]> granicus.if.org Git - curl/log
curl
9 years agoopenssl: adapt to 1.1.0+ name changes
Daniel Stenberg [Thu, 10 Dec 2015 16:30:31 +0000 (17:30 +0100)]
openssl: adapt to 1.1.0+ name changes

9 years agoscripts/makefile: add standard header
Daniel Stenberg [Tue, 8 Dec 2015 23:34:39 +0000 (00:34 +0100)]
scripts/makefile: add standard header

9 years agoscripts/Makefile: fix GNUism and survive no perl
Daniel Stenberg [Tue, 8 Dec 2015 23:32:42 +0000 (00:32 +0100)]
scripts/Makefile: fix GNUism and survive no perl

Closes #555

Reported-by: Thomas Klausner
9 years agofix b6d5cb40d7038fe
Daniel Stenberg [Tue, 8 Dec 2015 23:27:04 +0000 (00:27 +0100)]
fix b6d5cb40d7038fe

9 years agohttp2: Fix hanging paused stream
Tatsuhiro Tsujikawa [Fri, 4 Dec 2015 15:40:10 +0000 (00:40 +0900)]
http2: Fix hanging paused stream

When NGHTTP2_ERR_PAUSE is returned from data_source_read_callback, we
might not process DATA frame fully.  Calling nghttp2_session_mem_recv()
again will continue to process DATA frame, but if there is no incoming
frames, then we have to call it again with 0-length data.  Without this,
on_stream_close callback will not be called, and stream could be hanged.

Bug: http://curl.haxx.se/mail/lib-2015-11/0103.html
Reported-by: Francisco Moraes
9 years agobuild: fix compilation error with CURL_DISABLE_VERBOSE_STRINGS
Christian Stewart [Tue, 8 Dec 2015 15:04:52 +0000 (10:04 -0500)]
build: fix compilation error with CURL_DISABLE_VERBOSE_STRINGS

With curl disable verbose strings in http.c the compilation fails due to
the data variable being undefined later on in the function.

Closes #558

9 years agoconfig-win32: Fix warning HAVE_WINSOCK2_H undefined
Gisle Vanem [Tue, 8 Dec 2015 00:05:42 +0000 (19:05 -0500)]
config-win32: Fix warning HAVE_WINSOCK2_H undefined

9 years agoopenssl: BoringSSL doesn't have CONF_modules_free
Gisle Vanem [Mon, 7 Dec 2015 23:43:35 +0000 (18:43 -0500)]
openssl: BoringSSL doesn't have CONF_modules_free

9 years agolwip: Fix compatibility issues with later versions
Gisle Vanem [Mon, 7 Dec 2015 19:27:29 +0000 (14:27 -0500)]
lwip: Fix compatibility issues with later versions

The name of the header guard in lwIP's <lwip/opt.h> has changed from
'__LWIP_OPT_H__' to 'LWIP_HDR_OPT_H' (bug #35874 in May 2015).

Other fixes:

- In curl_setup.h, the problem with an old PSDK doesn't apply if lwIP is
used.

- In memdebug.h, the 'socket' should be undefined first due to lwIP's
lwip_socket() macro.

- In curl_addrinfo.c lwIP's getaddrinfo() + freeaddrinfo() macros need
special handling because they were undef'ed in memdebug.h.

- In select.c we can't use preprocessor conditionals inside select if
MSVC and select is a macro, as it is with lwIP.

http://curl.haxx.se/mail/lib-2015-12/0023.html
http://curl.haxx.se/mail/lib-2015-12/0024.html

9 years agoos400: define CURL_VERSION_PSL in ILE/RPG binding
Patrick Monnerat [Mon, 7 Dec 2015 09:09:33 +0000 (10:09 +0100)]
os400: define CURL_VERSION_PSL in ILE/RPG binding

9 years agoversion: Add flag CURL_VERSION_PSL for libpsl
Gisle Vanem [Thu, 3 Dec 2015 06:02:50 +0000 (01:02 -0500)]
version: Add flag CURL_VERSION_PSL for libpsl

9 years agoformdata: Check if length is too large for memory
Jay Satiro [Mon, 7 Dec 2015 07:43:24 +0000 (02:43 -0500)]
formdata: Check if length is too large for memory

- If the size of the length type (curl_off_t) is greater than the size
of the size_t type then check before allocating memory to make sure the
value of length will fit in a size_t without overflow. If it doesn't
then return CURLE_BAD_FUNCTION_ARGUMENT.

Bug: https://github.com/bagder/curl/issues/425#issuecomment-154518679
Reported-by: Steve Holme
9 years agotests: Corrected copy and pasted comments from commit e643c5c908
Steve Holme [Thu, 3 Dec 2015 23:31:24 +0000 (23:31 +0000)]
tests: Corrected copy and pasted comments from commit e643c5c908

9 years agocurl: remove keepalive #ifdef checks done on libcurl's behalf
Daniel Stenberg [Thu, 3 Dec 2015 08:53:53 +0000 (09:53 +0100)]
curl: remove keepalive #ifdef checks done on libcurl's behalf

They didn't match the ifdef logic used within libcurl anyway so they
could indeed warn for the wrong case - plus the tool cannot know how the
lib actually performs at that level.

9 years agotest947: Corrected typo in test name
Steve Holme [Wed, 2 Dec 2015 23:45:29 +0000 (23:45 +0000)]
test947: Corrected typo in test name

9 years agotests: Disable the OAUTHBEARER tests when using a non-default port number
Steve Holme [Wed, 2 Dec 2015 23:10:02 +0000 (23:10 +0000)]
tests: Disable the OAUTHBEARER tests when using a non-default port number

Tests 842, 843, 844, 845, 887, 888, 889, 890, 946, 947, 948 and 949 fail
if a custom port number is specified via the -b option of runtests.pl.

Suggested by: Kamil Dudka
Bug: http://curl.haxx.se/mail/lib-2015-12/0003.html

9 years agobump: towards next release
Daniel Stenberg [Wed, 2 Dec 2015 22:24:47 +0000 (23:24 +0100)]
bump: towards next release

for all we know now, it might be called 7.46.1

9 years agoRELEASE-NOTES: updated contributor count for 7.46.0 curl-7_46_0
Daniel Stenberg [Tue, 1 Dec 2015 22:10:16 +0000 (23:10 +0100)]
RELEASE-NOTES: updated contributor count for 7.46.0

9 years agoTHANKS: new contributors from the 7.46.0 release
Daniel Stenberg [Tue, 1 Dec 2015 21:53:24 +0000 (22:53 +0100)]
THANKS: new contributors from the 7.46.0 release

9 years agoTHANKS-filter: single Tim Rühsen spelling
Daniel Stenberg [Tue, 1 Dec 2015 21:51:50 +0000 (22:51 +0100)]
THANKS-filter: single Tim Rühsen spelling

9 years agodocs/examples: gitignore some more built examples
Daniel Stenberg [Tue, 1 Dec 2015 08:09:04 +0000 (09:09 +0100)]
docs/examples: gitignore some more built examples

9 years agoRELEASE-NOTES; this bug was never released
Daniel Stenberg [Mon, 30 Nov 2015 07:22:32 +0000 (08:22 +0100)]
RELEASE-NOTES; this bug was never released

9 years agoRELEASE-NOTES: synced with e55f15454efacb0
Daniel Stenberg [Mon, 30 Nov 2015 07:03:28 +0000 (08:03 +0100)]
RELEASE-NOTES: synced with e55f15454efacb0

9 years agoCurl_read_plain: clean up ifdefs that break statements
Flavio Medeiros [Sun, 29 Nov 2015 14:12:16 +0000 (11:12 -0300)]
Curl_read_plain: clean up ifdefs that break statements

Closes #546

9 years agohttp2: convert some verbose output into debug-only output
Daniel Stenberg [Sun, 29 Nov 2015 23:12:46 +0000 (00:12 +0100)]
http2: convert some verbose output into debug-only output

9 years agohttp2 push: add missing inits of new stream
Daniel Stenberg [Sun, 29 Nov 2015 23:11:42 +0000 (00:11 +0100)]
http2 push: add missing inits of new stream

- set the correct stream_id for pushed streams
- init maxdownload and size properly

9 years agohttp2 push: set weight for new stream
Daniel Stenberg [Sun, 29 Nov 2015 23:10:35 +0000 (00:10 +0100)]
http2 push: set weight for new stream

give the new stream the old one's stream_weight internally to avoid
sending a PRIORITY frame unless asked for it

9 years agocurl_setup.h: undef freeaddrinfo in c-ares block to fix build
Daniel Stenberg [Fri, 27 Nov 2015 22:24:59 +0000 (23:24 +0100)]
curl_setup.h: undef freeaddrinfo in c-ares block to fix build

Fixes warnings 78c25c854a added.

9 years agononblock: fix setting non-blocking mode for Amiga
Daniel Stenberg [Fri, 27 Nov 2015 22:29:30 +0000 (23:29 +0100)]
nonblock: fix setting non-blocking mode for Amiga

IoctlSocket() apparently wants a pointer to a long, passed as a char *
in its third parameter. This bug was introduced already back in commit
c5fdeef41d from October 1 2001!

Bug: http://curl.haxx.se/mail/lib-2015-11/0088.html
Reported-by: Norbert Kett
9 years agozsh install: fix DESTDIR support
Daniel Stenberg [Fri, 27 Nov 2015 22:24:07 +0000 (23:24 +0100)]
zsh install: fix DESTDIR support

Reported-by: Mohammad AlSaleh
9 years agolib: Only define curl_dofreeaddrinfo if struct addrinfo is available
Dan Fandrich [Fri, 27 Nov 2015 09:51:22 +0000 (10:51 +0100)]
lib: Only define curl_dofreeaddrinfo if struct addrinfo is available

9 years agotool_paramhlp: Fixed display of URL index in password prompt for --next
Steve Holme [Fri, 27 Nov 2015 06:59:02 +0000 (06:59 +0000)]
tool_paramhlp: Fixed display of URL index in password prompt for --next

Commit f3bae6ed73 added the URL index to the password prompt when using
--next. Unfortunately, because the size_t specifier (%zu) is not
supported by all sprintf() implementations we use the curl_off_t format
specifier instead. The display of an incorrect value arises on platforms
where size_t and curl_off_t are of a different size.

9 years agotimecond: do not add if-modified-since without timecondition
Daniel Stenberg [Wed, 25 Nov 2015 10:38:10 +0000 (11:38 +0100)]
timecond: do not add if-modified-since without timecondition

The RTSP code path didn't skip adding the if-modified-since for certain
RTSP code paths, even if CURLOPT_TIMECONDITION was set to
CURL_TIMECOND_NONE.

Also, an unknown non-zero CURLOPT_TIMECONDITION value no longer equals
CURL_TIMECOND_IFMODSINCE.

Bug: http://stackoverflow.com/questions/33903982/curl-timecond-none-doesnt-work-how-to-remove-if-modified-since-header

9 years agoRELEASE-NOTES: synced with 99d17a5e2ba77e58
Daniel Stenberg [Wed, 25 Nov 2015 10:29:00 +0000 (11:29 +0100)]
RELEASE-NOTES: synced with 99d17a5e2ba77e58

9 years agoexamples/README: cut out the incomplete list
Daniel Stenberg [Wed, 25 Nov 2015 09:49:22 +0000 (10:49 +0100)]
examples/README: cut out the incomplete list

... and add a generic explanation for them instead. Each example file
should contain its own description these days.

9 years agotest1513: make sure the callback is only called once
Daniel Stenberg [Tue, 24 Nov 2015 22:33:37 +0000 (23:33 +0100)]
test1513: make sure the callback is only called once

9 years agobuild: Install zsh completion
Daniel Shahaf [Fri, 20 Nov 2015 04:56:10 +0000 (04:56 +0000)]
build: Install zsh completion

Fixes #534
Closes #537

9 years agodone: make sure the final progress update is made
Daniel Stenberg [Tue, 24 Nov 2015 18:51:59 +0000 (19:51 +0100)]
done: make sure the final progress update is made

It would previously be skipped if an existing error was returned, but
would lead to a previous value being left there and later used.
CURLINFO_TOTAL_TIME for example.

Still it avoids that final progress update if we reached DONE as the
result of a callback abort to avoid another callback to be called after
an abort-by-callback.

Reported-by: Lukas Ruzicka
Closes #538

9 years agocurl: expanded the -XHEAD warning text
Daniel Stenberg [Tue, 24 Nov 2015 09:11:06 +0000 (10:11 +0100)]
curl: expanded the -XHEAD warning text

... to also mention the specific options used.

9 years agoRevert "cleanup: general removal of TODO (and similar) comments"
Daniel Stenberg [Tue, 24 Nov 2015 08:32:42 +0000 (09:32 +0100)]
Revert "cleanup: general removal of TODO (and similar) comments"

This reverts commit 64e959ffe37c436503f9fed1ce2d6ee6ae50bd9a.

Feedback-by: Dan Fandrich
URL: http://curl.haxx.se/mail/lib-2015-11/0062.html

9 years agoCURLOPT_HEADERFUNCTION.3: fix typo
Daniel Stenberg [Mon, 23 Nov 2015 15:11:57 +0000 (16:11 +0100)]
CURLOPT_HEADERFUNCTION.3: fix typo

Refer to _HEADERDATA not _WRITEDATA.

Reported-by: Michał Piechowski
9 years agoTODO: TCP Fast Open
Daniel Stenberg [Mon, 23 Nov 2015 09:58:12 +0000 (10:58 +0100)]
TODO: TCP Fast Open

9 years agoexamples: Added website parse-able descriptions to the e-mail examples
Steve Holme [Sun, 22 Nov 2015 11:55:10 +0000 (11:55 +0000)]
examples: Added website parse-able descriptions to the e-mail examples

9 years agoTODO: Added another 'multi-interface' idea
Steve Holme [Sat, 21 Nov 2015 17:59:14 +0000 (17:59 +0000)]
TODO: Added another 'multi-interface' idea

9 years agosmb.c: Fixed compilation warnings
Steve Holme [Sat, 21 Nov 2015 11:41:20 +0000 (11:41 +0000)]
smb.c: Fixed compilation warnings

smb.c:134:3: warning: conversion to 'short unsigned int' from 'int' may
             alter its value
smb.c:146:42: warning: conversion to 'unsigned int' from 'long long
              unsigned int' may alter its value
smb.c:146:65: warning: conversion to 'unsigned int' from 'long long
              unsigned int' may alter its value

9 years agoschannel: Corrected copy/paste error in commit 8d17117683
Steve Holme [Sat, 21 Nov 2015 02:54:44 +0000 (02:54 +0000)]
schannel: Corrected copy/paste error in commit 8d17117683

9 years agoschannel: Use GetVersionEx() when VerifyVersionInfo() isn't available
Steve Holme [Sat, 21 Nov 2015 02:43:17 +0000 (02:43 +0000)]
schannel: Use GetVersionEx() when VerifyVersionInfo() isn't available

Regression from commit 7a8e861a5 as highlighted in the msys autobuilds.

9 years agoexamples: Fixed compilation warnings
Steve Holme [Sat, 21 Nov 2015 01:47:02 +0000 (01:47 +0000)]
examples: Fixed compilation warnings

pop3-multi.c:96:5: warning: implicit declaration of function 'memset'
imap-multi.c:96:5: warning: implicit declaration of function 'memset'
http2-download.c:226:5: warning: implicit declaration of function 'memset'
http2-upload.c:290:5: warning: implicit declaration of function 'memset'
http2-upload.c:290:5: warning: implicit declaration of function 'memset'

9 years agoMakefile.inc: Fixed test run error
Steve Holme [Sat, 21 Nov 2015 00:04:57 +0000 (00:04 +0000)]
Makefile.inc: Fixed test run error

test845 not present in tests/data/Makefile.inc

9 years agoTODO: remove duplicated title
Daniel Stenberg [Fri, 20 Nov 2015 08:37:11 +0000 (09:37 +0100)]
TODO: remove duplicated title

9 years agoTODO: added two more libcurl ideas
Daniel Stenberg [Fri, 20 Nov 2015 07:39:59 +0000 (08:39 +0100)]
TODO: added two more libcurl ideas

Moved some ideas from "next major" to just ordinary ideas since we can
always add new things while keeping the old without doing a "next
major".

9 years agotests: Re-enabled tests 889 and 890 following POP3 fix
Steve Holme [Fri, 20 Nov 2015 07:04:19 +0000 (07:04 +0000)]
tests: Re-enabled tests 889 and 890 following POP3 fix

9 years agopop3: Differentiate between success and continuation responses
Steve Holme [Fri, 20 Nov 2015 07:01:01 +0000 (07:01 +0000)]
pop3: Differentiate between success and continuation responses

9 years agopop3: Added clarity on some server response codes
Steve Holme [Fri, 20 Nov 2015 06:41:53 +0000 (06:41 +0000)]
pop3: Added clarity on some server response codes

9 years agobuild: Fix theoretical infinite loops
Daniel Shahaf [Thu, 19 Nov 2015 22:23:11 +0000 (22:23 +0000)]
build: Fix theoretical infinite loops

Add error-checking to 'cd' in a few cases where omitting the checks
might result in an infinite loop.

Closes #535

9 years agocurl.h: s/#defien/#define/
Patrick Monnerat [Thu, 19 Nov 2015 12:53:47 +0000 (13:53 +0100)]
curl.h: s/#defien/#define/

9 years agoos400: synchronize ILE/RPG header file
Patrick Monnerat [Thu, 19 Nov 2015 12:52:06 +0000 (13:52 +0100)]
os400: synchronize ILE/RPG header file

9 years agoos400: Provide options for libssh2 use in compile scripts. Adjust README.
Patrick Monnerat [Thu, 19 Nov 2015 12:35:29 +0000 (13:35 +0100)]
os400: Provide options for libssh2 use in compile scripts. Adjust README.

9 years agozsh completion: Preserve single quotes in output
danielsh@apache.org [Wed, 18 Nov 2015 21:52:32 +0000 (21:52 +0000)]
zsh completion: Preserve single quotes in output

When an option's help string contains literal single quotes, those
single quotes would be stripped from the option's description in the
completion output (unless the zsh RC_QUOTES option were set while the
completion function was being sourced, which is not the default).  This
patch makes the completion output contain single quotes where the --help
output does.

Closes #532

9 years agoFAQ: Grammar changes
MaxGiting [Thu, 19 Nov 2015 04:00:24 +0000 (23:00 -0500)]
FAQ: Grammar changes

Closes https://github.com/bagder/curl/pull/533

9 years agohttp2: http_done: don't free already-freed push headers
Daniel Stenberg [Tue, 17 Nov 2015 08:47:58 +0000 (09:47 +0100)]
http2: http_done: don't free already-freed push headers

The push headers are freed after the push callback has been invoked,
meaning this code should only free the headers if the callback was never
invoked and thus the headers weren't freed at that time.

Reported-by: Davey Shafik
9 years agogetconnectinfo: Don't call recv(2) if socket == -1
Anders Bakken [Mon, 16 Nov 2015 20:50:41 +0000 (12:50 -0800)]
getconnectinfo: Don't call recv(2) if socket == -1

Closes #528

9 years agoCURLMOPT_PUSHFUNCTION.3: *_byname() returns only the first header
Daniel Stenberg [Mon, 16 Nov 2015 07:26:22 +0000 (08:26 +0100)]
CURLMOPT_PUSHFUNCTION.3: *_byname() returns only the first header

... if there are more than one using the same name

9 years agohttp2: minor comment typo
Daniel Stenberg [Mon, 16 Nov 2015 07:22:08 +0000 (08:22 +0100)]
http2: minor comment typo

9 years agosasl; fix checksrc warnings
Daniel Stenberg [Sun, 15 Nov 2015 22:15:00 +0000 (23:15 +0100)]
sasl; fix checksrc warnings

9 years agoRELEASE-NOTES: Adjusted for the recent OAuth 2.0 activity
Steve Holme [Sun, 15 Nov 2015 20:18:44 +0000 (20:18 +0000)]
RELEASE-NOTES: Adjusted for the recent OAuth 2.0 activity

9 years agotests: Disabled 889 and 890 until we support POP3 continuation responses
Steve Holme [Sun, 15 Nov 2015 20:14:46 +0000 (20:14 +0000)]
tests: Disabled 889 and 890 until we support POP3 continuation responses

As POP3 final and continuation responses both begin with a + character,
and both the finalcode and contcode variables in SASLprotoc are set as
such, we cannot tell the difference between them when we are expecting
an optional continuation from the server such as the following:

+ something else from the server
+OK final response

Disabled these tests until such a time we can tell the responses apart.

9 years agotests: Corrected typos from commit ba4d8f7eba
Steve Holme [Sun, 15 Nov 2015 17:39:05 +0000 (17:39 +0000)]
tests: Corrected typos from commit ba4d8f7eba

9 years agotests: Added OAUTHBEARER failure response tests
Steve Holme [Sun, 15 Nov 2015 17:36:59 +0000 (17:36 +0000)]
tests: Added OAUTHBEARER failure response tests

9 years agooauth2: Support OAUTHBEARER failures sent as continuation responses
Steve Holme [Sat, 5 Sep 2015 17:35:47 +0000 (18:35 +0100)]
oauth2: Support OAUTHBEARER failures sent as continuation responses

According to RFC7628 a failure message may be sent by the server in a
base64 encoded JSON string as a continuation response.

Currently only implemented for OAUTHBEARER and not XAUTH2.

9 years agoRELEASE-NOTES: synced with 808a17ee675
Daniel Stenberg [Sun, 15 Nov 2015 18:12:06 +0000 (19:12 +0100)]
RELEASE-NOTES: synced with 808a17ee675

9 years agotests: Renamed existing OAuth 2.0 (XOAUTH) tests
Steve Holme [Sat, 14 Nov 2015 11:19:57 +0000 (11:19 +0000)]
tests: Renamed existing OAuth 2.0 (XOAUTH) tests

9 years agotests: Added OAuth 2.0 (OAUTHBEARER) tests
Steve Holme [Sat, 14 Nov 2015 11:16:04 +0000 (11:16 +0000)]
tests: Added OAuth 2.0 (OAUTHBEARER) tests

9 years agooauth2: Added support for OAUTHBEARER SASL mechanism to IMAP, POP3 and SNMP
Steve Holme [Sat, 5 Sep 2015 17:09:40 +0000 (18:09 +0100)]
oauth2: Added support for OAUTHBEARER SASL mechanism to IMAP, POP3 and SNMP

OAUTHBEARER is now the official "registered" SASL mechanism name for
OAuth 2.0. However, we don't want to drop support for XOAUTH2 as some
servers won't support the new mechanism yet.

9 years agoRELEASE-NOTES: recounted curl_easy_setopt() options
Daniel Stenberg [Fri, 13 Nov 2015 22:40:40 +0000 (23:40 +0100)]
RELEASE-NOTES: recounted curl_easy_setopt() options

9 years agotypecheck-gcc.h: add missing slist-using options
Daniel Stenberg [Fri, 13 Nov 2015 22:19:19 +0000 (23:19 +0100)]
typecheck-gcc.h: add missing slist-using options

CURLOPT_RESOLVE and CURLOPT_PROXYHEADER were missing

Also sorted the list.

9 years agotypecheck-gcc.h: added CURLOPT_CLOSESOCKETDATA
Daniel Stenberg [Fri, 13 Nov 2015 22:12:18 +0000 (23:12 +0100)]
typecheck-gcc.h: added CURLOPT_CLOSESOCKETDATA

... and sorted curl_is_cb_data_option alphabetically

9 years agoopenssl: Free modules on cleanup
Sebastian Pohlschmidt [Fri, 13 Nov 2015 21:07:11 +0000 (16:07 -0500)]
openssl: Free modules on cleanup

Curl_ossl_init calls OPENSSL_load_builtin_modules() but
Curl_ossl_cleanup doesn't make a call to free these modules.

Bug: https://github.com/bagder/curl/issues/526

9 years agosymbols-in-versions: Added new CURLOPTTYPE_STRINGPOINT alias
Steve Holme [Fri, 13 Nov 2015 20:42:35 +0000 (20:42 +0000)]
symbols-in-versions: Added new CURLOPTTYPE_STRINGPOINT alias

...following commit aba281e762 to fix test 1119.

9 years agocurl: mark two more options strings for --libcurl output
Daniel Stenberg [Fri, 13 Nov 2015 15:59:09 +0000 (16:59 +0100)]
curl: mark two more options strings for --libcurl output

9 years agotypecheck-gcc.h: add some missing string types
Daniel Stenberg [Fri, 13 Nov 2015 15:53:17 +0000 (16:53 +0100)]
typecheck-gcc.h: add some missing string types

Also sorted that list alphabetically

9 years agocurl.h: introducing the STRINGPOINT alias
Daniel Stenberg [Fri, 13 Nov 2015 15:50:10 +0000 (16:50 +0100)]
curl.h: introducing the STRINGPOINT alias

As an alias for OBJECTPOINT. Provided to allow us to grep for all string
options easier.

9 years agocleanup: general removal of TODO (and similar) comments
Daniel Stenberg [Fri, 13 Nov 2015 15:15:26 +0000 (16:15 +0100)]
cleanup: general removal of TODO (and similar) comments

They tend to never get updated anyway so they're frequently inaccurate
and we never go back to revisit them anyway. We document issues to work
on properly in KNOWN_BUGS and TODO instead.

9 years agoftplistparser: remove empty function
Daniel Stenberg [Fri, 13 Nov 2015 15:12:18 +0000 (16:12 +0100)]
ftplistparser: remove empty function

9 years agoopenssl: remove #if check for 0.9.7 for ENGINE_load_private_key
Daniel Stenberg [Fri, 13 Nov 2015 14:55:52 +0000 (15:55 +0100)]
openssl: remove #if check for 0.9.7 for ENGINE_load_private_key

9 years agoopenssl: all supported versions have X509_STORE_set_flags
Daniel Stenberg [Fri, 13 Nov 2015 14:52:39 +0000 (15:52 +0100)]
openssl: all supported versions have X509_STORE_set_flags

Simplify by removing #ifdefs and macros

9 years agoopenssl: remove 0.9.3 check
Daniel Stenberg [Fri, 13 Nov 2015 14:51:46 +0000 (15:51 +0100)]
openssl: remove 0.9.3 check

9 years agoopenssl: remove #ifdefs for < 0.9.5 support
Daniel Stenberg [Fri, 13 Nov 2015 14:49:49 +0000 (15:49 +0100)]
openssl: remove #ifdefs for < 0.9.5 support

We only support >= 0.9.7

9 years agolib/vtls/openssl: remove unused traces of yassl ifdefs
Daniel Stenberg [Fri, 13 Nov 2015 14:48:04 +0000 (15:48 +0100)]
lib/vtls/openssl: remove unused traces of yassl ifdefs

9 years agounit1603: Demote hash mismatch failure to a warning
dfandrich [Thu, 12 Nov 2015 22:45:56 +0000 (23:45 +0100)]
unit1603: Demote hash mismatch failure to a warning

The hashes can vary between architectures (e.g. Sparc differs from x86_64).
This is not a fatal problem but just reduces the coverage of these white-box
tests, as the assumptions about into which hash bucket each key falls are no
longer valid.

9 years agounit1603: Added unit tests for hash functions
dfandrich [Thu, 12 Nov 2015 19:40:26 +0000 (20:40 +0100)]
unit1603: Added unit tests for hash functions

9 years agounit1602: Fixed failure in torture test
dfandrich [Thu, 12 Nov 2015 18:30:59 +0000 (19:30 +0100)]
unit1602: Fixed failure in torture test

9 years agosasl: Re-introduced XOAUTH2 in the default enabled authentication mechanism
Steve Holme [Thu, 12 Nov 2015 19:45:24 +0000 (19:45 +0000)]
sasl: Re-introduced XOAUTH2 in the default enabled authentication mechanism

Following the fix in commit d6d58dd558 it is necessary to re-introduce
XOAUTH2 in the default enabled authentication mechanism, which was
removed in commit 7b2012f262, otherwise users will have to specify
AUTH=XOAUTH2 in the URL.

Note: OAuth 2.0 will only be used when the bearer is specified.

9 years agosasl_sspi: fix identity memory leak in digest authentication
Stefan Bühler [Thu, 12 Nov 2015 10:10:32 +0000 (11:10 +0100)]
sasl_sspi: fix identity memory leak in digest authentication

9 years agosasl_sspi: fixed unicode build for digest authentication
Stefan Bühler [Thu, 12 Nov 2015 10:09:21 +0000 (11:09 +0100)]
sasl_sspi: fixed unicode build for digest authentication

Closes #525

9 years agooauth2: Re-factored OAuth 2.0 state variable
Steve Holme [Sat, 5 Sep 2015 16:32:37 +0000 (17:32 +0100)]
oauth2: Re-factored OAuth 2.0 state variable

9 years agosasl: Don't choose OAuth 2.0 if mechanism not advertised
Steve Holme [Sat, 5 Sep 2015 16:29:49 +0000 (17:29 +0100)]
sasl: Don't choose OAuth 2.0 if mechanism not advertised

Regression from commit 9e8ced9890 which meant if --oauth2-bearer was
specified but the SASL mechanism wasn't supported by the server then
the mechanism would be chosen.

9 years agoruntests: more compact "System characteristics" output
Daniel Stenberg [Thu, 12 Nov 2015 14:32:48 +0000 (15:32 +0100)]
runtests: more compact "System characteristics" output

- no point in repeating curl features that is already listed as features
  from the curl -V output

- remove the port numbers/unix domain path from the output unless
  verbose is used, as that is rarely interesting to users.

9 years agoruntests: rename conditional curl-features to $has_[name]
Daniel Stenberg [Thu, 12 Nov 2015 13:50:13 +0000 (14:50 +0100)]
runtests: rename conditional curl-features to $has_[name]

9 years agooauth2: Introduced support for host and port details
Steve Holme [Sat, 5 Sep 2015 16:08:54 +0000 (17:08 +0100)]
oauth2: Introduced support for host and port details

Added support to the OAuth 2.0 message function for host and port, in
order to accommodate the official OAUTHBEARER SASL mechanism which is
to be added shortly.