Twas brillig at 22:11:59 22.05.2009 UTC+02 when dan...@haxx.se did gyre and
gimble:
DS> I think I'll prefer to mimic the previous way of working for now,
DS> so that all existing committers can be given git push access to
DS> this master repo.
Anything need to be done for this?
Though I'm f
You ([EMAIL PROTECTED]) wrote:
DS> Ah, it seems totally suitable if that is the pointer set in
DS> libssh2_session_init_ex() ? If so, I'll retract my suggestion.
Yes, callback is called with abstract pointer set in
session_init_ex().
DS> We really need a set of man pages...
Mea culpa.
--
is mostly synchronous.
* src/packet.c 1.40 - improves detection of closed channel. Without
this change libssh2 hung when opening and closing many channels in
single SSH session.
--
Mikhail Gusarov
Senior Software Engineer
SWsoft, Inc.
JID: [EMAIL
You ([EMAIL PROTECTED]) wrote:
>> The patches:
DS> Sorry, but where exactly are these patches?
I just commited them.
--
Mikhail Gusarov
Senior Software Engineer
SWsoft, Inc.
JID: [EMAIL PROTECTED]
-
Take Surv
I'm going to add '--copy' option to the 'automake --add-missing'
clause in buildconf to make creating Debian packages a bit easier. Any
objections?
--
JID: [EMAIL PROTECTED]
-
Take Surveys. Earn Cash. Influence the Future o
Twas brillig at 01:49:35 03.01.2007 UTC-08 when Sara Golemon did gyre and
gimble:
SG> Go for it! Anything to make things easier on those Debian folk! :)
libssh2 in now on the way to the Debian unstable ;)
http://ftp-master.debian.org/new.html
--
Mikhail Gusarov
Senior Software Engin
est/ directory, compiles and runs everything found?
--
Mikhail Gusarov
Senior Software Engineer
SWsoft, Inc.
JID: [EMAIL PROTECTED]
-
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel an
you can
just drop new test_*.c to the directory and say 'make test'. DRY
principle.
--
Mikhail Gusarov
Senior Software Engineer
SWsoft, Inc.
JID: [EMAIL PROTECTED]
-
Take Surveys. Earn Cash. Influence the
Twas brillig at 17:13:15 when Simon Josefsson did gyre and gimble:
SJ> Ah, I understand. However, for 'make dist' to work properly (to
SJ> produce release *.tar.gz archives with all the required files),
SJ> the file has to be mentioned in Makefile.am. I don't see any
SJ> reliable way around
Several functions recently changed their contracts due to non-blocking
behavior, isn't it time to bump SONAME?
--
JID: [EMAIL PROTECTED]
-
Using Tomcat but need to do more? Need to support web services, security?
Get stuff
Twas brillig at 18:24:55 02.02.2007 UTC+01 when Daniel Stenberg did gyre and
gimble:
>> Several functions recently changed their contracts due to
>> non-blocking behavior, isn't it time to bump SONAME?
DS> There was no .so generated in the 0.14 release, it has been
DS> introduced in CVS sin
Twas brillig at 22:37:03 02.02.2007 UTC+01 when Daniel Stenberg did gyre and
gimble:
DS> So yes, let's do all the API tweaks we feel we have outstanding
DS> now and then we bump the soname for the real 0.15 release.
Could you do it? I'm not confident with handling sonames in autotools.
--
J
Twas brillig at 17:37:42 29.03.2007 UTC+07 when Elifant did gyre and gimble:
E> I have 2 patches for SFTP subsystem. Please review. These patches are based
E> on version 0.14. However CVS version still has the same bugs.
Applied. Please consider sending patches against CVS HEAD next time, and
Twas brillig at 22:50:04 28.06.2007 UTC+02 when Daniel Stenberg did gyre and
gimble:
DS> And yeah, I'm old and grumpy and I don't like web forums.
I'm old and grumpy 23-year one :), and +1 for disabling web forum.
--
JID: [EMAIL PROTECTED]
---
Twas brillig at 22:49:19 25.06.2007 UTC+02 when Daniel Stenberg did gyre and
gimble:
>> - the ./include/libssh2.h still has a #define LIBSSH2_VERSION "0.15-CVS"
DS> Oh right. I need to set up a little script that makes the release tarball
DS> so that this isn't forgotten next time...
Hmm. I
Twas brillig at 12:21:37 12.07.2007 UTC+02 when Guenter Knauf did gyre and
gimble:
GK> ../../curl/lib/transfer.c", line 335.47: 1506-207 (W) Integer constant
200706012030 out of range.
200706012030L ?
--
JID: [EMAIL PROTECTED]
---
Twas brillig at 23:27:38 19.07.2007 UTC-04 when Paul Thomas did gyre and gimble:
PT> And the strcasecmp (might have typed it wrong) should probably be something
PT> more like strcmp.
: strcasecmp, strncasecmp - compare two strings ignoring case
Does not look like it does exactly the same thin
Twas brillig at 10:31:20 20.07.2007 UTC+07 when Mikhail Gusarov did gyre and
gimble:
MG> This SF.net email is sponsored by: Microsoft
Damn. They are sneering. :)
--
JID: [EMAIL PROTECTED]
-
This SF.net email
Hi,
I just found that both 0.15 and 0.17 releases (did not check 0.16) contain
directories win32/test/CVS and win32/CVS.
--
JID: [EMAIL PROTECTED]
-
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log
"SourceForge.net" <[EMAIL PROTECTED]> writes:
> Bugs item #1816964, was opened at 2007-10-20 04:38
> Ok I found couples of things missing...
> I solved my problem with this :
[..]
Who has the necessary permissions on the SF account? Please close the bug.
--
JID: [EMAIL PROTECTED]
---
Daniel Stenberg <[EMAIL PROTECTED]> writes:
> Just FYI: I've now added you as a tech for the bug tracker so you too can
> close them in the future!
Thanks!
P.S.: looking at popcon.debian.org: woohoo! I'm in top 500 maintainers due to
libssh2-1 pulled by libcurl3 in unstable :)
--
JID: [EMAIL P
Daniel Stenberg <[EMAIL PROTECTED]> writes:
> Anything in particular I should pay attention to before I just proceed
> with this?
Release-making script leaves some CVS files in tarball (win32/CVS and
win32/test/CVS). Please fix it, so I no longer need to shut up the
lintian about them.
I'm sorry
Dan Fandrich <[EMAIL PROTECTED]> writes:
> You could store an array of functors for objects using libssh2 in a
> static variable, then choose one of those from your static callback
> function based on the libssh2 session ID that's passed in to the
> callback.
It's already done in the libssh2 itse
"Paul Thomas" <[EMAIL PROTECTED]> writes:
> So when I'm using the library, my class has to have a static member
> function so that it can be passed into you library as a function
> pointer. Unfortunately a non-static member function can not be passed
> into the library.
You've got "abstract" void
Twas brillig at 22:41:26 30.12.2008 UTC+01 when dan...@haxx.se did gyre and
gimble:
>> You are right, this works very well. The problem is, that if I call
>> "libssh2_channel_wait_closed" right before
>> "libssh2_channel_get_exit_status" the function
>> "libssh2_channel_wait_closed" never ret
Hello,
I was packaging 1.0 for Debian and found out that there are private
symbols exported by the library:
+ _libssh2_cipher_crypt
+ _libssh2_cipher_init
+ _libssh2_dsa_new
+ _libssh2_dsa_new_private
+ _libssh2_dsa_sha1_sign
+ _libssh2_dsa_sha1_verify
+ _libssh2_pem_decode_integer
+ _libssh2_pem
Twas brillig at 10:35:26 17.02.2009 UTC+01 when si...@josefsson.org did gyre
and gimble:
SJ> How about adding it and see if we run into problems? It would
SJ> solve Debian's symbol packaging concern.
Added to 1.0-1 in Debian, no problems reported yet.
--
pgpUiVbLMQO4Q.pgp
Description: PG
Twas brillig at 11:02:43 17.03.2009 UTC+01 when si...@josefsson.org did gyre
and gimble:
SJ> How about reducing the number of exported symbols?
SJ> Ok to commit?
ack
--
pgpGcrvxNFLIU.pgp
Description: PGP signature
--
Twas brillig at 12:57:41 19.03.2009 UTC+01 when pe...@stuge.se did gyre
and gimble:
PS> I like trac a lot. One gotcha is that it was very tied to
PS> subversion, but I understand that the vcs situation with trac is
PS> improving.
git and Mercurial plugins for Trac are very stable and usable s
Twas brillig at 13:01:45 19.03.2009 UTC+01 when si...@josefsson.org did
gyre and gimble:
SJ> I gave up maintaining a trac for gnutls since dealing with spam
SJ> took too much time from my hands, so be careful about how you set
SJ> up authentication.
I fought spam on all tracs I manage with th
Twas brillig at 13:19:44 19.03.2009 UTC+00 when sw...@lammy.co.uk did gyre and
gimble:
AL> For my own selfish reasons I like sticking to CVS/SVN. My build
AL> process automatically pulls in a copy of libssh2 from sourceforge
AL> via CVS
It should be easy to setup git-cvsserver running nearb
Twas brillig at 17:24:34 21.03.2009 UTC-04 when e...@nanobyte.org did gyre and
gimble:
ES> From a quick read of this excerpt from "C in a Nutshell"
Erm. Not quite a respectable source. Please provide another one: C99,
POSIX, SUS - nearly anything will be ok.
ES> The attached patch against th
Twas brillig at 13:17:51 23.03.2009 UTC+00 when
libssh2-...@lists.sourceforge.net did gyre and gimble:
Ccmfl> --- libssh2_session_flag.316 Mar 2009 15:00:45 -1.1
Ccmfl> +Its mere existence is an error
I propose to drop this function when SONAME will need to be bumped.
--
The following symbols have evaporated from the public interface since
1.0:
libssh2_channel_locate
libssh2_channel_nextid
libssh2_channel_packet_data_len
libssh2_htonu32
libssh2_htonu64
libssh2_ntohu32
libssh2_ntohu64
libssh2_packet_add
libssh2_packet_ask_ex
libssh2_packet_askv_ex
libssh2_packet_b
Twas brillig at 15:49:03 23.03.2009 UTC+01 when dan...@haxx.se did gyre and
gimble:
>> Probably most of them were exported by mistake, but it needs a double-check.
DS> That list looks entirely correct to me!
Okay, marked as un-exported in package symbols list.
--
pgpJAjUCtH7wA.pgp
Descrip
Twas brillig at 12:02:55 27.03.2009 UTC+01 when jean-louis.char...@oikialog.com
did gyre and gimble:
JC> I've made it possible to pass a NULL pointer for the publickey
JC> parameter of libssh2_userauth_publickey_fromfile and
JC> libssh2_userauth_hostbased_fromfile functions.
Thanks!
JC> I'
Twas brillig at 12:15:51 27.03.2009 UTC+01 when jean-louis.char...@oikialog.com
did gyre and gimble:
JC> +p_bytes = BN_num_bytes(dsa->p) + 1;
JC> +q_bytes = BN_num_bytes(dsa->q) + 1;
JC> +g_bytes = BN_num_bytes(dsa->g) + 1;
JC> +k_bytes = BN_num_bytes(dsa->pub_key) + 1;
Seems th
Twas brillig at 07:13:21 29.03.2009 UTC-07 when masott...@yahoo.com did gyre
and gimble:
PM> Hi
First of all, do not steal the threads.
PM> I'd like to know why is not libshh2 just distributed in a single
PM> package not depending on third party libraries??...
Because it's stupid.
--
p
Twas brillig at 08:49:34 02.04.2009 UTC+00 when
libssh2-...@lists.sourceforge.net did gyre and gimble:
Ccmfl> -#define LIBSSH2_VERSION "1.0.1_CVS"
Ccmfl> +#define LIBSSH2_VERSION "1.1_CVS"
1.1_CVS, really? Shouldn't it be just 1.1?
--
Twas brillig at 11:40:10 02.04.2009 UTC+02 when dan...@haxx.se did gyre and
gimble:
DS> On Thu, 2 Apr 2009, Mikhail Gusarov wrote:
>> 1.1_CVS, really? Shouldn't it be just 1.1?
DS> The actual file in the release tarball actually gets that line
DS> replaced by the &#x
Twas brillig at 15:45:35 30.04.2009 UTC+02 when nin...@gmx.at did gyre and
gimble:
d> After "making" the library, the created filename should be
d> "libssh2.so.1.0.2" instead of "libssh2.so.1.0.1" (everywhere, also
d> in libssh2.la).
filenames do not matter, SONAMES do.
$ readelf -a libssh2
Twas brillig at 13:38:44 04.05.2009 UTC+02 when dan...@haxx.se did gyre and
gimble:
DS> In my initial take I'll focus on OpenSSH complience, but I figure
DS> it could be useful to see what it takes to support putty's file
DS> format too. Anyone knows any details?
Erm... What?
LIBSSH2_API co
Twas brillig at 18:48:34 07.05.2009 UTC+02 when dan...@haxx.se did gyre and
gimble:
DS> I'd appreciate some comments/testing etc!
#ifndef WIN32
... obtaining OpenSSH agent socket
#else
... obtaining Putty agent socket
#endif
looks strange. Is putty agent the only one under Windows? Not even
m
43 matches
Mail list logo
