9
Version 7.21.2 (12 Oct 2010)
11
Daniel Stenberg (12 Oct 2010)
12
- RELEASE-NOTES: synced with ecd624b8e774a85
14
- [Julien Chaffraix brought this change]
18
Do not match the trailing '\n' in the regular expression as this would
19
make us dump a ) parenthesis on a new line.
21
This fixes the following error:
23
would get transformed into:
27
Bug: http://curl.haxx.se/mail/lib-2010-10/0065.html
28
Reported by: Dimitre Dimitrov
30
- header_callback: strip off file path separated with backslashes
32
If the filename contains a backslash, only use filename portion. The
33
idea is that even systems that don't handle backslashes as path
34
separators probably want that path removed for convenience.
36
This flaw is considered a security problem, see the curl security
37
vulnerability http://curl.haxx.se/docs/adv_20101013.html
39
Dan Fandrich (12 Oct 2010)
40
- Get the curl source files for Amiga from Makefile.inc
42
This is similar to how it's done in the lib directory.
43
The Amiga build appears to have been broken for a year because
44
of a missing homedir.c
46
- Added section on server-supplied names to security considerations
48
Guenter Knauf (12 Oct 2010)
49
- Fixed Watcom makefile.
51
- Added build bits for librtmp / libssh2 to Watcom makefiles.
53
- Added build bits for librtmp to NetWare makefiles.
55
Daniel Stenberg (12 Oct 2010)
56
- SFTP: more ignoring negative file sizes
58
As the change in 5f0ae7a0626cbe709 added a precaution against negative
59
file sizes that for some reason managed to get returned, this change now
60
introduces the same check at the second place in the code where the file
61
size from the libssh2 stat call is used.
63
This check might not be suitable for a 32 bit curl_off_t, but libssh2.h
64
assumes long long to work and to be 64 bit so I believe such a small
65
curl_off_t will be very unlikely to occur in the wild.
67
- SMTP: debug output for no known auth mechanisms supported
69
... and some minor source code whitespace edits
71
- test: urlglob error messages have no extra newline anymore
73
Guenter Knauf (11 Oct 2010)
74
- Added build bits for librtmp to MingW32 makefiles.
76
Daniel Stenberg (8 Oct 2010)
77
- RELEASE-NOTES: synced with 61f4cdb73ae4
79
- globbing: fix crash on unballanced open brace
81
Having an open brace without a closing brace caused a segfault.
83
Having a closing brace too many caused a silent error to occur, which
84
caused curl to bail out and return an error code but no error message
85
was shown. It does now!
87
All error message outputs no longer wrongly get _two_ newlines written
88
after the error message.
90
Reported by: Vlad Ureche
91
Bug: http://curl.haxx.se/bug/view.cgi?id=3083942
93
- [Dan Locks brought this change]
95
libcurl.m4: AC_PATH_PROG fixes
97
The invocation of autoconf's AC_PATH_PROG( ) is not quite right for
98
finding curl-config. This fix corrects the negative case (where
99
curl-config is not found).
101
- FAQ: added "How do I submit my patch?"
103
- examples: use example.com in example URLs
105
- TODO-RELEASE: libidn problem not repeatable
107
"261 - configure and libidn" is removed from the list since Julien
108
Chaffraix tried to repeat it but failed and the reporter did not return
109
to provide further details.
111
Reported by: Lyndon Hill
112
Bug: http://curl.haxx.se/mail/lib-2010-07/0029.html
114
- libcurl.m4: mention argument is PREFIX
116
The macro provides a --with-libcurl option that expects a PREFIX to be
117
specified and not actually a "directory" in which libcurl will be found.
118
This now spells that out more clearly.
120
Reported by: Dan Locks
121
Bug: http://curl.haxx.se/bug/view.cgi?id=3079891
123
Guenter Knauf (3 Oct 2010)
124
- Some NetWare makefile tweaks.
126
Renamed SDK_* to NDK_*; made NDK_* defines overwriteable from
127
environment; removed now obsolete YACC macro;
128
moved some curl_config.h defines to IPv6 section since they
129
are only needed when IPv6 is enabled - this makes libcurl compile
130
with older NDKs too which were not IPv6-aware.
132
Daniel Stenberg (2 Oct 2010)
133
- TODO-RELEASE: 416 error fixed
135
"3076808 Requests fail silently following a 416 error" done
137
Julien Chaffraix (2 Oct 2010)
138
- krb5-gssapi: Removed a memory leak in krb5_auth.
140
We forgot to release the buffer passed to gss_init_sec_context.
142
The previous logic was difficult to read as we were reusing the same
143
variable (gssbuf) for both input buffer and output buffer. Splitted the
144
logic in 2 variables to better underline who needs to be released.
145
Also made the code break at 80 lines.
147
- krb5-gssapi: Made the function always return a value.
149
kr5_auth missed a final 'return' statement. This is not an error in
150
gcc but can lead to potential bugs.
152
- krb5-gssapi: Delete the GSS-API context.
154
This fixes a memory leak related to the GSS-API code.
156
Added a krb5_init and krb5_end functions. Also removed a work-around
157
the lack of proper initialization of the GSS-API context.
159
Daniel Stenberg (2 Oct 2010)
160
- HTTP: remove special case for 416
162
It was pointed out that the special case libcurl did for 416 was
163
incorrect and wrong. 416 is not really different to other errors so the
164
response body must be handled like for other errors/http responses.
166
Reported by: Chris Smowton
167
Bug: http://curl.haxx.se/bug/view.cgi?id=3076808
169
- [Dan Fandrich brought this change]
171
sws: Added writedelay HTTP server command
173
This delays between write operations, hopefully making it easier
174
to spot problems where libcurl doesn't flush the socket properly
175
before waiting for the next response.
177
- TODO-RELEASE: no bug in ftp_nextconnect
179
The issue named "266 - Bug in ftp_nextconnect?" was deemed to not be a
180
bug and instead resulted in clarified docs.
182
- curl_easy_setopt.3: CURLOPT_DIRLISTONLY implies dir list
184
Make it explicit that setting CURLOPT_DIRLISTONLY to 1 will make libcurl
185
to list the directory.
187
- RELEASE-NOTES: synced up to 588402585bae
189
- TODO-RELEASE: move new features to next release
191
- README.ares: we know require c-ares 1.6.0
193
- SFTP: avoid downloading negative sizes!
195
It is still not clarified exactly why this happens, but libssh2
196
sometimes report a negative file size for the remote SFTP file and that
197
deeply confuses libcurl (or crashes it) so this precaution is added to
200
Reported by: Ernest Beinrohr
201
Bug: http://curl.haxx.se/bug/view.cgi?id=3076430
203
- TODO-RELEASE: drop curl_easy_setoptv
205
I haven't read any really convincing arguments for adding it
207
- [Dirk Manske brought this change]
209
multi & hiper examples: updates and cleanups
211
all multi and hiper examples:
213
* don't loop curl_multi_perform calls, that was <7.20.0 style, currently
214
the exported multi functions will not return CURLM_CALL_MULTI_PERFORM
217
* renamed check_run_count to check_multi_info
218
* don't compare current running handle count with previous value, this
219
was the wrong way to check for finished requests, simply call
221
* it's also safe to call curl_multi_remove_handle inside the
222
curl_multi_info_read loop.
225
* replaced curl_multi_socket (that function is marked as obsolete) calls
226
with curl_multi_socket_action calls (as in hiperfifo.c and
229
ghiper.c and evhiperfifo.c:
230
* be smart as hiperfifo.c, don't do uncessary curl_multi_* calls in
233
- TODO-RELEASE: one fixed, one postponed, one added
235
As we're already in feature freeze, I pushed the feature onwards.
237
Dan Fandrich (29 Sep 2010)
238
- Renamed test1204 to test1117 to move it into the normal range
240
Patrick Monnerat (29 Sep 2010)
241
- Add gopher protocol definition to ILE/RPG binding.
242
OS400 compile script in test dir updated for chkhostname.
244
Julien Chaffraix (28 Sep 2010)
245
- krb5-gssapi: Remove several memory leaks.
247
Remove a leak seen on Kerberos/MIT (gss_OID is copied internally and
248
we were leaking it). Now we just pass NULL as advised in RFC2744.
250
|tmp| was never set back to buf->data.
252
Cleaned up Curl_sec_end to take into account failure in Curl_sec_login
253
(where conn->mech would be NULL but not conn->app_data or
254
conn->in_buffer->data).
256
- security.c: Remove Curl_sec_fflush_fd.
258
The current implementation would make us send wrong data on a closed
259
socket. We don't buffer our data so the method can be safely removed.
261
- security.c: We should always register the socket handler.
263
Following a change in the way socket handler are registered, the custom
264
recv and send method were conditionaly registered.
265
We need to register them everytime to handle the ftp security
268
Re-added the clear text handling in sec_recv.
270
- security.c: Fix Curl_sec_login after rewrite.
272
Curl_sec_login was returning the opposite result that the code in ftp.c
273
was expecting. Simplified the return code (using a CURLcode) so to see
274
more clearly what is going on.
276
- security.c: Readd the '\n' to the infof() calls.
278
They are not automatically added and make the output of the verbose
279
mode a lot more readable.
281
- security.c: Fix typo (PSBZ -> PBSZ)
283
- security.c: Fix ftp_send_command.
285
My use of va_args was completely wrong. Fixed the usage so that
286
we send the right commands!
288
Daniel Stenberg (28 Sep 2010)
289
- curl_easy_escape: don't escape "unreserved" characters
291
According to RFC3986 section 2.3 the letters -, ., _ and ~ should not be
294
Reported by: Miguel Diaz
295
Bug: http://curl.haxx.se/mail/lib-2010-09/0227.html
297
- multi: don't expire timeouts at disonnect or done
299
The functions Curl_disconnect() and Curl_done() are both used within the
300
scope of a single request so they cannot be allowed to use
301
Curl_expire(... 0) to kill all timeouts as there are some timeouts that
302
are set before a request that are supposed to remain until the request
305
The timeouts are now instead cleared at curl_easy_cleanup() and when the
306
multi state machine changes a handle to the complete state.
308
Dan Fandrich (27 Sep 2010)
309
- Changed the TPF make file to get source files from Makefile.inc
311
Patch was fixed and validated by David McCreedy.
313
- Added test case 1204 to test HTTP range failure
315
This is an attempt to reproduce bug #3076808
317
Daniel Stenberg (27 Sep 2010)
318
- [Dirk Manske brought this change]
320
multi_runsingle: set timeout error messages
322
With the latest changes to fix the timeout handling with multi interface
323
we lost the timeout error messages. This patch brings them back.
325
- TODO-RELEASE: updated list of issues to work on
327
- parsedate: allow time specified without seconds
329
The date format in RFC822 allows that the seconds part of HH:MM:SS is
330
left out, but this function didn't allow it. This change also includes a
331
modified test case that makes sure that this now works.
333
Reported by: Matt Ford
334
Bug: http://curl.haxx.se/bug/view.cgi?id=3076529
336
- TFTP: re-indented the source code
338
Just made sure that the good old curl indentation style is used all over
341
- [Tim Newsome brought this change]
343
TFTP: Work around tftpd-hpa upload bug
345
tftpd-hpa has a bug where it will send an incorrect ack when the block
346
counter wraps and tftp options have been sent. Work around that by
347
accepting an ack for 65535 when we're expecting one for 0.
349
- Revert "security.c: buffer_read various fixes."
351
This reverts commit fbb38de415b7bb7d743e53a7b4b887ffb12b3e5b.
353
- security.c: removed superfluous parentheses
355
And also removed the FIXME where memory was zeroed just before freed,
356
and some other minor whitespace changes.
358
- [Julien Chaffraix brought this change]
360
security.c: Update the #include statements after the rewrite.
362
- [Julien Chaffraix brought this change]
364
security.c: sec_write tweaks
366
- |fd| is now a curl_socket_t and |len| a size_t to avoid conversions.
367
- Added 2 FIXMEs about the 2 unsigned -> signed conversions.
368
- Included 2 minor changes to Curl_sec_end.
370
- [Julien Chaffraix brought this change]
372
security.c: _sec_send tweaks
374
- Renamed the method to sec_send now that we
375
renamed sec_send to do_sec_send.
376
- Some more variable renaming.
378
- [Julien Chaffraix brought this change]
380
security.c: sec_read tweaks
382
- Renamed the function to sec_recv.
383
- Renamed the parameters and variable to match the rest of the code.
385
- [Julien Chaffraix brought this change]
387
security.c: Curl_sec_fflush_fd tweaks
389
- Use an early return as it makes the code more readable.
390
- Added a FIXME about a conversion.
392
- [Julien Chaffraix brought this change]
394
security.c: sec_send tweaks
396
- Renamed it to do_sec_send as it is the function doing the actual
398
- Do not return any values as no one was checking it and it never
399
reported a failure (added a FIXME about checking for errors).
400
- Renamed the variables to make their use more specific.
401
- Removed some casts (int -> curl_socket_t, ...)
402
- Avoid doing the htnl <-> nthl twice by caching the 2 results.
404
- [Julien Chaffraix brought this change]
406
security.c: Curl_sec_read_msg tweaks
408
- Renamed the variables name to better match their intend.
409
- Unified the |decoded_len| checks.
410
- Added some FIXMEs to flag some improvement that did not go in this
413
- [Julien Chaffraix brought this change]
415
security.c: Curl_sec_set_protection_level tweaking
417
- Removed sec_prot_internal as it is now inlined in the function (this removed
419
- Changed the prototype to return an error code.
420
- Updated the method to use the new ftp_send_command function.
421
- Added a level_to_char helper method to avoid relying on the compiler's
422
bound checks. This default to the maximum security we have in case of a
425
- [Julien Chaffraix brought this change]
427
security.c: factored the logic from Curl_sec_login into a dedicated method that better reflect its intent.
429
Introduced a helper method ftp_send_command that synchronously send
432
- [Julien Chaffraix brought this change]
434
security.c: Remove out_buffer as it was never written into.
436
- [Julien Chaffraix brought this change]
438
security.c: buffer_read various fixes.
440
Tighten the type of the |data| parameter to avoid a cast. Also made
441
it const as we should not modify it.
443
Added a DEBUGASSERT on the size to be written while changing it.
445
- [Julien Chaffraix brought this change]
447
security.c: Made block_write return a CURLcode.
449
While doing so, renamed it to socket_write to better match its
452
- [Julien Chaffraix brought this change]
454
security.c: Made block_read and sec_get_data return CURLcode.
456
To do so, made block_read call Curl_read_plain instead of read.
458
While changing them renamed block_read to socket_read and sec_get_data
459
to read_data to better match their function.
461
Also fixed a potential memory leak in block_read.
463
- [Julien Chaffraix brought this change]
465
Security.c: Fix headers guard to match the rest of the code.
467
- [Julien Chaffraix brought this change]
469
configure: Fix the LDAPS disable message
471
... for example when LDAP is not compiled.
473
Fixed the logic to match the rest of the options' message that is we
474
update the default message only if the option is not disabled after the
477
Reported by: Guenter Knauf
479
- RELEASE-NOTES: sync with 8665d4e5 and c-ares >= 1.6.0 note
481
- parse_remote_port: ignore colons without port number
483
Obviously, browsers ignore a colon without a following port number. Both
484
Firefox and Chrome just removes the colon for such URLs. This change
485
does not remove the colon for URLs sent over a HTTP proxy, so we should
486
consider doing that change as well.
488
Reported by: github user 'kreshano'
490
- RELEASE-NOTES: in sync with 19f45eaa799
492
- duphandle: use ares_dup()
494
curl_easy_duphandle() was not properly duping the ares channel. The
495
ares_dup() function was introduced in c-ares 1.6.0 so by starting to use
496
this function we also raise the bar and require c-ares >= 1.6.0
497
(released Dec 9, 2008) for such builds.
499
Reported by: Ning Dong
500
Bug: http://curl.haxx.se/mail/lib-2010-08/0318.html
502
- [Hendrik Visage brought this change]
504
MacOSX-Framework: updates for Snowleopard
506
1) PPC64 appears to be an 10.5 only supported architecture, so I
507
forced 10.5 for 64bit if there is a need for PPC64, else 64bit only
510
2) proper "make clean" after every ./configure. fixes a bug where
511
subsequent runs the 32bit do not get compiled
513
3) Added a version numbering curl-$VERSION} rather than the "stock standard" A
515
- RELEASE-NOTES: synced with 5fcc4332d62fe
517
Removed the duplicate entry of Kamil in the credits.
519
- configure: don't enable RTMP if the lib detect fails
521
librtmp is often statically linked and using sub dependencies like
522
OpenSSL, so we need to make sure we can actually link with it properly
523
before enabling it. Otherwise we easily end up trying to link with a
526
- TODO: added 8.4 non-gcrypt under GnuTLS
528
We must not assume gcrypt just because of GnuTLS
530
- configure: check for gcrypt if using GnuTLS
532
1 - libcurl assumes that there are gcrypt functions available when
535
2 - GnuTLS can be built to use libnettle instead as crypto library,
536
which breaks assumption (1)
538
This change makes configure make sure that if GnuTLS is requested and
539
detected, it also makes sure that gcrypt is present or it errors
540
out. This is mostly a way to make the user more aware of this flaw, the
541
correct fix would be to detect which crypto layer that is in use and
542
adapt our code to use that instead of blindly assuming gcrypt.
544
Reported by: Michal Gorny
545
Bug: http://curl.haxx.se/bug/view.cgi?id=3071038
547
- RELEASE-NOTES: sync from d2a7fd2fe65b to HEAD
549
- FTP: fix bad check of Curl_timeleft() return code
551
When it returns 0 it means no timeout. Only a negative value means that
554
- LDAP: moved variable declaration to avoid compiler warn
556
If built without HTTP or proxy support it would cause a compiler warning
557
due to the unused variable. I moved the declaration of it into the only
560
Tor Arntsen (18 Sep 2010)
561
- LDAP: Use FALSE instead of bool_false when setting bits.close
563
bool_false is the internal name used in the setup_once.h definition
564
we fall back to for non-C99 non-stdbool systems, it's not the actual
565
name to use in assignments (we use bool_false, bool_true there to
566
avoid global namespace problems, see comment in setup_once.h).
567
The correct C99 value to use is 'false', but let's use FALSE as
568
used elsewhere when assigning to bits.close. FALSE is set equal
569
to 'false' in setup_once.h when possible.
571
This fixes a build problem on C99 targets.
573
- LDAP: Add missing declaration for 'result'
575
Daniel Stenberg (18 Sep 2010)
576
- [Mauro Iorio brought this change]
578
LDAP: Support for tunnelling queries through HTTP proxy
580
As of curl-7.21.1 tunnelling ldap queries through HTTP Proxies is not
581
supported. Actually if --proxytunnel command-line option (or equivalent
582
CURLOPT_HTTPPROXYTUNNEL) is used for ldap queries like
583
ldap://ldap.my.server.com/... You are unable to successfully execute the
584
query. In facts ldap_*_bind is executed directly against the ldap server
585
and proxy is totally ignored. This is true for both openLDAP and
588
Step to reproduce the error:
589
Just launch "curl --proxytunnel --proxy 192.168.1.1:8080
590
ldap://ldap.my.server.com/dc=... "
592
This fix adds an invocation to Curl_proxyCONNECT against the provided
593
proxy address and on successful "CONNECT" it tunnels ldap query to the
594
final ldap server through the HTTP proxy. As far as I know Microsoft
595
LDAP APIs don't permit tunnelling in any way so the patch provided is
596
for OpenLDAP only. The patch has been developed against OpenLDAP 2.4.23
597
and has been tested with Microsoft ISA Server 2006 and works properly
598
with basic, digest and NTLM authentication.
600
- timeout: use the correct start value as offset
602
Rodric provide an awesome recipe that proved libcurl didn't timeout at
603
the requested time - it instead often timed out at [connect time] +
604
[timeout time] instead of the documented and intended [timeout time]
605
only. This bug was due to the code using the wrong base offset when
606
comparing against "now". I could also take the oppurtinity to simplify
607
the code by properly using of the generic help function for this:
610
Reported by: Rodric Glaser
611
Bug: http://curl.haxx.se/bug/view.cgi?id=3061535
613
- Curl_timeleft: avoid returning "no timeout" by mistake
615
As this function uses return code 0 to mean that there is no timeout, it
616
needs to check that it doesn't return a time left value that is exactly
617
zero. It could lead to libcurl doing an extra 1000 ms select() call and
618
thus not timing out as accurately as it should.
620
I fell over this bug when working on the bug 3061535 but this fix does
621
not correct that problem alone, although this is a problem that needs to
624
Reported by: Rodric Glaser
625
Bug: http://curl.haxx.se/bug/view.cgi?id=3061535
627
- whitespace: unified source
632
and some other changes in the similar spirit, trying to make the
633
whole file use the same style
635
- remote-header-name: don't output filename when NULL
637
- [James Bursa brought this change]
639
TheArtOfHttpScripting: use long options
641
- [James Bursa brought this change]
643
getinmemory: make the example easier to follow
645
1. Remove the comment warning that it's "not been verified to work". It
646
works with no problems in my testing.
648
2. Remove 2 unnecessary includes.
650
3. Remove the myrealloc(). Initialize chunk.memory with malloc() instead
651
of NULL. The comments for these two parts contradicted each other.
653
4. Handle out of memory from realloc() instead of continuing.
655
5. Print a brief status message at the end.
657
- multi: don't do extra expire calls for the connection
659
The timeout is set for the connect phase already at the start of the
660
request so we should not add a new one, and we MUST not set expire to 0
661
as that will remove any other potentially existing timeouts.
663
- [Peter Pentchev brought this change]
665
Fix a bashism: test a = b is more portable than ==.
667
- glob_word: remove a check that is always false
669
- inflate_stream: remove redundant check that is always true
671
- digest: make it clear the condition is always true
673
- ssluse: removed redundant check that is always true
675
Dan Fandrich (11 Sep 2010)
676
- Link curl and the test apps with -lrt explicitly when necessary
678
When curl calls a function from that library then it needs to
679
explicitly link to the library instead of piggybacking on
680
libcurl's own dependency. Without this, GNU ld with the
681
--no-add-needed flag fails when linking (which Fedora now does
684
Reported by: Quanah Gibson-Mount
685
Bug: http://curl.haxx.se/mail/lib-2010-09/0085.html
687
- Mention the Debian Popularity Contest
689
Tor Arntsen (9 Sep 2010)
690
- test565: Don't hardcode IP:PORT
692
Use %HOSTIP:%HTTPPORT instead of 127.0.0.1:8990 so that
693
verification works if the baseport change option is used
694
when executing runtests.pl.
696
Daniel Stenberg (9 Sep 2010)
697
- curl.1: updated protocols and polished language
699
- FAQ: CURL_STATICLIB for visual studio users
701
Clarified as it isn't used with a -D option for them.
703
Reported by: Artfunkel
704
Bug: http://curl.haxx.se/bug/view.cgi?id=3060381
706
- FAQ: updated and added host with custom IP question
708
Added "3.19 How do I get HTTP from a host using a specific IP address?"
709
and updated some stuff about certs etc.
711
- chunky parser: only rewind if needed
713
The code reading chunked encoding attempts to rewind the code if it had
714
read more data than the chunky parser consumes. The rewinding can fail
715
and it will then cause an error. This change now makes the rewinding
716
only happen if pipelining is in use - as that's the only time it really
719
Bug: http://curl.haxx.se/mail/lib-2010-08/0297.html
720
Reported by: Ron Parker
722
Kamil Dudka (6 Sep 2010)
723
- rtsp: avoid SIGSEGV on malformed header
725
- rtsp: avoid SIGSEGV on malformed header
727
Daniel Stenberg (6 Sep 2010)
728
- warning: fix conversion to 'int' from 'size_t'
730
- portabilty: use proper variable type to hold sockets
732
Curl_getconnectinfo() is changed to return a proper curl_socket_t for
733
the last socket so that it'll work more portably (and cause less
736
Guenter Knauf (3 Sep 2010)
737
- Trial to fix another compiler warning with braces.
739
Dan Fandrich (2 Sep 2010)
740
- Use checkprefix() to compare protocol-specific strings
742
Otherwise, there could be problems running in certain locales.
744
Guenter Knauf (2 Sep 2010)
745
- Moved S_ISREG define to setup as suggested by Dan.
747
- Use own typedef as workaround for broken sspi.h header (f.e. Watcom).
749
- Added some hacks in order to build with VC from git.
751
Adam Light posted this patch to the list which enables builds from
752
git with VC versions other than vc6; also he added a vc10 target.
754
- Added S_ISREG define for Win32.
756
Daniel Stenberg (1 Sep 2010)
757
- multi: fixes for timing out handles
759
Add a timeout check for handles in the state machine so that they will
760
timeout in all states disregarding what actions that may or may not
763
Fixed a bug in socket_action introduced recently when looping over timed
764
out handles: it wouldn't assign the 'data' variable and thus it wouldn't
765
properly take care of handles.
767
In the update_timer function, the code now checks if the timeout has
768
been removed and then it tells the application. Previously it would
769
always let the remaining timeout(s) just linger to expire later on.
771
- threaded resolver: no more expire 0 calls
773
Curl_expire() set to 0 expires ALL timeouts so it should only be called
774
if we truly and really want to remove all timeouts for the handle.
776
- resolve_server: simplify code
778
Make use of the helper function Curl_timeleft() instead of duplicating
781
- multi: make sure the next timeout is used when one expires
783
Each easy handle has a list of timeouts, so as soon as the main timeout
784
for a handle expires, we must make sure to get the next entry from the
785
list and re-add the handle to the splay tree.
787
This was attempted previously but was done poorly in my commit
790
Dan Fandrich (30 Aug 2010)
791
- Added proxy keyword to allow skipping test in proxyless configs
793
Daniel Stenberg (29 Aug 2010)
794
- multi: set timeouts when transfer begins
796
When a new transfer is about to start we now set the proper timeouts to
797
expire for the multi interface if they are set for the handle. This is a
798
follow-up bugfix to make sure that easy handles timeout properly when
799
the times expire and the multi interface is used. This also improves
800
curl_multi_timeout().
802
- CURLOPT_DIRLISTONLY: don't use with CURLOPT_WILDCARDMATCH
804
- FAQ: update list of supported protocols
806
- [Fabian Keil brought this change]
808
In the m4 detection line, factor out the 2>dev/null
810
- [Fabian Keil brought this change]
812
If m4 doesn't support --version, try if gm4 does.
814
- [Fabian Keil brought this change]
816
If the m4 version isn't recognized at all, just say so
818
'm4 version found. You need a GNU m4 installed!' is a bit confusing.
820
- HISTORY: mention the gopher story
822
Dan Fandrich (25 Aug 2010)
823
- Tweaked some test data files
825
Fixed some issues that caused xmllint failures, added features
826
and keywords, fixed some quotes and removed some <strip> sections
827
that unnecessarily limited test checking.
829
- Added new source files to Symbian and TPF makefiles
831
Daniel Stenberg (25 Aug 2010)
832
- RELEASE-NOTES: sync from b980c9a02 to HEAD
834
- Makefile: add gopher.c file to build
836
As the VC and RISCOS makefiles don't use the .inc file
838
- runtests: fix uninitialized variable warning
840
- gopher tests: revert parts of gopher in the pingpong server
842
Introduced in the initial gopher commits, there was added logic to do
843
GOPHER test serving in the pingpong server but as it resembles HTTP much
844
more than FTP or SMTP, the gopher testing has been moved over to instead
845
use the sws (HTTP) server. This change simply removes unused code.
847
- gopher tests: use sws and adjusted to more standard style
849
- sws: added basic gopher support
851
- gopher: enable the header callback/verbosity
853
- gopher: fix test case line endings
855
Patches over email very easily lose CRLF line endings in files otherwise
856
LF-only so I had to put them back where needed.
858
- gopher: fix memory leak and busyloop
860
The fix for the busyloop really only is a temporary work-around. It
861
causes a BLOCKING behavior which is a NO-NO. This function should rather
862
be split up in a do and a doing piece where the pieces that aren't
863
possible to send now will be sent in the doing function repeatedly until
864
the entire request is sent.
866
- [Cameron Kaiser brought this change]
868
Gopher using Curl_write; test suite (4 tests)
870
- [Cameron Kaiser brought this change]
874
- [Cameron Kaiser brought this change]
876
Forgot gopher.h in Makefile.inc
878
- [Cameron Kaiser brought this change]
880
Gopher protocol support (initial release)
882
- http: handle trailer headers in all chunked responses
884
HTTP allows that a server sends trailing headers after all the chunks
885
have been sent WITHOUT signalling their presence in the first response
886
headers. The "Trailer:" header is only a SHOULD there and as we need to
887
handle the situation even without that header I made libcurl ignore
890
Test case 1116 was added to verify this and to make sure we handle more
891
than one trailer header properly.
893
Reported by: Patrick McManus
894
Bug: http://curl.haxx.se/bug/view.cgi?id=3052450
896
- TODO: we now support RTMP
898
- TODO: done "NTLM with other crypto functions"
900
Since NTLM was made to work with the NSS API as well, the primary SSL
901
alternatives will be built with NTLM support in libcurl.
903
- TODO: fixed "Make curl_multi_info_read faster"
905
It is really fast now
907
Dan Fandrich (24 Aug 2010)
908
- Fixed a NULL pointer dereference in form posting
910
It was introduced in commit eeb2cb05 along with the -F type=
911
change. Also fixed a typo in the name of the magic filename=
912
parameter. Tweaked tests 39 and 173 to better test this path.
914
Daniel Stenberg (24 Aug 2010)
915
- [Ben Greear brought this change]
917
multi: Fix compile warning on 64-bit systems
919
Dan Fandrich (23 Aug 2010)
920
- Mention PolarSSL in tutorial & add some URLs to INSTALL
922
Daniel Stenberg (23 Aug 2010)
923
- RESUME_FROM: clarify what ftp uploads do
925
The numerical value passed to CURLOPT_RESUME_FROM for FTP uploads is
926
interpreted and used as position where to resume the _reading_ of the
927
local file and it will "blindly" append that data on the remote
928
file. This was certainly not clear in the docs previously.
931
Bug: http://curl.haxx.se/bug/view.cgi?id=3048174
933
- [Dirk Manske brought this change]
935
Curl_is_connected: use correct errno
937
The correctly extracted errno contents were mistakenly overwritten by a newer
938
value that wasn't the correct error value.
940
Bug: http://curl.haxx.se/mail/lib-2010-08/0242.html
942
- cmdline: make -F type= accept ;charset=
944
The -F option allows some custom parameters within the given string, and
945
those strings are separated with semicolons. You can for example specify
946
"name=daniel;type=text/plain" to set content-type for the
947
field. However, the use of semicolons like that made it not work fine if
948
you specified one within the content-type, like for:
949
"name=daniel;type=text/plain;charset=UTF-8"
950
... as the second one would be seen as a separator and "charset" is no
951
parameter curl knows anything about so it was just silently discarded.
953
The new logic now checks if the semicolon and following keyword looks
954
like a parameter it knows about and if it isn't it is assumed to be
955
meant to be used within the content-type string itself.
957
I modified test case 186 to verify that this works as intended.
959
Reported by: Larry Stone
960
Bug: http://curl.haxx.se/bug/view.cgi?id=3048988
962
Guenter Knauf (20 Aug 2010)
963
- Added mk-ca-bundle.vbs script.
965
The script works exactly same as the Perl one except for one thing:
966
when the text descriptions generated with openssl are included then
967
the md5 fingerprints are missing; seems openssl has either a bug or
968
a feature which prints the md5 fingerprint output to stdout instead
969
of writing them to specified file; this script could here do the same
970
as what the Perl scripr does (redirect stdout into file) but this
971
makes the script take up double the time because it needs to launch
972
cmd.exe 140 times (fo each openssl call). So I think for now we just
973
ommit the md5 fingerprints, and see if openssl will be fixed.
975
- Trial to fix win32 autobuilds.
977
It seems that its time to look at some better ideas for the win32
978
non-configure builds; probably a prebuild target which copies
979
config-win32.h to curl_config.h and appends also then feature
980
defines like USE_ARES.
982
Dan Fandrich (19 Aug 2010)
983
- Use the S_ISREG macro to determine what is a regular file
985
Kamil Dudka (19 Aug 2010)
986
- AC_INIT: avoid a warning with autoconf 2.66
988
It was complaining about the '=>' operator, introduced in e3fc0d5.
990
Dan Fandrich (18 Aug 2010)
991
- Fixed a memory leak during OOM in the multi timeout code
993
- Removed a C99ism & made an array const
995
Daniel Stenberg (19 Aug 2010)
996
- [Julien Chaffraix brought this change]
998
test: added test 579 to verify progress callback for chunked post
1000
The 66 bytes checked are those 38 bytes with the chunked encoding
1001
headers added: 8+8+10+35+5 = 66
1003
The three-letter words become 8 bytes on the wire because they are sent
1004
like: "3\r\none\r\n"
1006
... and there's the trailing 5 bytes write after the four lines since
1007
the final chunk is sent (which is "0\r\n\r\n").
1009
- multi: avoid sending multiple complete messages
1011
I fell over this bug report that mentioned that libcurl could wrongly
1012
send more than one complete messages at the end of a transfer. Reading
1013
the code confirmed this, so I've added a new multi state to make it not
1014
happen. The mentioned bug report was made by Brad Jorsch but is (oddly
1015
enough) filed in Debian's bug tracker for the "wmweather+" tool.
1017
Bug: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=593390
1019
- FAQ: update the list of supported protocols
1021
- FAQ: added blurb about ECCN
1023
"1.13 curl's ECCN number" is a new section mostly made up from
1024
Alessandro Vesely's very informative ML post on the subject:
1025
http://curl.haxx.se/mail/lib-2008-03/0251.html
1027
Guenter Knauf (18 Aug 2010)
1028
- It is sufficient to pipe stderr to NUL to get rid of the nasty messages.
1030
- Added SSPI build to Watcom makefile.
1032
Daniel Stenberg (16 Aug 2010)
1033
- [Julien Chaffraix brought this change]
1035
progress: callback for POSTs less than MAX_INITIAL_POST_SIZE
1037
Add a call to Curl_pgrsSetUploadSize in this case valided by a test
1040
Reported by: ŠŠøŠŗŠøŃŠ° ŠŠ¾ŃŠ¾Ń
ŠøŠ½.
1041
Bug: http://curl.haxx.se/mail/lib-2010-04/0173.html
1043
Dan Fandrich (16 Aug 2010)
1044
- Make the LD_PRELOAD path absolute in the tests that use it
1046
In some situations, libtool will change directories and perform
1047
a link step before executing the libtest test app. Since
1048
LD_PRELOAD is in effect for this entire process, the path to the
1049
binary must be absolute so it will be valid no matter in which
1050
directory the app is running.
1052
Daniel Stenberg (16 Aug 2010)
1053
- negotiation: Wrong proxy authorization
1055
There's an error in http_negotiation.c where a mistake is using only
1056
userpwd even for proxy requests. Ludek provided a patch, but I decided
1057
to write the fix slightly different using his patch as inspiration.
1059
Reported by: Ludek Finstrle
1060
Bug: http://curl.haxx.se/bug/view.cgi?id=3046066
1062
Dan Fandrich (16 Aug 2010)
1063
- Clear stdout and stderr files on each test run
1065
This allows a test to be run several times in the same test
1066
session even when the -k option is given.
1068
Guenter Knauf (15 Aug 2010)
1069
- Syncroniszed vclean target; fixed some comments.
1071
Daniel Stenberg (15 Aug 2010)
1072
- THANKS: added contributors from 7.21.1
1074
- multi: two fixes done
1076
- multi: use timeouts properly for MAX_RECV/SEND_SPEED
1078
When detecting that the send or recv speed, the multi interface changes
1079
state to TOOFAST and previously there was no timeout set that would
1080
force a recheck but it would rely on the application to somehow call
1081
libcurl anyway. This now sets a timeout for a suitable future time to
1082
check again if the average transfer speed is then below the threshold
1085
- multi: support timeouts
1087
Curl_expire() is now expanded to hold a list of timeouts for each easy
1088
handle. Only the closest in time will be the one used as the primary
1089
timeout for the handle and will be used for the splay tree (which sorts
1090
and lists all handles within the multi handle).
1092
When the main timeout has triggered/expired, the next timeout in time
1093
that is kept in the list will be moved to the main timeout position and
1094
used as the key to splay with. This way, all timeouts that are set with
1095
Curl_expire() internally will end up as a proper timeout. Previously any
1096
Curl_expire() that set a _later_ timeout than what was already set was
1097
just silently ignored and thus missed.
1099
Setting Curl_expire() with timeout 0 (zero) will cancel all previously
1102
Corrects known bug #62.
1104
- Curl_llist_insert_next: allow insertion first in the list
1106
When we specify the "insert after" entry as NULL, this function now
1107
inserts the new entry first in the list.
1109
- multi: make curl_multi_info_read perform O(1)
1111
Instead of looping over all attached easy handles, this now keeps a list
1112
of messages in the multi handle. It allows curl_multi_info_read() to
1113
perform O(1) no matter how many easy handles that are handled. This is
1114
of importance since this function may be polled very frequently by apps
1115
using the multi interface.
1117
Kamil Dudka (15 Aug 2010)
1118
- curl -T: ignore file size of special files
1120
original bug report at https://bugzilla.redhat.com/622520
1122
Dan Fandrich (13 Aug 2010)
1123
- Reset environment variables before starting servers
1125
Otherwise, variables from tests could affect the servers
1128
Kamil Dudka (12 Aug 2010)
1129
- typecheck-gcc: work around gcc upstream bug #32061
1131
original bug report at https://bugzilla.redhat.com/617757
1133
Daniel Stenberg (11 Aug 2010)
1134
- release cycle loop: start over toward 7.21.2
9
1136
Version 7.21.1 (11 Aug 2010)
11
1138
Daniel Stenberg (11 Aug 2010)
3353
4480
- Attempt to enhance stunnel 3.X logging
3355
4482
- Prevent running stunnel unless its version can be determined.
3357
- Log sws IPv version, port and pid when exiting due to SIGINT or SIGTERM.
3359
- To allow remote log inspection avoid redirecting messages to stderr.
3361
Set 0600 file permissions on certificate pem files.
3363
- Preserve empty line following last target
3365
- Log info for start up failures.
3367
Daniel Stenberg (26 Nov 2009)
3368
- - Extended and fixed the change I did on Dec 11 for the the progress
3369
meter/callback during FTP command/response sequences. It turned out it was
3370
really lame before and now the progress meter SHOULD get called at least
3373
Yang Tse (26 Nov 2009)
3374
- Only attempt to clear the server-logs lock when previously set by this same server.
3376
- signal handling to cleanup on SIGINT and SIGTERM
3378
- - Larry Lansing fixed ares_parse_srv_reply to properly parse replies
3379
which might contain non-SRV answers, skipping over potential non-SRV
3380
ones such as CNAMEs.
3382
- When using icc, compile with -fpic and link with intel dynamic libraries.
3384
Daniel Stenberg (24 Nov 2009)
3385
- clarify what a modern version of libcurl means
3387
Yang Tse (24 Nov 2009)
3388
- Added 'currently' in italics to insist on transient situation.
3392
Daniel Stenberg (24 Nov 2009)
3395
Yang Tse (23 Nov 2009)
3396
- - David Byron modified Makefile.dist vc8 and vc9 targets in order to allow
3397
finer granularity control when generating src and lib makefiles.
3399
- Enhance some debug messages for initialization failures.
3401
Fix compiler warning: conditional expression is constant.
3403
Daniel Stenberg (23 Nov 2009)
3404
- - Bjorn Augustsson reported a bug which made curl not report any problems even
3405
though it failed to write a very small download to disk (done in a single
3406
fwrite call). It turned out to be because fwrite() returned success, but
3407
there was insufficient error-checking for the fclose() call which tricked
3408
curl to believe things were fine.
3410
Yang Tse (23 Nov 2009)
3411
- Daniel wants upcoming release to be 1.7.0
3413
- Mention last changes
3415
- - Removed from external interface preprocessor symbol definition for
3416
CARES_HAVE_ARES_FREE_DATA. Current functionality of ares_free_data()
3417
makes it unnecessary.
3421
- Changed c-ares naming conventions when using MSVC as described in README.msvc
3423
Daniel Stenberg (23 Nov 2009)
3424
- fixed CURLOPT_TFTP_BLKSIZE typo
3426
Yang Tse (23 Nov 2009)
3427
- - Mention other recent changes
3429
- - Jakub Hrozek renamed addrttl and addr6ttl structs to ares_addrttl and
3430
ares_addr6ttl in order to prevent name space pollution, along with
3431
necessary changes to code base and man pages.This change does not break
3432
ABI, there is no need to recompile existing applications. But existing
3433
applications using these structs with the old name will need source code
3434
adjustments when recompiled using c-ares 1.6.1.
3436
- - Jakub Hrozek fixed more function prototypes in man pages to sync them
3437
with the ones declared in ares.h
3439
- Make configure remove the curlbuild.h file included in distribution tarballs
3440
for use by non-configure systems. As intended, configure would overwrite the
3441
distributed one when doing in-tree builds. But VPATH builds would end having
3442
two curlbuild.h files, one in the source tree and another in the build tree.
3444
- Make configure remove the ares_build.h file included in distribution tarballs.
3446
- Fix macro redefinition.
3448
- Fix name space pollution.
3450
- Allow using different extra import libraries for debug and release builds.
3452
- Add manifest stuff to msvc makefile
3454
Daniel Stenberg (20 Nov 2009)
3455
- - Constantine Sapuntzakis identified a write after close, as the sockets were
3456
closed by libcurl before the SSL lib were shutdown and they may write to its
3457
socket. Detected to at least happen with OpenSSL builds.
3459
Yang Tse (20 Nov 2009)
3460
- Sync man page with reality
3462
- Add missing external API decoration for ares_set_socket_callback()
3464
Daniel Stenberg (20 Nov 2009)
3465
- - Jad Chamcham pointed out a bug with connection re-use. If a connection had
3466
CURLOPT_HTTPPROXYTUNNEL enabled over a proxy, a subsequent request using the
3467
same proxy with the tunnel option disabled would still wrongly re-use that
3468
previous connection and the outcome would only be badness.
3470
Yang Tse (20 Nov 2009)
3471
- Add ares_free_data() man page.
3473
- - Provide in external interface preprocessor symbol definitions for
3474
CARES_HAVE_ARES_FREE_DATA as an indication of function availability.
3480
- Add ares_data.c and ares_data.h
3482
- Jakub Hrozek modified ares_parse_srv_reply() and ares_parse_txt_reply() API
3483
to return a linked lists of results. These were also modified to internally
3484
use the ares_data memory struct and as such its result must be free'ed with
3487
- Initial support for the generic ares_free_data() function that will allow
3488
applications to free memory allocated and returned by some c-ares funtions.
3490
- Explicit initialization of all FTP struct members in ftp_init()
3492
- Fix compiler warning: ISO C90 forbids mixed declarations and code
3494
- Make memory tracking system intolerant with zero sized malloc(),
3495
calloc() and realloc() function calls.
3497
- Make usage of calloc()'s arguments consistent with rest of code base
3499
- workaround icc 9.1 optimizer issue
3501
Daniel Stenberg (17 Nov 2009)
3502
- - Constantine Sapuntzakis provided another fix for the DNS cache that could
3503
end up with entries that wouldn't time-out:
3505
1. Set up a first web server that redirects (307) to a http://server:port
3507
2. Have curl connect to the first web server using curl multi
3509
After the curl_easy_cleanup call, there will be curl dns entries hanging
3510
around with in_use != 0.
3512
(http://curl.haxx.se/bug/view.cgi?id=2891591)
3514
- - Marc Kleine-Budde fixed: curl saved the LDFLAGS set during configure into
3515
its pkg-config file. So -Wl stuff ended up in the .pc file, which is really
3516
bad, and breaks if there are multiple -Wl in our LDFLAGS (which are in
3517
PTXdist). bug #2893592 (http://curl.haxx.se/bug/view.cgi?id=2893592)
3519
Yang Tse (16 Nov 2009)
3520
- Add icc fvisibility bug test
3522
- Fix icc 9.0 compiler warning: external definition with no prior declaration
3524
- Update serial number
3526
- Fix three var names
3528
- Add check for assert.h header file
3530
Kamil Dudka (15 Nov 2009)
3531
- - David Byron improved the configure script to use pkg-config to find OpenSSL
3532
(and in particular the list of required libraries) even if a path is given
3533
as argument to --with-ssl
3535
Yang Tse (15 Nov 2009)
3536
- getaddrinfo is fully thread safe on solaris versions which
3537
implement the function even when h_errno is not a macro.
3539
The h_errno macro test now only done on systems for which there
3540
is no hard coded knowledge about getaddrinfo's thread safeness.
3542
- Client certificate ENG file type requires OpenSSL 0.9.7 or newer
3544
- Remove files generated on previous buildconf/configure run
3546
- Add a couple of renamed files to the removal list
3548
- Remove enable-thread / disable-thread configure option. These were only placebo
3549
options. The library is always built as thread safe as possible on every system.
3551
Claes Jakobsson (14 Nov 2009)
3552
- Added '--configure' option to curl-config to display original configure arguments when curl was built
3554
Yang Tse (14 Nov 2009)
3555
- Refactor how preprocessor symbol _THREAD_SAFE definition is done.
3557
- OpenSSL 0.9.7 or newer required for ENGINE_CTRL_GET_CMD_FROM_NAME definition
3559
Daniel Stenberg (14 Nov 2009)
3560
- - Claes Jakobsson restored the configure functionality to detect NSS when
3561
--with-nss is set but not "yes".
3563
I think we can still improve that to check for pkg-config in that path etc,
3564
but at least this patch brings back the same functionality we had before.
3566
- - Camille Moncelier added support for the file type SSL_FILETYPE_ENGINE for
3567
the client certificate. It also disable the key name test as some engines
3568
can select a private key/cert automatically (When there is only one key
3569
and/or certificate on the hardware device used by the engine)
3571
Yang Tse (14 Nov 2009)
3572
- - Constantine Sapuntzakis provided the fix that ensures that an SSL connection
3573
won't be reused unless protection level for peer and host verification match.
3575
- Assume that getaddrinfo is thread safe, unless hard coded
3576
knowledge says the contrary or h_errno is not defined.
3578
- Related with the threadsafe capability of getaddrinfo:
3580
- Constantine Sapuntzakis reported that Darwin 6.0 a.k.a. MAC OS X 10.2
3581
and newer have a threadsafe getaddrinfo.
3583
- Fix Dragonfly BSD triplet detection.
3585
- In case the hard-coded knowledge says that getaddrinfo is threadsafe,
3586
an additional check is done to verify that h_errno is also defined.
3587
If h_errno isn't defined, we finally assume that it isn't threadsafe.
3588
Jamie Lokier provided the inspiration for this extra check.
3590
- AIX 5.2 and newer have threadsafe getaddrinfo.
3592
Add some comments to better understand what the regex's pretend to achieve.
3594
- HP-UX 11.11 and later have threadsafe getaddrinfo
3596
- Constantine Sapuntzakis patch for hostip.c
3598
No need for a separate variable ndns.
3600
The memory leak detection will detect code that fails to release a dns reference.
3602
The DEBUGASSERT will detect code that releases too many references.
3604
- Check if getaddrinfo is threadsafe when function check allows it to be used
3606
- Add missing variable initialization
3608
Kamil Dudka (12 Nov 2009)
3609
- - libcurl-NSS now tries to reconnect with TLS disabled in case it detects
3610
a broken TLS server. However it does not happen if SSL version is selected
3611
manually. The approach was originally taken from PSM. Kaspar Brand helped me
3612
to complete the patch. Original bug reports:
3613
https://bugzilla.redhat.com/525496
3614
https://bugzilla.redhat.com/527771
3616
- - Kevin Baughman provided a fix preventing libcurl-NSS from crash on doubly
3617
closed NSPR descriptor. The issue was hard to find, reported several times
3618
before and always closed unresolved. More info at the RH bug:
3619
https://bugzilla.redhat.com/534176
3621
Yang Tse (12 Nov 2009)
3622
- Make hostip.c compile again
3624
Dan Fandrich (11 Nov 2009)
3625
- Made the path to the binaries absolute
3627
Yang Tse (11 Nov 2009)
3628
- Add last change credit
3630
- - Marco Maggi reported that compilation failed when configured --with-gssapi
3631
and GNU GSS installed due to a missing mutual exclusion of header files in
3632
the Kerberos 5 code path. He also verified that my patch worked for him.
3634
Daniel Stenberg (11 Nov 2009)
3635
- NSS SSL connections are also blocking
3637
- - Constantine Sapuntzakis posted bug #2891595
3638
(http://curl.haxx.se/bug/view.cgi?id=2891595) which identified how an entry
3639
in the DNS cache would linger too long if the request that added it was in
3640
use that long. He also provided the patch that now makes libcurl capable of
3641
still doing a request while the DNS hash entry may get timed out.
3643
- - Christian Schmitz noticed that the progress meter/callback was not properly
3644
used during the FTP connection phase (after the actual TCP connect), while
3645
it of course should be. I also made the speed check get called correctly so
3646
that really slow servers will trigger that properly too.
3648
Yang Tse (11 Nov 2009)
3649
- Renamed fpGetNetworkParams and fpSystemFunction036 to avoid namespace pollution with static library
3653
- Mention last changes
3655
- Reinstate copyright symbol lost in previous commit
3657
- Make some strings different in resource file for debug or release builds
3659
- Ignore more subdirs
3661
- Fix compiler warning: conditional expression is constant
3663
- Sync linker and resource compiler options with Makefile.msvc
3665
- Follow Makefile.msvc subdirectory naming scheme, and sync compiler options
3667
- Updated MSVC makefile that allows building dynamic and static
3668
c-ares libraries in debug and release flavours.
3670
Additionally each of the three sample programs is built against
3671
each of the four possible c-ares libraries, generating all this
3672
a total number of 12 executables and 4 libraries.
3674
- Test for USE_WINSOCK since it is more restrictive than WIN32
3676
- Make header inclusion depend on HAVE_*_H definition
3678
- Remove unneeded preprocessor directives
3680
Daniel Stenberg (7 Nov 2009)
3681
- Joshua Kwan added paragraph 4.18 about file:// URLs on windows etc
3683
Yang Tse (5 Nov 2009)
3684
- Adjust c-ares include paths for memory tracking enabled (--enable-curldebug) builds
3686
Kamil Dudka (5 Nov 2009)
3687
- - Dropped misleading timeouts in libcurl-NSS and made sure the SSL socket works
3688
in non-blocking mode.
3690
Yang Tse (5 Nov 2009)
3691
- I removed leading 'curl' path on the 'curlbuild.h' include statement in
3692
curl.h, adjusting auto-makefiles include path, to enhance portability to
3693
OS's without an orthogonal directory tree structure such as OS/400.
3695
- avoid division by zero
3697
Daniel Stenberg (4 Nov 2009)
3698
- - I fixed several problems with the transfer progress meter. It showed the
3699
wrong percentage for small files, most notable for <1000 bytes and could
3700
easily end up showing more than 100% at the end. It also didn't show any
3701
percentage, transfer size or estimated transfer times when transferring
3702
less than 100 bytes.
3704
Yang Tse (4 Nov 2009)
3705
- source files used by sample programs
3707
Daniel Stenberg (4 Nov 2009)
3708
- start over towards 7.19.8
3710
- updated with names from the 7.19.7 release notes
3712
Version 7.19.7 (4 Nov 2009)
3714
Daniel Stenberg (4 Nov 2009)
3715
- fix the find command line
3717
- avoid fiddling in the ares subdir since that isn't part of the release
3724
- I have next to no hope of fixing more bugs before 7.19.7 so all are now
3725
moved to 7.19.8. I removed the bugs already in KNOWN_BUGS (but they should
3726
of course still get fixed).
3728
Added three recent bugs. 7.19.8 is targetted to get shipped in Janurary 2010
3730
Dan Fandrich (4 Nov 2009)
3731
- Put the Symbian binaries into the more modern locations in ROM
3733
- Updated systems, bindings, NTLM support
3735
- Export the Symbian .iby files to make it easier to build a ROM
3737
Daniel Stenberg (2 Nov 2009)
3738
- strerror() => curl_easy_strerror()
3740
Yang Tse (2 Nov 2009)
3741
- include progress.h for Curl_pgrsSetDownloadCounter function prototype
3743
Daniel Stenberg (2 Nov 2009)
3744
- - As reported independent by both Stan van de Burgt and Didier Brisebourg,
3745
CURLINFO_SIZE_DOWNLOAD (the -w variable size_download) didn't work when
3746
getting data from ldap!
3748
- describe the cmake situation
3750
Yang Tse (2 Nov 2009)
3751
- Renamed c-ares setup.h to ares_setup.h
3753
- Adjust include paths to take in account that currently:
3755
c-ares with --enable-curldebug uses memdebug.h from libcurl's lib subdirectory.
3757
memdebug.h needs access to libcurl's setup.h from libcurl's lib subdirectory
3758
and also needs access to libcurl's generated curl_config.h
3760
- Undo old temporary change once used for testing purposes
3762
Daniel Stenberg (31 Oct 2009)
3763
- - Gabriel Kuri reported a problem with CURLINFO_CONTENT_LENGTH_DOWNLOAD if the
3764
download was 0 bytes, as libcurl would then return the size as unknown (-1)
3765
and not 0. I wrote a fix and test case 566 to verify it.
3767
Yang Tse (31 Oct 2009)
3768
- Mention many changes
3770
- Mention --enable-symbol-hiding configure option
3772
- Symbol hiding configure options renamed to the hopefully less ambiguous
3773
--enable-symbol-hiding and --disable-symbol-hiding as well as related
3774
macro names and some internal variables used for them.
3776
Related configuration file preprocessor symbols named to
3777
CARES_SYMBOL_HIDING and CARES_SYMBOL_SCOPE_EXTERN.
3779
Daniel Stenberg (30 Oct 2009)
3780
- Added test case 566 in an attempt to repeat bug 2884561
3781
(http://curl.haxx.se/bug/view.cgi?id=2884561) but it seems to work for me...
3783
- - Liza Alenchery mentioned a problem with re-used SCP connection when a bad
3784
auth is used, as it caused a crash. I failed to repeat the issue, but still
3785
made a change that now forces the TCP connection used for a freed SCP
3786
session to get closed and not be re-used.
3788
- - "Tom" posted a bug report that mentioned how libcurl did wrong when doing a
3789
POST using a read callback, with Digest authentication and
3790
"Transfer-Encoding: chunked" enforced. I would then cause the first request
3791
to be wrongly sent and then basically hang until the server closed the
3792
connection. I fixed the problem and added test case 565 to verify it.
3794
Yang Tse (30 Oct 2009)
3795
- Header inclusion depending on HAVE_* symbol.
3798
- Comparison of the Initial revision of this file with ares_parse_a_reply.c
3799
shows that this one is actually a modified copy of ares_parse_a_reply.c.
3801
In order to comply with ares_parse_a_reply.c's M.I.T. license, the old
3802
1998 M.I.T. copyright notice is now also preserved in this file the same
3803
as it is done in other ares_parse_*.c files.
3806
Fix identation of some license lines.
3812
- In no particular order, changed/fixed all of the following in
3813
ares_parse_txt_reply() current version:
3815
- Fixed a couple of potential double free's.
3817
- Fixed memory leaks upon out of memory condition.
3819
- Fixed pointer arithmetic.
3821
- Setting ntxtreply to zero upon entry for all failure cases.
3823
- Changed data type to size_t for variables substr_len, str_len and
3824
the length member of ares_txt_reply struct.
3826
- Avoided a couple of memcpy() calls.
3828
- Changed i data type to unsigned int to prevent compiler warnings.
3830
- Adjusted a comment.
3832
- Use ARES_SUCCESS literal for successfull completion.
3836
- Add c-ares DLL resource file to distribution archive
3842
- Updated MSVC 6.0 workspace and project files that allows building
3843
dynamic and static c-ares libraries in debug and release flavours.
3845
Additionally each of the three sample programs is built against
3846
each of the four possible c-ares libraries, generating all this
3847
a total number of 12 executables and 4 libraries.
3849
Daniel Stenberg (29 Oct 2009)
3850
- no need to check for NULL pointers before dereferencing, as the pointers
3851
MUST be valid and they are dereferenced further down in the function
3854
- shorten the descriptions somewhat
3856
- update to the new struct name
3858
- Jakub Hrozek added ares_parse_txt_reply() for TXT parsing
3860
- use 'ares_srv_reply' for proper name-spacing
3862
Yang Tse (29 Oct 2009)
3863
- Fix compiler warning: control reaches end of non-void function
3865
- Take in account c-ares 1.6.1 will use __declspec function decoration
3866
for Win32 and Symbian unless CARES_STATICLIB is defined to use static
3869
- Add reference for ares_parse_srv_reply.pdf
3871
- Add reference for ares_parse_srv_reply docs
3873
- External API function linkage decoration adjustment
3875
- External API function linkage decoration adjustment
3877
Daniel Stenberg (28 Oct 2009)
3878
- Since the NSS lib closes the socket the memory tracking system wrongly gets a
3879
false positive on a leaked socket, so this introduces a way to tell the system
3880
that the socket is indeed closed without explicitly closing it!
3882
Yang Tse (28 Oct 2009)
3883
- Initial step towards the ability to reduce c-ares exported symbols
3884
based on the 'visibility' attribute for GNUC and __global for Sun
3885
compilers, taking also in account __declspec function decoration
3886
for Win32 and Symbian DLL's.
3888
Introducing configure options --enable-hidden-symbols and
3889
--disable-hidden-symbols following libcurl's naming.
3891
- libssh2 detection changes mentioned in http://curl.haxx.se/mail/lib-2009-10/0343.html
3897
- Fix Pelles C Win32 target compilation issues
3899
Daniel Stenberg (26 Oct 2009)
3900
- grrr, removed the conflict markers
3902
- spell fix by Michael Wood
3904
- - Dima Barsky made the curl cookie parser accept cookies even with blank or
3905
unparsable expiry dates and then treat them as session cookies - previously
3906
libcurl would reject cookies with a date format it couldn't parse. Research
3907
shows that the major browser treat such cookies as session cookies. I
3908
modified test 8 and 31 to verify this.
3910
Dan Fandrich (24 Oct 2009)
3911
- Mention that Android works
3913
Daniel Stenberg (24 Oct 2009)
3914
- Fabian Keil's suggested wording
3916
Yang Tse (23 Oct 2009)
3917
- John Engelhart noticed an unreleased problem relative to a duplicate
3918
ARES_ECANCELLED error code value and missing error code description.
3920
Patrick Monnerat (23 Oct 2009)
3921
- OS400 install: Duplicate H include file to file CURL to support more include path forms.
3923
Daniel Stenberg (22 Oct 2009)
3924
- update the company sponsors part, as one company is no longer sponsoring
3925
and another is sponsoring more
3927
Yang Tse (21 Oct 2009)
3928
- Update MSVC 6 caveats section
3930
- Update memory tracking/debugging reference
3932
Daniel Stenberg (21 Oct 2009)
3933
- - Attempt to use pkg-config for finding out libssh2 installation details
3936
- mention bug 2848436 in regard to the TFTP problems
3938
- s/koresh/Johan van Selst/
3940
- add --crlfile to the man page
3942
- - A patch in bug report #2883177 (http://curl.haxx.se/bug/view.cgi?id=2883177)
3943
by user 'koresh' introduced the --crlfile option to curl, which makes curl
3944
tell libcurl about a file with CRL (certificate revocation list) data to
3947
- spelling corrections
3949
Yang Tse (21 Oct 2009)
3950
- Detect when msvc 6.0 is in use without PSDK installed, intentionally
3951
fail to build when this happens, and show an appropriate error.
3953
The brave of heart can circumvect this. Defining ALLOW_MSVC6_WITHOUT_PSDK
3954
in lib/config-win32.h, although absolutely discouraged and unsupported,
3955
this will allow the die hard MSVC hacker to build in such a discouraged
3958
The actually supported 'fix' is to install 'February 2003 Platform SDK'
3959
a.k.a. 'Windows Server 2003 PSDK' which can be freely downloaded from
3960
http://www.microsoft.com/msdownload/platformsdk/sdkupdate/psdk-full.htm
3962
Gunter Knauf (19 Oct 2009)
3963
- added cast macros to silent compiler warnings with 64-bit systems.
3965
Yang Tse (19 Oct 2009)
3966
- Fix compiler warning: local variable may be used without having been initialized
3968
- Use *_CHECK_PATH_SEPARATOR_REQUIRED to ensure that *_CHECK_PATH_SEPARATOR
3969
is only expanded and included once in the configure script.
3971
- Our _AS_PATH_SEPARATOR_PREPARE override is now m4_defun'd instead of m4_define'd
3972
due to autoconf 2.64 m4_require'ing it in _AS_SHELL_SANITIZE indirectly through
3975
Daniel Stenberg (18 Oct 2009)
3976
- - Ray Dassen provided a patch in Debian's bug tracker (bug number #551461)
3977
that now makes curl_getdate(3) actually handles RFC 822 formatted dates that
3978
use the "single letter military timezones".
3979
http://www.rfc-ref.org/RFC-TEXTS/822/chapter5.html has the details.
3981
Yang Tse (18 Oct 2009)
3982
- Check for basename() is now done the same as other function checks
3984
Daniel Stenberg (18 Oct 2009)
3985
- - Fixed memory leak in the SCP/SFTP code as it never freed the knownhosts
3988
- John Dennis filed bug report #2873666
3989
(http://curl.haxx.se/bug/view.cgi?id=2873666) which identified a problem which
3990
made libcurl loop infinitely when given incorrect credentials when using HTTP
3991
GSS negotiate authentication.