On Fri, Feb 06, 2015 at 02:40:52PM +0100, Daniel Stenberg wrote:
On Fri, 6 Feb 2015, Daniel Stenberg wrote:
I made the MD4_* and MD5_* functions in libcurl static now and pushed. I
think that fixes this problem.
Hehe, and when I did that it struck me that we got this problem because
axTLS
On Fri, Feb 06, 2015 at 01:12:51PM -0500, Jorge Hernandez wrote:
Hey guys,
What can cause this and how can I fix it, I can see they are all
related to OpenLDAP, which I have version 2.4.40 of:
=
../lib/.libs/libcurl.so: undefined reference to `ldap_msgtype@OPENLDAP_2.4_2'
On Thu, Feb 05, 2015 at 02:32:47PM -0500, Sachin Nikumbh wrote:
Hello,
I am building curl with openssl. I am using the following configuration
options:
--wish-ssl=PATH
--enable-shared
--enable-static
As a result, I am getting both static and dynamic libraries linked against
openssl.
The new md5 code added in the last 3 days has caused builds using axTLS to
segfault. Here's a trace of curl from git HEAD run against the test suite
server under valgrind using axTLS 1.4.9 on x86 Linux:
$ LD_LIBRARY_PATH=lib/.libs valgrind --num-callers=16 src/.libs/curl --max-time
13 --output
On Mon, Feb 02, 2015 at 05:24:30PM -0600, Craig A. Berry wrote:
On Feb 2, 2015, at 4:37 PM, Daniel Stenberg dan...@haxx.se wrote:
2. The requirement to always add new ones last in this generated list will
be seriously awkward: right now multi.h is scanned last in the test (and I
On Fri, Jan 23, 2015 at 12:26:51PM +0100, Daniel Stenberg wrote:
On Fri, 23 Jan 2015, Daniel Stenberg wrote:
I'll get the libressl build done as well.
Test 2025 blew up in my face when trying this. I've had troubles with that
test using normal OpenSSL before and now with libressl my test
On Wed, Jan 21, 2015 at 03:24:58PM +0530, Darshan Ghumare wrote:
I am using curl's latest version (curl-7.40.0).
I am getting SIGSEGV at transfer.c:1075 when I am trying to put (HTTP PUT
Every time?
request) a file to swift.
I don't know what swift is.
Following is the stack (Here, http
On Thu, Jan 15, 2015 at 05:29:56PM -0700, Michael Petch wrote:
As for OS/X you can build a variant with Macports that relies on GNUTLS
rather than OpenSSL. I haven't tested it, but I'll lay odds it works.
But I won't know until I get around to doing the next Mac builds.
If these are the only
On Thu, Jan 15, 2015 at 10:55:12AM +0530, Heena Sirwani wrote:
I am using libcurl in C with the mutli socket interface along with libuv. I
add
easy handles to the multi handle for each request and start the libuv event
loop together with adding sockets for polling for appropriate activity to
On Wed, Jan 14, 2015 at 05:59:00PM +0100, Daniel Stenberg wrote:
On Tue, 13 Jan 2015, Dan Fandrich wrote:
The interpretation of these bits as a signed or unsigned number will
differ, but since libcurl treats the result as a raw set of bits, it
doesn't really matter.
That's also why
On Tue, Jan 13, 2015 at 04:27:13PM -0700, Michael Petch wrote:
On 2015-01-13 2:47 PM, Dan Fandrich wrote:
My view is only that the OpenSSL license
is a non-issue when developing an app using libcurl.
This isn't my view. Hypothetically assume we use OpenSSL on the back end
of libcurl
On Tue, Jan 13, 2015 at 11:17:20AM -0700, Michael Petch wrote:
The downside is that we add some more dependencies to GNUbg. We'll
If you want https support without rolling your own crypto, there's no way
around another dependency. Even if that dependency is normally supplied with
the OS. And
On Tue, Jan 13, 2015 at 01:21:56PM -0700, Michael Petch wrote:
On 2015-01-13 1:11 PM, Dan Fandrich wrote:
An end
user compiling his own gnubg doesn't even need to concern himself with the
license of the SSL library (as long as he doesn't redistribute the code)
while
packagers can
On Tue, Jan 13, 2015 at 11:49:44AM -0200, Jonathan C. wrote:
Currently it's 1 31 (2147483648), which is greater than the biggest 32 bit
signed int. I'm asking that because javascript will cast all of its double
values into signed 32-bit integers to do the bitwise operation, causing given
On Fri, Jan 02, 2015 at 02:45:32AM -0500, Ray Satiro wrote:
CURL_SSLVERSION_SSLv3 is exclusively SSLv3 [1]. There is no longer a define
for
SSLv3 or later. The default is now TLS 1.0 or later. I have attached a patch
that adds legacy support for CURL_SSLVERSION_SSLv3_OR_LATER to all backends;
On Tue, Dec 30, 2014 at 02:58:52PM +, Jones, Rick wrote:
I have packaged cURL for my enterprise developers to use when needed but am
having some issues.
The installation part is easy, but I need to modify my package due to a
configuration issue. Please let me explain.
We are in a
On Mon, Dec 15, 2014 at 12:27:24PM -0800, Jeroen Ooms wrote:
On Mon, Dec 15, 2014 at 3:34 AM, Gisle Vanem gva...@yahoo.no wrote:
Jeroen Ooms wrote:
#define MAKE_OPTION(a) {#a, CURLOPT_##a}
typedef struct {
char name[40];
int val;
} keyval;
keyval curl_options[] = {
On Wed, Dec 10, 2014 at 10:52:46PM -0800, Jeroen Ooms wrote:
I'm writing some bindings to libcurl for a scripting language. I was
wondering if there is an API call or some other method to list the
options (CURLOPT_* values and their integer) that the linked version
of libcurl supports?
That
On Thu, Dec 11, 2014 at 07:39:54PM +, Steve Holme wrote:
My only objection is that we are then taking away the ability to allow the
user to purposely send LF characters to the mail server for whatever reason
- they may have a non RFC compliant mail server that requires the line
ending to
On Mon, Dec 08, 2014 at 10:46:44PM -0500, John Coffey wrote:
after much debugging and sprinkling of printouts on our target platform, it
turns out that the source of the bug was 75% an application problem (mine) and
25% (in my opinion) an libCurl issue due to the weakness of using loosely
On Thu, Nov 27, 2014 at 11:59:15PM +0100, Peter Wu wrote:
This patch series introduces support for performing requests over UNIX
domain sockets. The main use case is debugging nginx servers that listen
on said sockets. For that goal, two tests have been added.
This patch series has broken
On Thu, Dec 04, 2014 at 08:36:06PM +0100, Niels Larsen wrote:
Libssh2 people,
On stock Ubuntu 14.04 (Linux 64 bit) I get:
~/BION/Software/Package_sources/Utilities/libssh2-1.4.3 ./configure
checking whether to enable maintainer-specific portions of Makefiles...
no
checking for sed...
On Wed, Dec 03, 2014 at 02:56:34PM -0500, Dave Reisner wrote:
---
docs/curl.1 | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
I've pushed this now. Thanks for the patch!
Dan
---
List admin:
On Tue, Dec 02, 2014 at 02:57:53PM +0100, Guenter wrote:
Hi all,
its now more than 2 years since the release of 1.4.3, and since then
a couple of things were fixed or new introduced; last commits were
about 3 months ago, so things should have settled ...
anything against a soonish release?
On Mon, Dec 01, 2014 at 12:05:46PM +0530, Mohanraj V wrote:
Hi,
I have following questions regarding lib-curl. Please answer my questions.
1. How to reuse connections using multi-handle and multi-perform API ?
I tried with easy handle by continuous calling of
curl_easy_perform. So
On Mon, Dec 01, 2014 at 03:46:20PM +0530, Arif Ali wrote:
Hello there,
This is in response to
http://curl.haxx.se/mail/lib-2014-08/0282.html
If I do not want to build and use only binaries.
How do you run your software then if not in binary form?
Is there any set easy
On Fri, Nov 28, 2014 at 10:20:17PM +0530, Praveen Pvs wrote:
Can I use CURL in multi threaded application?
Yes, but you need to read http://curl.haxx.se/libcurl/features.html#thread and
the links from there. Running into problems under a multithreaded environment
that don't show up in a single
On Tue, Nov 25, 2014 at 11:55:10AM -0600, John E. Malmberg wrote:
In building the Nov 24 daily snapshot on VMS, I am getting the
following warnings because a const pointer is being assigned to a
member of a structure that is not const.
Generally with DECC, a cast of (char *) is needed here,
On Tue, Nov 25, 2014 at 11:55:25AM -0600, John E. Malmberg wrote:
Building on VMS currently requires having the lib/curl_config.h.in
and the configure because the tools to generate them are not yet
working on VMS and will not be in the near future.
Is there anyway to pull a git branch with
On Tue, Nov 18, 2014 at 10:57:57AM -0800, bch wrote:
Inspired by one of Daniels recent podcasts. Note that I did *NOT*
Thanks for sending in a patch!
patch the .html, because it looks like it's generated by a roff-tool
of Daniels.
Correct; the web version is built from the troff source in
On Tue, Nov 18, 2014 at 12:00:51PM -0800, bch wrote:
On 11/18/14, Dan Fandrich d...@coneharvesters.com wrote:
I'm not sure I agree with this change. Adding By default implies that
there
are some conditions under which the data IS copied by the library, but there
aren't. CURLOPT_POSTFIELDS
On Mon, Nov 17, 2014 at 06:06:26PM +0100, Peter Wu wrote:
test916 would be detect as fail when ran under gdb. Fix this by
returning the child exit code from gdb rather than assuming 0.
Signed-off-by: Peter Wu pe...@lekensteyn.nl
---
tests/runtests.pl | 10 ++
1 file changed, 6
On Mon, Nov 17, 2014 at 06:06:23PM +0100, Peter Wu wrote:
This is the second revision of the earlier patchset[1] which fixes tests
which broke when ran under LD_PRELOAD=libhostname.so with ASAN enabled
(cmake without CURLDEBUG). To facilitate that change, it was necessary
to change from a
On Sun, Nov 16, 2014 at 12:19:26AM +, Steve Holme wrote:
However, if we were to remove it (as per my patch) is that such a hardship
especially as krb4 has gone - never (probably) to return? At the end of the
day the output you mentioned is purely textual and it saves having an usused
On Sat, Nov 15, 2014 at 03:43:09PM +, Steve Holme wrote:
The curl command line tool supports --krb LEVEL and in its usages sets the
CURLOPT_KRBLEVEL option in libcurl - but only if KRB4 support has been
detected if(curlinfo-features CURL_VERSION_KERBEROS4) in
tool_getparam.c:696.
I
On Sat, Nov 15, 2014 at 05:57:19PM +, Steve Holme wrote:
On Sat, 15 Nov 2014, Dan Fandrich wrote:
I have prepared a patch to remove this (see attached), however, from
reading the libcurl code (security.c) and associated comments it seems
more of a generic Kerberos option. Does
On Sat, Nov 15, 2014 at 10:37:11PM +, Steve Holme wrote:
On Sat, 15 Nov 2014, Dan Fandrich wrote:
From a curl command line point of view my patch really doesn't do anything
different as the current code checks for the presence of
CURL_VERSION_KERBEROS4 which won't be there (when
On Sat, Nov 15, 2014 at 09:25:32PM +0100, Michael Osipov wrote:
There is no such thing as --krb 5. If you take a look at the man
page, you'll see that you are allowed to pass: The level must be
entered and should be one of 'clear', 'safe', 'confidential', or
'private'.
Ok, it doesn't matter
On Thu, Nov 13, 2014 at 03:19:17PM -0800, Mark Juric wrote:
Hello all,
I have an issue with the curl plugin. I'm collecting at 60 second intervals
from a web page on localhost, and it appears the connection is being held open
in between queries. For most situations this is okay, but I have a
On Thu, Nov 06, 2014 at 09:09:40PM +, Steve Holme wrote:
On Thu, 6 Nov 2014, GitHub wrote:
Added !SSPI to the features list of the HTTP digest tests, as SSPI
based builds now use the Windows SSPI messaging API rather than the
internal functions, and we can't control the random numbers
On Tue, Nov 04, 2014 at 03:09:46PM -0800, David J Ryan wrote:
I have a multi-threaded application that appears to have no problem FTPing
files to a large number of QNX boxes however I am getting a segmentation fault
that appears to show the following call stack. I'm not sure how to handle
Package: metastore
Version: 1+20080623+debian-4
The Debian copyright file claims:
License: GPL-2+
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version
On Wed, Oct 22, 2014 at 03:41:37PM +0200, Guenter wrote:
Hi all,
due to failure of one of my autobuilds [1] I wonder if we have
somewhere a document which lists the minimal versions of
dependencies?
Yes: docs/INTERNALS But it looks like it's not quite up-to-date.
I see few things mentioned
On Thu, Oct 16, 2014 at 07:30:39AM +, Bruno Thomsen wrote:
From a security aspect SSLv3 should be dropped completely due to its many
weaknesses.
I think it would be a good move to follow in the footsteps of libressl.
Legacy systems are most likely also using an old version of curl.
On Thu, Oct 16, 2014 at 10:16:03AM +0200, Kamil Dudka wrote:
I agree that SSLv3 should be disabled by default but I see no point in hiding
SSLv3 behind a compile-time option, or creating new run-time options to
enable
it. There already are (lib)curl options to require SSLv3 explicitly.
lmms 1.0.3 sets the -Werror flag by default which turns every compile warning,
no matter how innocuous, into a compile failure. I can see the value of this
flag to developers if the goal is to provide a warning-free build, but it's
incredibly annoying to users/packagers who just want to build the
On Tue, Oct 07, 2014 at 08:17:03PM +0200, Daniel Stenberg wrote:
On Tue, 7 Oct 2014, TJ Saunders wrote:
That part is understandable, but why do these #defines include a
trailing space character in the format specification?
Blah, I didn't even notice that because I was too busy looking at
On Sat, Oct 04, 2014 at 04:55:40PM -0400, K. Frank wrote:
Dan Fandrich commented in the thread:
using libcurl in an enterprise application server
as follows:
As documented, curl_global_init may not be called
from a thread but from the main function before any
threads
On Fri, Oct 03, 2014 at 05:27:42AM -0400, Dan Donahue wrote:
As stated in my original mail, this issue first manifested using the
default write callback in libcurl (I did not register one). During debug
I did register a callback but the problem persisted.
The set up of the file handle is in
On Thu, Oct 02, 2014 at 08:32:14AM +0200, Romain Rivière wrote:
On 02/10/2014 00:21, Dan Fandrich wrote:
Have you tried using pkg-config? That is the most reliable method. Just
set PKG_CONFIG_PATH to the right location for your OpenSSL installation's
openssl.pc file and it should just work
On Thu, Oct 02, 2014 at 02:28:54PM -0400, Michael Mueller wrote:
Summary of problem : libcurl config.log shows openssl 1.0.1i is used;
binaries produced by make and make install are using much older
openssl libraries
The curl build compiles against the special OpenSSL library you provided
at
On Thu, Oct 02, 2014 at 02:42:58PM -0400, Dan Donahue wrote:
This is my first post to the libcurl mailing list. My hope is this mail
is etiquettely correct according to your mailing list rules, which I
read in their entirety.
Bonus points!
Please review the info below and provide feedback
On Wed, Oct 01, 2014 at 05:16:50PM -0400, Michael Mueller wrote:
On Wed, Oct 1, 2014 at 1:07 PM, Romain Rivière lecoy...@lecoyote.org wrote:
As you can see from the second line, the path to Openssl is not the
system path: it is Optware's staging dir, and libopenssl is there:
-rw-r--r-- 1
On Tue, Sep 23, 2014 at 11:30:19AM +0100, KALLEL Mohamed wrote:
In fact I do not have a big idea about http protocol neither about
http over udp cases. But I see this request in the standard protocol
TR-069 cwmp (Amendment 5) and in this standard they require to send
HTTP messages over UDP and
On Tue, Sep 09, 2014 at 09:31:16PM -0400, Keith Broerman wrote:
I'm trying to use libcurl to build a HTTP client that subscribes one time to
an
event service and waits to receive events on a persistent connection. I'm
starting from the sendrecv.c example code. My libcurl-based client
On Thu, Sep 04, 2014 at 09:37:50AM +0200, Daniel Stenberg wrote:
On Thu, 4 Sep 2014, Dan Fandrich wrote:
When I reviewed the security vulnerabilities associated with version
7.21.4, I
7.21.4 isn't a released curl version.
Sorry, but I felt I had to correct you here. 7.21.4 was a real
On Tue, Sep 02, 2014 at 03:12:45PM -0400, Cathey, Jim wrote:
This is/was part of a proprietary system, but I do know that it
kept track of _when_ a time correction was applied, and how much
was applied, from which it extrapolated the amount of drift the
clock would have at any point. In
On Thu, Aug 21, 2014 at 10:34:37AM -0400, dev wrote:
First thing, after the zeroth thing which is the lib1513, is that I see
a row of yellow text for all manner of protocols. At the top of the
page I see :
Text looking like this is for servers that couldn't be started
for the test
On Fri, Aug 08, 2014 at 07:00:48PM +0200, Pierre Joye wrote:
On Aug 7, 2014 4:17 PM, clin...@elemtech.com wrote:
What encoding does libcurl expect from the caller when setting the
CURLOPT_URL option? UTF-8 or the Windows ANSI code page?
The answer to that would help make a correct patch.
On Mon, Aug 04, 2014 at 09:40:26AM +0200, Michael Osipov wrote:
I am aware of that wrapper from SAP because this is at our company's SAP core
but I did not know that is actually open source. It is a tremendous amount of
work to bridge both as they have different approaches in conventions. I am
On Sun, Jul 27, 2014 at 01:32:51PM +0200, Fabian Keil wrote:
Dan Fandrich d...@coneharvesters.com wrote:
On Fri, Jul 04, 2014 at 03:38:39PM +0200, Fabian Keil wrote:
Subject: [PATCH 1/6] Allow to overwrite $TESTDIR through the environment
This one seems like it could be useful. I
On Wed, Aug 06, 2014 at 12:24:45PM +0530, shine wrote:
I am using curl 0.7.24 library and it is got stuck sometimes during
Do you mean version 7.24.0? That versions is already 2½ years old and has had
many bug fixes applied in that time. Have you tried the latest version?
Un-initialization.
On Wed, Aug 06, 2014 at 12:40:54AM +0530, Girish Aher wrote:
I am using libcurl in a multi-threaded program wherein each thread has its own
curl session handle as per the documentation. All the sessions talk to the
same
destination web server to upload data over http and they pass through a
On Sun, Aug 03, 2014 at 11:02:21PM +0200, Michael Osipov wrote:
On Sun, Aug 03, 2014 at 08:53:55PM +0200, Michael Osipov wrote:
Am 2014-08-03 um 11:27 schrieb Dan Fandrich:
On Sun, Aug 03, 2014 at 10:50:21AM +0200, Michael Osipov wrote:
Am 2014-08-03 um 10:27 schrieb Dan Fandrich
On Sat, Aug 02, 2014 at 02:18:29PM +, Michael Osipov wrote:
@@ -180,7 +180,8 @@ FOOTNOTES
*1 = requires OpenSSL, GnuTLS, NSS, yassl, axTLS, PolarSSL, WinSSL (native
Windows), Secure Transport (native iOS/OS X) or qssl (native IBM i)
*2 = requires OpenLDAP
- *3 = requires a
On Sun, Aug 03, 2014 at 10:50:21AM +0200, Michael Osipov wrote:
Am 2014-08-03 um 10:27 schrieb Dan Fandrich:
On Sat, Aug 02, 2014 at 02:18:29PM +, Michael Osipov wrote:
@@ -180,7 +180,8 @@ FOOTNOTES
*1 = requires OpenSSL, GnuTLS, NSS, yassl, axTLS, PolarSSL, WinSSL
(native
On Sun, Aug 03, 2014 at 08:53:55PM +0200, Michael Osipov wrote:
Am 2014-08-03 um 11:27 schrieb Dan Fandrich:
On Sun, Aug 03, 2014 at 10:50:21AM +0200, Michael Osipov wrote:
Am 2014-08-03 um 10:27 schrieb Dan Fandrich:
On Sat, Aug 02, 2014 at 02:18:29PM +, Michael Osipov wrote:
@@ -180,7
On Wed, Jul 30, 2014 at 12:12:12AM +0200, Daniel Stenberg wrote:
This action is made to highlight the need for attention and to stress
that we really can use good random, even for a few non-TLS features
and while we have a more stupid default implemenation for when
there's really no TLS
On Mon, Jul 28, 2014 at 06:35:44PM +, Valluri, Sathish wrote:
I have implemented SFTP transport using CURL, which in turn uses libssh2.
The following are the corresponding versions of libraries that I am using
Openssl 1.0.0-20 FIPS certified
Libssh2 1.4.3
Curl 7.19
I want know
On Sat, Jul 26, 2014 at 12:36:23PM -0700, Thomas Dineen wrote:
The code shown below works about 98% of the time!
About 2% of the time it will hang in url_fgets() maybe for 10
minutes or
more and sometimes longer.
if the read fails that is ok, I just need it to return in an
Hi, Fabian. I've applied a number of this batch of patches (the short,
trivial ones). That's not to say that the others don't have value, just that
I haven't thought about them much, noone else has chimed in one way or the
other, and they touch thousands of files and it's too late in the day for
On Fri, Jul 25, 2014 at 09:52:11AM -0400, Jon Torrey wrote:
I am interested in cross-compiling on Linux for ARM INTEGRITY as it seems most
similar to what I am running.
The example provided within install.html leaves me a tad confused on how to
get
prepared. I've done a file grep inside
On Thu, Jul 24, 2014 at 03:20:02PM +, Valluri, Sathish wrote:
We are facing issue in curl sftp file transfer when FIPS mode is enabled in
the
openssl-1.0.0-20 version.
Is libssh2 using the FIPS version of OpenSSL as well?
* Hostname was NOT found in DNS cache
* Trying
On Thu, Jul 24, 2014 at 09:39:00AM -0400, Jon Torrey wrote:
I am currently attempting to port over libcurl to an ARM-based embedded system
with a proprietary operating system. After practicing building on both Linux
Sounds like fun!
and Windows and capturing the output of the process in text
On Sun, Jul 20, 2014 at 11:20:13AM +0300, Timo Teras wrote:
On Sat, 19 Jul 2014 21:31:33 +0200
Dan Fandrich d...@coneharvesters.com wrote:
On Fri, Jul 18, 2014 at 10:07:06PM +0300, Timo Teräs wrote:
and do it in smaller code:
diff --git a/procps/top.c b/procps/top.c
index
I looked into why test 310 is failing to run on a CyaSSL build with a
failed starting HTTPS server (stunnel) error. The answer is that the server
verification is failing with
* SSL_connect failed with error -283: Key Use digitalSignature not set Error
CyaSSL doesn't like that the certificate
On Mon, Jul 21, 2014 at 06:24:14PM +0200, Stéphane Graber wrote:
On Mon, Jul 21, 2014 at 04:14:43PM +, Serge Hallyn wrote:
Quoting Stéphane Graber (stgra...@ubuntu.com):
This forces wget to retry if it gets a network error.
Signed-off-by: Stéphane Graber stgra...@ubuntu.com
I
On Sun, Jul 20, 2014 at 12:38:59PM -0400, dev wrote:
On July 14, 2014 at 1:55 PM Daniel Stenberg dan...@haxx.se wrote:
On Mon, 14 Jul 2014, Dan Fandrich wrote:
I have a patch ready to to set _POSIX_PTHREAD_SEMANTICS for Solaris
very
similar to Mischa Salle's suggestion, and his
On Sat, Jul 19, 2014 at 05:02:01PM +0200, Alessandro Ghedini wrote:
Just a small typo I noticed (transfered - transferred).
Thanks! I've just committed it.
Dan
---
List admin: http://cool.haxx.se/list/listinfo/curl-library
On Tue, Jul 15, 2014 at 08:05:13PM +0100, Steve Holme wrote:
I don't know whether it is worth fixing these up before the release but I
noticed that the number of warnings in the OpenSuSE autobuilds has increased
from 18 to 30 at some point over the last few weeks.
It seems that they are
On Thu, Jul 17, 2014 at 01:01:23PM +0200, Daniel Stenberg wrote:
On Thu, 17 Jul 2014, Alessandro Ghedini wrote:
Is it just me or the announce for release 7.37.1 was not sent to
this mailing list? I only noticed because of the GitHub tag.
There's something wrong with the mail server atm
On Sun, Jul 13, 2014 at 03:05:04PM +0200, Tor Arntsen wrote:
On 13 July 2014 09:33, Dan Fandrich d...@coneharvesters.com wrote:
The Solaris builds failed overnight because their default getpwuid_r uses
four
arguments instead of the POSIX standard 5. They do provide a
POSIX-conformant
On Mon, Jul 14, 2014 at 07:18:22PM +0200, Daniel Stenberg wrote:
On Mon, 14 Jul 2014, Dan Fandrich wrote:
Problem is, this change broke curl compiles out-of-the-box for
Solaris. We need to do something by the time of this release.
My gut says we back out that particular commit
On Sun, Jul 13, 2014 at 07:14:33AM +0100, David Woodhouse wrote:
On Sun, 2014-07-13 at 01:09 +0200, Dan Fandrich wrote:
On Sat, Jul 12, 2014 at 05:59:56PM +0100, David Woodhouse wrote:
The cipher list problem was because Fedora's GnuTLS doesn't have SRP
support. Given
On Mon, Jul 14, 2014 at 09:12:41PM +0200, Daniel Stenberg wrote:
On Sun, 13 Jul 2014, David Woodhouse wrote:
This is rather unfortunate. I'll improve the error message as you
suggest, but I wonder what the best way is to determine whether
SRP is supported or not. Is there a compile-time
On Mon, Jul 14, 2014 at 11:45:06AM +0200, Tor Arntsen wrote:
I did a very quick scan of the Solaris header files,
_POSIX_PTHREAD_SEMANTICS has a lot of overlap with _POSIX_C_SOURCE
but the latter definitely has more impact. From what I could gather
with the minutes I had available this morning
The Solaris builds failed overnight because their default getpwuid_r uses four
arguments instead of the POSIX standard 5. They do provide a POSIX-conformant
version if you compile with the macro _POSIX_PTHREAD_SEMANTICS set. We could do
a getnameinfo-style configure check on the number of
On Sat, Jul 12, 2014 at 11:22:38AM +0100, David Woodhouse wrote:
I note that for finding the home directory in both lib/netrc.c and
src/tool_homedir.c we use $HOME *before* getpwuid(). And we actually use
That adds flexibility to the user who can switch the location easily without
changing
On Sat, Jul 12, 2014 at 05:59:56PM +0100, David Woodhouse wrote:
The cipher list problem was because Fedora's GnuTLS doesn't have SRP
support. Given that gnutls_set_priority_direct() actually *gives* us a
pointer to the part of the string that it objected to, our error
handling could stand to
On Fri, Jul 11, 2014 at 08:38:00PM +, Alona Rossen wrote:
Our C++ application uses libcurl 7.21.0 linked to openssl-1.0.0m to implement
a
variety of functionalities, including SMTP mail transfer.
The application runs on numerous platforms of different versions, including
Windows,
On Fri, Jul 11, 2014 at 11:36:46AM +0100, David Woodhouse wrote:
From: David Woodhouse david.woodho...@intel.com
---
lib/curl_ntlm_wb.c | 15 +++
1 file changed, 15 insertions(+)
diff --git a/lib/curl_ntlm_wb.c b/lib/curl_ntlm_wb.c
index 52d1323..ac05fbb 100644
---
On Thu, Jul 10, 2014 at 10:17:42AM -0400, Glen A Johnson Jr. wrote:
The patch you suggest would allow the code I’m working on to function however
if you consider the case where use_netrc is set to CURL_NETRC_REQUIRED this
would lead to a successful return from override_login if home is not
On Wed, Jul 09, 2014 at 08:22:42PM +0200, Michael Osipov wrote:
I have several patches pending since end of May no one has taken any
action to review them at least.
I think a big part of the problem is that lots of people are taking summer
vacations at the moment, and fewer people want to code.
On Wed, Jul 09, 2014 at 03:34:28PM -0400, Richard Gray wrote:
Taking the hint from above, I re-ran using:
CFLAGS=-D_XOPEN_SOURCE_EXTENDED ./configure and I seem to have
gotten a working curl build. I wonder if configure should always
set this flag for HP-UX systems? It (or something
On Wed, Jul 09, 2014 at 03:25:12PM -0400, Glen A Johnson Jr. wrote:
I ran into this issue using curl 7.36.0 and also confirmed with curl 7.37.0.
The issue is that when use_netrc is set to CURL_NETRC_OPTIONAL that
CURLE_OUT_OF_MEMORY is returned from override_login if the netrc file is not
On Mon, Jul 07, 2014 at 05:48:46PM -0400, Eric Fortin wrote:
On Jun 26, 2014, at 7:19 AM, Sandeep Sastry sandeep.a.sas...@gmail.com
wrote:
Though there are some existing discussion related to the topic
Content-Length Vs Chunked Transfer Encoding on various non-curl sites, I
wanted to
On Thu, Jul 03, 2014 at 02:45:04PM +, Rick Breen wrote:
First, thank you (all) for your work and effort in supplying the world with
very functional open source code.
This is probably a newb question, but I'm in the thick of porting code and
bringing-up an Android-based Freescale iMX6
On Thu, Jul 03, 2014 at 12:12:34AM +0530, Ajay Garg wrote:
On Wed, Jul 2, 2014 at 1:11 AM, Dan Fandrich d...@coneharvesters.com wrote:
If that's followed there will never be an
opportunity to mix up received data since the handle passed in to the
callback
in each thread
On Mon, Jun 30, 2014 at 09:43:53AM -0400, Sachin Nikumbh wrote:
Thanks a lot for the response. I am going to investigate the cause behind
sockets getting into the TIME_WAIT state. But in the mean time, I am trying to
understand how CURLOPT_SOCKOPTFUNCTION and CURLOPT_OPENSOCKETFUNCTION can be
On Tue, Jul 01, 2014 at 09:14:24PM +0530, Ajay Garg wrote:
Is it safe to pass the same callback-function as the parameter to
CURLOPT_WRITEFUNCTION in different threads?
Note that the threads may be highly concurrent, sending and receiving
requests/
responses simultaneously.
Or we run the
401 - 500 of 1406 matches
Mail list logo