Hi Gün,
2009/10/11, Guenter wrote:
isnt the change
http://article.gmane.org/gmane.comp.web.curl.cvs/12383
now obsolete after you changed the type of addrlen with
http://article.gmane.org/gmane.comp.web.curl.cvs/12384
??
Nah! It is still required to silence an HP-UX bogus compiler warning
2009/10/6, Alexey Simak wrote:
In c-ares' setup.h, on 64-bit WIN32 platform, ssize_t will be redefined
as int (previously defined as __int64 in config-win32.h).
Thanks. Fixed now in CVS.
--
-=[Yang]=-
2009/10/1, Joshua Kwan wrote:
I poked around the source tree and noticed that cURL does not even try to
link against iconv unless you manually configure it like that (even though
99% of the time it's going to be in C library anyway.) Why is this?
Not to GPL infect default configuration
2009/9/24, Joshua Kwan wrote:
[...]
I genericized the code in hostthre.c to be compilable on both Windows
and Mac, by using macros to abstract away the competing threading
idioms. (Despite certain wrinkles, they are actually quite similar.)
Taking in account that the next 'C Standard
Hi Simon,
IE6 has a problem rendering http://autobuild.josefsson.org/libssh2/
Table rows with tr bgcolor=lightgray are actually rendered with a
background color which is DarkBlue or navy from here
http://en.wikipedia.org/wiki/Web_colors#HTML_color_names
This makes me nearly blind to table
Hi Kamil,
2009/9/17, Kamil Dudka wrote:
I am experiencing random test-suite failures of SCP/SFTP tests. It can be
reproduced by running the SCP/SFTP test-suite in the loop:
$ while ./runtests.pl -k -p -v SFTP SCP; do echo ok; done
[...]
Any idea what's happening?
Read this thread
Patch just thrown into the daily build arena.
http://cool.haxx.se/cvs.cgi/curl/lib/select.c.diff?r1=1.52r2=1.53
--
-=[Yang]=-
28 Jul 2009 Joshua Kwan http://curl.haxx.se/mail/lib-2009-07/0292.html wrote:
We have a Glib-style event loop implementation into which the
multi API is hooked. Presumably due to the network disconnection,
polling on the FD responsible for name resolution is in a
'POLLERR' state (indicating
Hi,
Win64 target (MingW64) autobuilds done using x86_64-pc-mingw32-gcc
should not be used as a guide to fix nearly anything in the libcurl
code base. These autobuilds should be considered simply an
x86_64-pc-mingw32-gcc compiler exerciser, nearly nothing compiled with
this compiler which needs to
Hi Gün,
2009/7/20, Guenter li...@gknw.net:
Hi, Yang,
Yang Tse schrieb:
Oops, In previous message where I have written 'size of long' I really
wanted to say 'size of size_t'.
x86_64-pc-mingw32-gcc has size of long = 4 and size of size_t = 8.
You really should be more specific with what
Hi Gün,
Relative to the current regression state in which it is impossible to
build libcurl with c-ares and memory tracking enabled, you have
several options...
1) The easy way. Revert the renaming of config.h to curl_config.h and
ares_config.h.
2) Modify c-ares API in order to make c-ares
Hi Gün,
seems that your recent configure changes break all my MingW autobuilds:
Yep I broke'em with last commit. There's no progress without struggle ;-)
I've already committed a follow-up that should fix recently broken builds.
Autobuilds will show if it is good enough now.
Cheers,
--
Hi John,
2009/6/20, John E. Malmberg wrote:
[...]
The build and kitting procedures required an additional 32 files. The
readme. file in that list has been updated for building using the GNV
package.
[...]
Without knowing the real contents and purpose of each new file you
mention I can
2009/6/18, Cogentree CT cogent...@gmail.com:
I am using CURLOPT_RESUME_FROM_LARGE for a file which is more than 4 GB
(4441787445 bytes) in size.
I wanted download to restart; at offset: 4441487245 bytes.
But, whenever I specify curl_easy_setopt(curl, CURLOPT_RESUME_FROM_LARGE,
4441487245L)
2009/6/16, Igor Novoseltsev wrote:
#ifdef _WRS_KERNEL
extern int open (const char *, int, int);
#else
extern int open (const char *, int, ...);
#endif /* _WRS_KERNEL */
Then why doesn't the 2-arg version work just fine?
The answer is: the ccppc
Hi jvp,
When you get that error which are the contents of src\config.h ?
The only Line is:
'#include ../lib/config.dos'
Ok, we had it right in front of our eyes. Look at those nasty single
quotes surrounding the whole line, these are actually inside
src\config.h converting it into
2009/6/10, John E. Malmberg wrote:
This patch against the 7.19.6-20090609 daily snapshot makes things work.
[...]
+++ packages/vms/curlmsg_vms.h Tue Jun 9 01:27:48 2009
[...]
+++ src/main.c Tue Jun 9 23:56:29 2009
[...]
+++ src/os-specific.c Tue Jun 9 01:26:49 2009
[...]
+++
2009/6/10, jvp wrote:
Although it has created the Libcurl.a there are errors during the next
phase.
Hope this is not to long.
---
make.exe[1]: Leaving directory `c:/Curl/lib'
c:/djgpp/bin/make.exe -C src -f Makefile.dj
make.exe[1]:
2009/6/10, John E. Malmberg wrote:
docs/examples/anyauthput.c needs a typedef for intptr_t.
Fixed in CVS, thanks.
--
-=[Yang]=-
2009/6/9, Gregor Jasny wrote:
[...] I'll prepare a visibility patch within the next days for the GCC
compiler.
That's very nice.
Related with all this stuff there is another possibility...
c-ares' configure script could lack something or have a bug somewhere
that only gets triggered when
2009/6/9, jvp wrote:
Now get:
[...]
setup.h:358:21: error: tcp.h: No such file or directory (ENOENT)
file.c:40:24: error: sys/socket.h: No such file or directory (ENOENT)
file.c:52:19: error: netdb.h: No such file or directory (ENOENT)
file.c:55:23: error: arpa/inet.h: No such file or
2009/6/9, jvp wrote:
Indeed progressed to the point where 37 files were compiled till here:
---
gcc -g -gcoff -O2 -I. -I../include -I../lib -IC:/NET/WATT/inc -Wall
-DHAVE_CONFIG_H -o djgpp/content_encoding.o -c
2009/6/3, Igor Novoseltsev wrote:
Please find the attached patch that enables compilation of the libcurl
in VS2005 for Windows Mobile 6 (ARMV4I) platform.
Note it requires the wcecompat library (open source,
http://www.essemer.com.au/windowsce).
Which functions, struct definitions, or
Committed with 'long' typecasts.
Thanks,
--
-=[Yang]=-
Igor,
Disregard my previous message.
After some further adjustments your work has been committed to CVS.
Verify if everything is ok using daily snapshot dated June 5, 2009 or
later from http://cool.haxx.se/curl-daily/
I yet have a couple of comments.
In Makefile.vxworks you should probably
2009/6/1, Igor Novoseltsev wrote:
Please review the patch.
In no particular order...
strerror_r() on VxWorks might have two arguments but the return type
is not the error converted to a string. It is something like OK or
ERROR. You should probably write an VxWorks wrapper for that function.
I
2009/6/1, John E. Malmberg wrote:
The header files are now supplied by the operating system in a text library
independent of the compiler.
Thanks for the explanation and clean-up.
Committed to CVS.
--
-=[Yang]=-
2009/6/1, John E. Malmberg wrote:
I just was bringing this up in case someone was interested in making the
configure scripts more robust.
We are always interested in improving everything. Simply wanted to
make sure we were speaking the same language.
I've just committed a change to CVS that
2009/5/31, John E. Malmberg wrote:
On VMS, sys/types.h and types.h are the same file.
By default, the VMS C compiler ignores the sys/ path. This is the case
for all platforms on VMS that curl is currently being built for.
So there is no need for VMS specific code here.
Is this the same
2009/5/26, Daniel Stenberg wrote:
One of these days I'd like to see us move the c-ares repo off the piggy-
backing of the curl one we use now, and then we'll need to have things
running on their own...
For now, the decoupling of --enable-curldebug from --enable-debug yet
needs some
2009/5/27, Matthew Fincham wrote:
I got the compilation working. Thanks for you the replies.
Maybe you would like to contribute your changes or tips.
--
-=[Yang]=-
2009/5/26, Matthew Fincham wrote:
I am trying to compile version c-ares 1.6.0 on Windows using the Borland
compiler - bcc32 5.6.4 (an old compiler, I know!).
You will need to write a Borland specific makefile 'Makefile.b32'.
Existing makefiles for mingw 'Makefile.m32' and for msvc
Hi,
Would it be possible to somehow have commitdiff's or diffs between
versions show up on the web interface as full-file dif's?
IMO Many times just the colored diff -u is less usefull than the whole
file diff.
Or is this standard when using git?
Thanks,
--
-=[Yang]=-
Hi Peter,
2009/5/22, Peter Stuge wrote:
Yang Tse wrote:
Would it be possible to somehow have commitdiff's or diffs between
versions show up on the web interface as full-file dif's?
I'm sorry, but I'm not sure I understand exactly what you are
requesting here. Can you please clarify
2009/5/19, Vlad Dinulescu wrote:
Yes, you are right. So just linking against iphlpapi.lib seems to be
the best way after all.
Dynamic linking of iphlpapi allows the use of any version of PSDK or
WSDK to build c-ares and result in a library that works all the way up
from Win95 to the very last
2009/5/13, Daniel Stenberg wrote:
(http://support.microsoft.com/default.aspx?scid=kb;en-us;836769)
Quoting the above link...
This problem may occur when the following conditions are both true:
Your program uses LoadLibrary calls and FreeLibrary calls to load
and to unload the Iphlpapi.dll
2009/5/12, Daniel Stenberg wrote:
Blurgh. Sometimes I wish I had a mac...
And more time to sleep ;-)
The ftp involved change is the one resumed by this two diffs:
http://cool.haxx.se/cvs.cgi/curl/lib/ftp.c.diff?r1=1.511r2=1.513
http://cool.haxx.se/cvs.cgi/curl/lib/url.c.diff?r1=1.795r2=1.797
2009/5/8, Yang Tse wrote:
2009/5/8, David McCreedy wrote:
How about a new strippartconv (or some similar name) that only applies for
systems that use CURL_DOES_CONVERSIONS?
I've tried this out and it works.
It would be ignored on systems that don't use CURL_DOES_CONVERSIONS.
And it's
2009/5/10, Daniel Stenberg wrote:
I committed changes last night that fixed the problems in my end.
I confirm that these changes also solved the problems on my side.
Thanks,
--
-=[Yang]=-
2009/5/8, Guenter wrote:
Hi,
Hi Guen,
it seems to me that since the last commits the tests
533, 546, 702, 703 fail now
The commit done for Constantine Sapuntzakis's bug report #2784055 is
also breaking things on my side, and other daily builds. Not much of a
surprise as the change log
2009/5/8, David McCreedy wrote:
How about a new strippartconv (or some similar name) that only applies for
systems that use CURL_DOES_CONVERSIONS?
I've tried this out and it works.
It would be ignored on systems that don't use CURL_DOES_CONVERSIONS.
And it's more flexible than a
2009/5/8, David McCreedy wrote:
On Thu, May 7, 2009 at 11:59 AM, Yang Tse yangs...@gmail.com wrote:
[...]
testcurl.pl tricks and test5XX patches on different files and later on.
Will do.
I meant runtests.pl that's the one that would need modification for this.
--
-=[Yang]=-
2009/5/4, Yang Tse wrote:
2009/5/1, David McCreedy wrote:
This patch changes tests/ftpserver.pl and tests/runtests.pl to pass the
client IP address on FTP Server tests.
Committed.
And reverted now.
The origin of the problem with test case #251 was the --ftp-port not
being the CLIENTIP
2009/5/1, David McCreedy wrote:
That causes two problems:
Committed,
Thanks
--
-=[Yang]=-
Hi friends,
Work done for ares_socklen_t has been done in parallel with the one
done for curl_socklen_t. Everything done for libcurl also applies to
c-ares.
Those not subscribed to the curl-library mailing list please read
thread linked below. It also applies to c-ares but just changing
'c-ares'
2009/4/30, john blair wrote:
I have successfully built a fat binary for curl 7.18.2. Has any one tried
building it with 7.19.4?
No personal experience on OS X. But...
Just in case you want to test it, there's an improved
'MacOSX-Framework' shell script in CVS that will ship with upcomming
2009/4/24, flourishing wrote:
and i searched google, many people says use __declspec(dllexport) but i
searched the full folder, didn'f find this.
my question is , how libcurl make this happend ? :)
It is there. Search for __declspec in include\curl\curl.h
--
-=[Yang]=-
Hi,
Any chance that OpenSSL-1.0.0 would include 'uninstall' as a valid make target?
--
-=[Yang]=-
__
OpenSSL Project http://www.openssl.org
Development Mailing List
2009/4/23, Jack Liptrap wrote:
I wanted to make the solution files available for those who use Microsoft
Visual Studio 2008. I spent the morning figuring out why VS9 wouldn't
compile the solution after it auto-upgraded the VS6 workspace. libcurl
project wasn't outputing to the correct
2009/4/14, Piotr Dobrogost wrote
Yang Tse wrote:
Hi Guen ;-)
Yang: I think you should proceed with your nmake tweaks
I respectfully listen your opinion, and in a friendly manner I tell
you that, for now, I won't follow it.
Yang
Have you changed your mind so we can finish
2009/4/8, Daniel Stenberg wrote:
We need to make the test case either conditional on using a static lib (when
on windows) or we need to make the test case use the code from the lib
separately (as opposed to just expecting the lib to export it). Or
something...
Skip test #558 and #559 also
Hi Gisle,
This http://cool.haxx.se/cvs.cgi/curl/tests/server/tftpd.c.diff?r1=1.48r2=1.49
change seems to have disgusted some daily builds.
Just in case you haven't noticed
--
-=[Yang]=-
2009/4/3, Stefan Teleman wrote:
So how often do you guys break ABI ? Inquiring minds want to know.
It sounds like you don't care at all that 7.19.x breaks ABI with
earlier releases.
Cool down.
The ABI breakage takes place only in some very specific cases for some
platforms. Depending on how
Daniel,
No computer this weekend for me, so I won't be able to answer any
message until Monday. In any case I want to leave something more
clear right away...
The purpose of curlbuild.h.dist is mostly for platforms which lack the
capability of running the configure script. This is mentioned
Hi,
Just a heads up for those using Sun compilers.
I've just committed a change to CVS with the following description:
Fix curl_off_t definition for builds done using Sun compilers and a
non-configured libcurl. In this case curl_off_t data type was gated
to the off_t data type which depends on
2009/4/2, Stefan Teleman wrote:
Your patch breaks ABI with earlier versions of CURL.
Yes. It breaks ABI for the very specific case which is mentioned in
the commit message, the RELEASE-NOTES document for next release and
the message I've posted to this list warning of such fact.
And it breaks
Hi Guen ;-)
Yang: I think you should proceed with your nmake tweaks
I respectfully listen your opinion, and in a friendly manner I tell
you that, for now, I won't follow it.
Contrary to what someone has said, Win32 lib/curl already can be built
out of the box in many ways and with a good bunch
2009/3/31, Daniel Stenberg wrote:
Please update from CVS again and see if this might be what happened.
Yes. it fixes it for me. Now the 6XX tests pass with no problem.
I'll start a full rebuild shortly and submit the log to the */auto page.
Cheers,
--
-=[Yang]=-
2009/3/27 Daniel Stenberg wrote:
Yes, that looks like the right thing to do. Will you commit?
Committed !
--
-=[Yang]=-
--
___
libssh2-devel mailing list
2009/3/17, Simon Josefsson wrote:
Adding -o 'StrictModes no' in ssh2.sh sshd options allows the test to
succeed even with the wrong $HOME permissions.
I've added this in CVS now, please confirm that it works.
Confirmed. It now works flawlessly.
--
-=[Yang]=-
Hendrik,
I've been unable, after 30 successive executions on Windows, to crash
any of your test programs. But there's still some hope ;-) Please
answer all the following...
Have you built libcurl or downloaded from somewhere? If downloaded
provide link to it.
Are you linking your test program
2009/3/11, Pierre Brico wrote:
[...] my first issue (I tested it under Windows and
it works). However, the behavior is different of what I was waiting for.
Indeed, I though the timeout was computed between two calls of the
write_callback function and not computed by subtracting current time
2009/3/10, Cleaver, Japheth wrote:
- u_long ip;
+ unsigned long ip;
Thanks, applied.
--
-=[Yang]=-
2008/10/20, Charles Hardin wrote:
On some networking stacks, a UDP connect can also return EINPROGRESS
or EWOULDBLOCK.
Out of curiosity, which system's connectionless UDP connect() returns
the above errors?
--
-=[Yang]=-
2008/9/29, Brad House wrote:
It appears as though a manual check for AR is being performed
which is infact causing issues if you want to use a different AR
than exists in your path, it's always overwriting it without ever
checking for the variable to have already been set.
2008/9/26, Brad House wrote:
In configure.ac:
It appears as though a manual check for AR is being performed
which is infact causing issues if you want to use a different AR
than exists in your path, it's always overwriting it without ever
checking for the variable to have already been set.
2008/9/12, E L [EMAIL PROTECTED] wrote:
The libssh2 code is not getting the real socket errors in most cases.
After calling send() or recv(), it is assuming errno will be set, but
Windows does not set errno for socket functions. If the connection
gets reset, the call to send() or recv() will
2008/8/26, Brad House wrote:
this may warrant a new release, even if its a bit early.
Sure, why not! Releasing should be fairly light work. Let's just say that
I'll put together a release in 48 hours unless someone speaks up with
something serious to address first!
Just need the
2008/8/3, Eric Blake wrote:
[Please keep the list in the loop]
[Sorry for that. I usually simply hit 'reply' and forget to double
check if the mailing list address is actually in the 'from' or 'reply
to' field of original message.]
So you are saying that the presence of the file containing
Hello!
Last addition of some macros to the cURL (libcurl) project have
triggered an aclocal failure which is only happening on the continuous
build system on Solaris systems. On other platforms there is no
problem related with this. http://curl.haxx.se/auto/
The failing system log can be viewed
Hello!
Since I'm not sure if this is actually an automake or autoconf issue,
and sure you can also help I'm also posting this here as well as on
the autoconf mailin-list. Excuse me if this bothers someone.
Last addition of some macros to the cURL (libcurl) project have
triggered an aclocal
2008/7/30, Ralf Wildenhues wrote:
aclocal endless loops are a pain to debug. If you can go back to a
known-good version (of configure.ac and *.m4 files), you may be able to
bisect.
One of the many things that have already been tried is to comment out
the new macro invocations leaving in
2008/7/6, Daniel Stenberg wrote:
I think it looks fine (and very curl-like ! :-) so I think it works for both
snapshots and releases.
Committed as posted ;-)
--
-=[Yang]=-
Platforms which define OPENSSL_SYS_WINDOWS or OPENSSL_SYS_WIN32 are
not properly built following the zlib configuration parameter
[[no-]zlib|zlib-dynamic] when static zlib support is required into
openssl.
Even if static linking of zlib.lib is desired specifying option 'zlib'
to 'Configure' what
This patch fixes the misplacement of a generated .pdb file when
building a debug version of openssl with Visual C.
Attached file 0.9.7-stable-SNAP-20050922-yang-diffs.patch is the patch
for 0.9.7-stable snapshot 20050922.
Attached file 0.9.8-stable-SNAP-20050922-yang-diffs.patch is the patch
for
401 - 475 of 475 matches
Mail list logo