Lundell filed bug report #1604956
499
(http://curl.haxx.se/bug/view.cgi?id=1604956) which identified setting
500
CURLOPT_MAXCONNECTS to zero caused libcurl to SIGSEGV. Starting now, libcurl
501
will always internally use no less than 1 entry in the connection cache.
503
- Sh Diao reported that CURLOPT_FORBID_REUSE no works, and indeed it broke in
506
- Martin Skinner brought back bug report #1230118 to haunt us once again.
507
(http://curl.haxx.se/bug/view.cgi?id=1230118) curl_getdate() did not work
508
properly for all input dates on Windows. It was mostly seen on some TZ time
509
zones using DST. Luckily, Martin also provided a fix.
511
- Alexey Simak filed bug report #1600447
512
(http://curl.haxx.se/bug/view.cgi?id=1600447) in which he noted that active
513
FTP connections don't work with the multi interface. The problem is here
514
that the multi interface state machine has a state during which it can wait
515
for the data connection to connect, but the active connection is not done in
516
the same step in the sequence as the passive one is so it doesn't quite work
517
for active. The active FTP code still use a blocking function to allow the
518
remote server to connect.
520
The fix (work-around is a better word) for this problem is to set the
521
boolean prematurely that the data connection is completed, so that the "wait
522
for connect" phase ends at once.
524
The proper fix, left for the future, is of course to make the active FTP
525
case to act in a non-blocking way too.
527
- Matt Witherspoon fixed a problem case when the CPU load went to 100% when a
528
HTTP upload was disconnected:
530
"What appears to be happening is that my system (Linux 2.6.17 and 2.6.13) is
531
setting *only* POLLHUP on poll() when the conditions in my previous mail
532
occur. As you can see, select.c:Curl_select() does not check for POLLHUP. So
533
basically what was happening, is poll() was returning immediately (with
534
POLLHUP set), but when Curl_select() looked at the bits, neither POLLERR or
535
POLLOUT was set. This still caused Curl_readwrite() to be called, which
536
quickly returned. Then the transfer() loop kept continuing at full speed
539
Daniel (1 December 2006)
540
- Toon Verwaest reported that there are servers that send the Content-Range:
541
header in a third, not suppported by libcurl, format and we agreed that we
542
could make the parser more forgiving to accept all the three found
545
Daniel (25 November 2006)
546
- Venkat Akella found out that libcurl did not like HTTP responses that simply
547
responded with a single status line and no headers nor body. Starting now, a
548
HTTP response on a persistent connection (i.e not set to be closed after the
549
response has been taken care of) must have Content-Length or chunked
550
encoding set, or libcurl will simply assume that there is no body.
552
To my horror I learned that we had no less than 57(!) test cases that did bad
553
HTTP responses like this, and even the test http server (sws) responded badly
554
when queried by the test system if it is the test system. So although the
555
actual fix for the problem was tiny, going through all the newly failing test
556
cases got really painful and boring.
558
Daniel (24 November 2006)
559
- James Housley did lots of work and introduced SFTP downloads.
561
Daniel (13 November 2006)
562
- Ron in bug #1595348 (http://curl.haxx.se/bug/view.cgi?id=1595348) pointed
563
out a stack overwrite (and the corresponding fix) on 64bit Windows when
564
dealing with HTTP chunked encoding.
566
Daniel (9 November 2006)
567
- Nir Soffer updated libcurl.framework.make:
568
o fix symlinks, should link to Versions, not to ./Versions
569
o indentation improvments
571
- Dmitriy Sergeyev found a SIGSEGV with his test04.c example posted on 7 Nov
572
2006. It turned out we wrongly assumed that the connection cache was present
573
when tearing down a connection.
575
- Ciprian Badescu found a SIGSEGV when doing multiple TFTP transfers using the
576
multi interface, but I could also repeat it doing multiple sequential ones
577
with the easy interface. Using Ciprian's test case, I could fix it.
579
Daniel (8 November 2006)
580
- Bradford Bruce reported that when setting CURLOPT_DEBUGFUNCTION without
581
CURLOPT_VERBOSE set to non-zero, you still got a few debug messages from the
582
SSL handshake. This is now stopped.
584
Daniel (7 November 2006)
585
- Olaf fixed a leftover problem with the CONNECT fix of his that would leave a
586
wrong error message in the error message buffer.
588
Daniel (3 November 2006)
589
- Olaf Stueben provided a patch that I edited slightly. It fixes the notorious
590
KNOWN_BUGS #25, which happens when a proxy closes the connection when
591
libcurl has sent CONNECT, as part of an authentication negotiation. Starting
592
now, libcurl will re-connect accordingly and continue the authentication as
595
Daniel (2 November 2006)
596
- James Housley brought support for SCP transfers, based on the libssh2 library
597
for the actual network protocol stuff.
599
Added these new curl_easy_setopt() options:
601
CURLOPT_SSH_AUTH_TYPES
602
CURLOPT_SSH_PUBLIC_KEYFILE
603
CURLOPT_SSH_PRIVATE_KEYFILE
605
Version 7.16.0 (30 October 2006)
607
Daniel (25 October 2006)
608
- Fixed CURLOPT_FAILONERROR to return CURLE_HTTP_RETURNED_ERROR even for the
609
case when 401 or 407 are returned, *IF* no auth credentials have been given.
610
The CURLOPT_FAILONERROR option is not possible to make fool-proof for 401
611
and 407 cases when auth credentials is given, but we've now covered this
614
You might get some amounts of headers transferred before this situation is
615
detected, like for when a "100-continue" is received as a response to a
616
POST/PUT and a 401 or 407 is received immediately afterwards.
618
Added test 281 to verify this change.
620
Daniel (23 October 2006)
621
- Ravi Pratap provided a major update with pipelining fixes. We also no longer
622
re-use connections (for pipelining) before the name resolving is done.
624
Daniel (21 October 2006)
625
- Nir Soffer made the tests/libtest/Makefile.am use a proper variable for all
626
the single test applications' link and dependences, so that you easier can
627
override those from the command line when using make.
629
- Armel Asselin separated CA cert verification problems from problems with
630
reading the (local) CA cert file to let users easier pinpoint the actual
631
problem. CURLE_SSL_CACERT_BADFILE (77) is the new libcurl error code.
633
Daniel (18 October 2006)
634
- Removed the "protocol-guessing" for URLs with host names starting with FTPS
635
or TELNET since they are practically non-existant. This leaves us with only
636
three different prefixes that would assume the protocol is anything but
637
HTTP, and they are host names starting with "ftp.", "dict." or "ldap.".
639
Daniel (17 October 2006)
640
- Bug report #1579171 pointed out code flaws detected with "prefast", and they
641
were 1 - a too small memory clear with memset() in the threaded resolver and
642
2 - a range of potentially bad uses of the ctype family of is*() functions
643
such as isdigit(), isalnum(), isprint() and more. The latter made me switch
644
to using our own set of these functions/macros using uppercase letters, and
645
with some extra set of crazy typecasts to avoid mistakingly passing in
646
negative numbers to the underlying is*() functions.
648
- With Jeff Pohlmeyer's help, I fixed the expire timer when using
649
curl_multi_socket() during name resolves with c-ares and the LOW_SPEED
650
options now work fine with curl_multi_socket() as well.
652
Daniel (16 October 2006)
653
- Added a check in configure that simply tries to run a program (not when
654
cross-compiling) in order to detect problems with run-time libraries that
655
otherwise would occur when the sizeof tests for curl_off_t would run and
656
thus be much more confusing to users. The check of course should run after
657
all lib-checks are done and before any other test is used that would run an
658
executable built for testing-purposes.
660
Dan F (13 October 2006)
661
- The tagging of application/x-www-form-urlencoded POST body data sent
662
to the CURLOPT_DEBUGFUNCTION callback has been fixed (it was erroneously
663
included as part of the header). A message was also added to the
664
command line tool to show when data is being sent, enabled when
667
Daniel (12 October 2006)
668
- Starting now, adding an easy handle to a multi stack that was already added
669
to a multi stack will cause CURLM_BAD_EASY_HANDLE to get returned.
671
- Jeff Pohlmeyer has been working with the hiperfifo.c example source code,
672
and while doing so it became apparent that the current timeout system for
673
the socket API really was a bit awkward since it become quite some work to
674
be sure we have the correct timeout set.
676
Jeff then provided the new CURLMOPT_TIMERFUNCTION that is yet another
677
callback the app can set to get to know when the general timeout time
678
changes and thus for an application like hiperfifo.c it makes everything a
679
lot easier and nicer. There's a CURLMOPT_TIMERDATA option too of course in
680
good old libcurl tradition.
682
Jeff has also updated the hiperfifo.c example code to use this news.
684
Daniel (9 October 2006)
685
- Bogdan Nicula's second test case (posted Sun, 08 Oct 2006) converted to test
686
case 535 and it now runs fine. Again a problem with the pipelining code not
687
taking all possible (error) conditions into account.
689
Daniel (6 October 2006)
690
- Bogdan Nicula's hanging test case (posted Wed, 04 Oct 2006) was converted to
691
test case 533 and the test now runs fine.
693
Daniel (4 October 2006)
694
- Dmitriy Sergeyev provided an example source code that crashed CVS libcurl
695
but that worked nicely in 7.15.5. I converted it into test case 532 and
698
Daniel (29 September 2006)
699
- Removed a few other no-longer present options from the header file.
701
- Support for FTP third party transfers was removed. Here's why:
703
o The recent multi interface changes broke it and the design of the 3rd party
704
transfers made it very hard to fix the problems
705
o It was still blocking and thus nasty for the multi interface
706
o It was a lot of extra code for a very rarely used feature
707
o It didn't use the same code as for "plain" FTP transfers, so it didn't work
708
fine for IPv6 and it didn't properly re-use connections and more
709
o There's nobody around who's willing to work on and improve the existing
712
This does not mean that third party transfers are banned forever, only that
713
they need to be done better if they are to be re-added in the future.
715
The CURLOPT_SOURCE_* options are removed from the lib and so are the --3p*
716
options from the command line tool. For this reason, I also bumped the
717
version info for the lib.
719
Daniel (28 September 2006)
720
- Reported in #1561470 (http://curl.haxx.se/bug/view.cgi?id=1561470), libcurl
721
would crash if a bad function sequence was used when shutting down after
722
using the multi interface (i.e using easy_cleanup after multi_cleanup) so
723
precautions have been added to make sure it doesn't any more - test case 529
726
Daniel (27 September 2006)
727
- The URL in the cookie jar file is now changed since it was giving a 404.
728
Reported by Timothy Stone. The new URL will take the visitor to a curl web
729
site mirror with the document.
731
Daniel (24 September 2006)
732
- Bernard Leak fixed configure --with-gssapi-libs.
734
- Cory Nelson made libcurl use the WSAPoll() function if built for Windows
735
Vista (_WIN32_WINNT >= 0x0600)
737
Daniel (23 September 2006)
738
- Mike Protts added --ftp-ssl-control to make curl use FTP-SSL, but only
739
encrypt the control connection and use the data connection "plain".
741
- Dmitriy Sergeyev provided a patch that made the SOCKS[45] code work better
742
as it now will read the full data sent from servers. The SOCKS-related code
743
was also moved to the new lib/socks.c source file.
745
Daniel (21 September 2006)
746
- Added test case 531 in an attempt to repeat bug report #1561470
747
(http://curl.haxx.se/bug/view.cgi?id=1561470) that is said to crash when an
748
FTP upload fails with the multi interface. It did not, but I made a failed
749
upload still assume the control connection to be fine.
751
Daniel (20 September 2006)
752
- Armel Asselin fixed problems when you gave a proxy URL with user name and
753
empty password or no password at all. Test case 278 and 279 were added to
756
Daniel (12 September 2006)
757
- Added docs/examples/10-at-a-time.c by Michael Wallner
759
- Added docs/examples/hiperfifo.c by Jeff Pohlmeyer
761
Daniel (11 September 2006)
762
- Fixed my breakage from earlier today so that doing curl_easy_cleanup() on a
763
handle that is part of a multi handle first removes the handle from the
766
- Added CURLOPT_SSL_SESSIONID_CACHE and --no-sessionid to disable SSL
767
session-ID re-use on demand since there obviously are broken servers out
768
there that misbehave with session-IDs used.
770
- Jeff Pohlmeyer presented a *multi_socket()-using program that exposed a
771
problem with it (SIGSEGV-style). It clearly showed that the existing
772
socket-state and state-difference function wasn't good enough so I rewrote
773
it and could then re-run Jeff's program without any crash. The previous
774
version clearly could miss to tell the application when a handle changed
775
from using one socket to using another.
777
While I was at it (as I could use this as a means to track this problem
778
down), I've now added a 'magic' number to the easy handle struct that is
779
inited at curl_easy_init() time and cleared at curl_easy_cleanup() time that
780
we can use internally to detect that an easy handle seems to be fine, or at
781
least not closed or freed (freeing in debug builds fill the area with 0x13
782
bytes but in normal builds we can of course not assume any particular data
785
Daniel (9 September 2006)
786
- Michele Bini fixed how the hostname is put in NTLM packages. As servers
787
don't expect fully qualified names we need to cut them off at the first dot.
789
- Peter Sylvester cleaned up and fixed the getsockname() uses in ftp.c. Some
790
of them can be completetly removed though...
792
Daniel (6 September 2006)
793
- Ravi Pratap and I have implemented HTTP Pipelining support. Enable it for a
794
multi handle using CURLMOPT_PIPELINING and all HTTP connections done on that
795
handle will be attempted to get pipelined instead of done in parallell as
796
they are performed otherwise.
798
As a side-effect from this work, connections are now shared between all easy
799
handles within a multi handle, so if you use N easy handles for transfers,
800
each of them can pick up and re-use a connection that was previously used by
801
any of the handles, be it the same or one of the others.
803
This separation of the tight relationship between connections and easy
804
handles is most noticable when you close easy handles that have been used in
805
a multi handle and check amount of used memory or watch the debug output, as
806
there are times when libcurl will keep the easy handle around for a while
807
longer to be able to close it properly. Like for sending QUIT to close down
810
This is a major change.
812
Daniel (4 September 2006)
813
- Dmitry Rechkin (http://curl.haxx.se/bug/view.cgi?id=1551412) provided a
814
patch that while not fixing things very nicely, it does make the SOCKS5
815
proxy connection slightly better as it now acknowledges the timeout for
816
connection and it no longer segfaults in the case when SOCKS requires
817
authentication and you did not specify username:password.
819
Daniel (31 August 2006)
820
- Dmitriy Sergeyev found and fixed a multi interface flaw when using asynch
821
name resolves. It could get stuck in the wrong state.
823
Gisle (29 August 2006)
824
- Added support for other MS-DOS compilers (desides djgpp). All MS-DOS
825
compiler now uses the same config.dos file (renamed to config.h by
826
make). libcurl now builds fine using Watcom and Metaware's High-C
827
using the Watt-32 tcp/ip-stack.
829
Daniel (29 August 2006)
830
- David McCreedy added CURLOPT_SOCKOPTFUNCTION and CURLOPT_SOCKOPTDATA to
831
allow applications to set their own socket options.
833
Daniel (25 August 2006)
834
- Armel Asselin reported that the 'running_handles' counter wasn't updated
835
properly if you removed a "live" handle from a multi handle with
836
curl_multi_remove_handle().
838
Daniel (22 August 2006)
839
- David McCreedy fixed a remaining mistake from the August 19 TYPE change.
841
- Peter Sylvester pointed out a flaw in the AllowServerConnect() in the FTP
842
code when doing pure ipv6 EPRT connections.
844
Daniel (19 August 2006)
845
- Based on a patch by Armel Asselin, the FTP code no longer re-issues the TYPE
846
command on subsequent requests on a re-used connection unless it has to.
848
- Armel Asselin fixed a crash in the FTP code when using SINGLECWD mode and
849
files in the root directory.
851
- Andrew Biggs pointed out a "Expect: 100-continue" flaw where libcurl didn't
852
send the whole request at once, even though the Expect: header was disabled
853
by the application. An effect of this change is also that small (< 1024
854
bytes) POSTs are now always sent without Expect: header since we deem it
855
more costly to bother about that than the risk that we send the data in
858
Daniel (9 August 2006)
859
- Armel Asselin made the CURLOPT_PREQUOTE option work fine even when
860
CURLOPT_NOBODY is set true. PREQUOTE is then run roughly at the same place
861
in the command sequence as it would have run if there would've been a
864
Daniel (8 August 2006)
865
- Fixed a flaw in the "Expect: 100-continue" treatment. If you did two POSTs
866
on a persistent connection and allowed the first to use that header, you
867
could not disable it for the second request.
869
Daniel (7 August 2006)
870
- Domenico Andreolfound a quick build error which happened because
871
src/config.h.in was not a proper duplcate of lib/config.h.in which it
872
should've been and this was due to the maketgz script not doing the cp
9
875
Version 7.15.5 (7 August 2006)
11
877
Daniel (2 August 2006)
added
removed
CHANGES
