headers
Darren Tucker | 1 Sep 2006 05:31
Picon
Picon

Re: Testing for the 4.4p1 release

On Wed, Aug 30, 2006 at 06:35:21PM +0200, Corinna Vinschen wrote:
> Cygwin 1.5.21, OpenSSL 0.9.8b

Thanks for testing.

[...]
> Actually, OpenSSH didn't use Cygwin's glob() implementation before
> (which is a relatively old NetBSD derived implementation), because the
> configure test for gl_matchc failed up to 4.3p2.  The AC_EGREP_CPP
> autoconf test failed, while the new AC_TRY_COMPILE test in 4.4p1 now
> works, so starting with 4.4p1, OpenSSH uses Cygwin's glob function.
> 
> But why does it core dump?  The reason is that the old glob implementation
> in Cygwin doesn't know about the GLOB_NOMATCH return code.  In case there's
> no match, it returns 0, with gl_matchc set to 0 and gl_pathv set to NULL.

I'm wondering if we should test for GLOB_NOMATCH in configure and use
the glob in openbsd-compat if it's not found.  This would avoid having
to carry additional diffs in -portable.

Index: configure.ac
===================================================================
RCS file: /usr/local/src/security/openssh/cvs/openssh_cvs/configure.ac,v
retrieving revision 1.356
diff -u -p -r1.356 configure.ac
--- configure.ac	30 Aug 2006 17:24:41 -0000	1.356
+++ configure.ac	1 Sep 2006 00:59:14 -0000
 <at>  <at>  -982,6 +982,8  <at>  <at>  AC_TRY_COMPILE(
 	]
 )
(Continue reading)

Permalink | Reply | 1 comment |
headers
Damien Miller | 1 Sep 2006 07:42
Favicon

Re: Testing for the 4.4p1 release

On Thu, 31 Aug 2006, santhi wrote:

> /usr/ccs/bin/ld: Unsatisfied symbols:
>    ntohl (first referenced in sshconnect.o) (code)
>    ntohs (first referenced in ssh.o) (code)
>    htonl (first referenced in sshconnect.o) (code)
>    htons (first referenced in
> openbsd-compat//libopenbsd-compat.a(rresvport.o)) (code)
> *** Error exit code 1
> 
> Stop.
> 
> Problem Description:
> ----------------------
> * The ntohl, ntohs, htonl, htons routines are defined as macros instead of
> functions in <netinet/in.h> , linker raises "functions" not found error.
> 
> Fix:
> -----
> Including the <arpa/inet.h> in includes.h file may fix this problem

Thanks for the report - does the following diff help?

Index: openbsd-compat/bindresvport.c
===================================================================
RCS file: /var/cvs/openssh/openbsd-compat/bindresvport.c,v
retrieving revision 1.7
diff -u -p -r1.7 bindresvport.c
--- openbsd-compat/bindresvport.c	24 Jul 2006 04:51:01 -0000	1.7
+++ openbsd-compat/bindresvport.c	1 Sep 2006 05:41:51 -0000
(Continue reading)

Permalink | Reply | 2 comments |
headers
Damien Miller | 1 Sep 2006 07:49
Favicon

Re: Testing for the 4.4p1 release

On Thu, 31 Aug 2006, Iain Morgan wrote:

> Sometime ago, Damien Miller wrote:
> [...]
> > Testing on suitable non-production systems is also appreciated. Please send
> > reports of success or failure to openssh-unix-dev <at> mindrot.org, including 
> > details of your platform, compiler and configure options.
> > 
> 
> Configures, builds and tests OK on Solaris 9 with Sun Forte 7 C 5.4
> compiler. However, there are a lot of warnings. A breakdown of the
> warnings follows:

Thanks for the report - could you please try tonight's (20060902)
snapshot? These should be fixed.

-d
Permalink | Reply | 1 comment |
headers
santhi | 1 Sep 2006 09:01
Picon

Re: Testing for the 4.4p1 release

> On Thu, 31 Aug 2006, santhi wrote:
>
> > /usr/ccs/bin/ld: Unsatisfied symbols:
> >    ntohl (first referenced in sshconnect.o) (code)
> >    ntohs (first referenced in ssh.o) (code)
> >    htonl (first referenced in sshconnect.o) (code)
> >    htons (first referenced in
> > openbsd-compat//libopenbsd-compat.a(rresvport.o)) (code)
> > *** Error exit code 1
> >
> > Stop.
> >
> > Problem Description:
> > ----------------------
> > * The ntohl, ntohs, htonl, htons routines are defined as macros instead
of
> > functions in <netinet/in.h> , linker raises "functions" not found error.
> >
> > Fix:
> > -----
> > Including the <arpa/inet.h> in includes.h file may fix this problem
>
> Thanks for the report - does the following diff help?
>
> Index: openbsd-compat/bindresvport.c
> ===================================================================
> RCS file: /var/cvs/openssh/openbsd-compat/bindresvport.c,v
> retrieving revision 1.7
> diff -u -p -r1.7 bindresvport.c
> --- openbsd-compat/bindresvport.c 24 Jul 2006 04:51:01 -0000 1.7
(Continue reading)

Permalink | Reply | 1 comment |
headers
Corinna Vinschen | 1 Sep 2006 10:44
Picon
Favicon

Re: Testing for the 4.4p1 release

On Sep  1 13:31, Darren Tucker wrote:
> On Wed, Aug 30, 2006 at 06:35:21PM +0200, Corinna Vinschen wrote:
> > Cygwin 1.5.21, OpenSSL 0.9.8b
> 
> Thanks for testing.
> 
> [...]
> > Actually, OpenSSH didn't use Cygwin's glob() implementation before
> > (which is a relatively old NetBSD derived implementation), because the
> > configure test for gl_matchc failed up to 4.3p2.  The AC_EGREP_CPP
> > autoconf test failed, while the new AC_TRY_COMPILE test in 4.4p1 now
> > works, so starting with 4.4p1, OpenSSH uses Cygwin's glob function.
> > 
> > But why does it core dump?  The reason is that the old glob implementation
> > in Cygwin doesn't know about the GLOB_NOMATCH return code.  In case there's
> > no match, it returns 0, with gl_matchc set to 0 and gl_pathv set to NULL.
> 
> I'm wondering if we should test for GLOB_NOMATCH in configure and use
> the glob in openbsd-compat if it's not found.  This would avoid having
> to carry additional diffs in -portable.

Good idea!  I tested your patch and it works, if another glob related
expression is changed in includes.h:

Index: includes.h
===================================================================
RCS file: /cvs/openssh/includes.h,v
retrieving revision 1.125
diff -p -u -r1.125 includes.h
--- includes.h  1 Sep 2006 05:48:19 -0000       1.125
(Continue reading)

Permalink | Reply | 0 comments |
headers
Corinna Vinschen | 1 Sep 2006 11:10
Picon
Favicon

[PATCH] Cygwin: Avoid implicit declaration warnings

Hi,

I have left this slip through already too long.  When compiling
openbsd-compat/bsd-cygwin_util.c, the following warnings appear:

  openbsd-compat/bsd-cygwin_util.c: In function `binary_open':
  openbsd-compat/bsd-cygwin_util.c:67: warning: implicit declaration of function `open'
  openbsd-compat/bsd-cygwin_util.c: In function `binary_pipe':
  openbsd-compat/bsd-cygwin_util.c:73: warning: implicit declaration of function `pipe'

The below patch fixes that.

Index: openbsd-compat/bsd-cygwin_util.c
===================================================================
RCS file: /cvs/openssh/openbsd-compat/bsd-cygwin_util.c,v
retrieving revision 1.18
diff -p -u -r1.18 bsd-cygwin_util.c
--- openbsd-compat/bsd-cygwin_util.c    5 Aug 2006 09:08:17 -0000       1.18
+++ openbsd-compat/bsd-cygwin_util.c    1 Sep 2006 08:42:22 -0000
 <at>  <at>  -31,6 +31,13  <at>  <at> 

 #ifdef HAVE_CYGWIN

+#if defined(open) && open == binary_open
+# undef open
+#endif
+#if defined(pipe) && open == binary_pipe
+# undef pipe
+#endif
+
(Continue reading)

Permalink | Reply | 0 comments |
headers
Iain Morgan | 1 Sep 2006 19:19
Picon
Favicon

Re: Testing for the 4.4p1 release

Sometime ago, Damien Miller wrote:
> On Thu, 31 Aug 2006, Iain Morgan wrote:
> 
> > Sometime ago, Damien Miller wrote:
> > [...]
> > > Testing on suitable non-production systems is also appreciated. Please send
> > > reports of success or failure to openssh-unix-dev <at> mindrot.org, including 
> > > details of your platform, compiler and configure options.
> > > 
> > 
> > Configures, builds and tests OK on Solaris 9 with Sun Forte 7 C 5.4
> > compiler. However, there are a lot of warnings. A breakdown of the
> > warnings follows:
> 
> Thanks for the report - could you please try tonight's (20060902)
> snapshot? These should be fixed.
> 

I just tried the 0902 snapshot. The vast majority of the warnigns are
now fixed. There are still 14 warnings, but I can live with that.
Thanks.

"bsd-cray.c", line 819: warning: empty translation unit
"/usr/include/iso/stddef_iso.h", line 73: warning: macro redefined: offsetof
"hostfile.c", line 92: warning: argument #2 is incompatible with prototype:
"hostfile.c", line 130: warning: argument #2 is incompatible with prototype:
"hostfile.c", line 131: warning: argument #2 is incompatible with prototype:
"hostfile.c", line 134: warning: argument #1 is incompatible with prototype:
"hostfile.c", line 135: warning: argument #1 is incompatible with prototype:
"/usr/include/iso/stddef_iso.h", line 73: warning: macro redefined: offsetof
(Continue reading)

Permalink | Reply | 0 comments |
headers
Iain Morgan | 1 Sep 2006 21:14
Picon
Favicon

Re: Testing for the 4.4p1 release

Sometime ago, Iain Morgan wrote:
> Sometime ago, Damien Miller wrote:
> [...]
> > Testing on suitable non-production systems is also appreciated. Please send
> > reports of success or failure to openssh-unix-dev <at> mindrot.org, including 
> > details of your platform, compiler and configure options.
> > 
> 
> The 20060830 snapshot configures, but fails to build on IRIX 6.5.29
> usign the MIPSpro 7.4 compilers.
> 
> 
> export CC=c99
> export CFLAGS="-g -O3 -mips3 -r14000"
> 
> PREFIX=/usr/prg/pkg/openssh/4.4p1
> OPENSSL=$HOME/build
> ZLIB=$HOME/build
> 
> ./configure --prefix=$PREFIX \
>         --sysconfdir=/usr/prg/etc \
>         --with-ssl-dir=$OPENSSL \
>         --with-zlib=$ZLIB \
>         --with-tcp-wrappers=/usr/local \
>         --with-md5-passwords \
>         --with-pam
> 
> There are numberous warnings, but here's the part that causes the
> compile to fail:
>
(Continue reading)

Permalink | Reply | 11 comments |
headers
Roger Cornelius | 2 Sep 2006 01:45
Favicon

Re: Testing for the 4.4p1 release

[ SCO OSR6 w/mp2 using system compiler, SNAP-20060902 ]

On 08/31/2006 10:12, Roger Cornelius wrote:
> [4.4p1 on SCO OSR6 w/MP2]
> 
> Thanks.  The build now completes but 'make tests' fails at:
> 
> run test login-timeout.sh ...
> /u1/src/rac/openssh/openssh/regress/login-timeout.sh: warning: line 18: `...` ob
> solete, use $(...)
> ssh_exchange_identification: Connection closed by remote host^M
> ssh connect after login grace timeout failed with privsep
> failed connect after login grace timeout
> make[1]: *** [t-exec] Error 1
> make[1]: Leaving directory `/u1/src/rac/openssh/openssh/regress'
> make: *** [tests] Error 2
> 
> I won't be able to look into why it's failing until possibly tomorrow.
> I'll also need to test that updwtmpx() works correctly.  4.3p1 required
> me to build with -DBROKEN_WTMPX (which I didn't report), and I see it is
> not defined by configure for this platform.

OK, I apparently have a dns issue that is causing
regress/login-timout.sh to timeout and fail.  'make tests' completes
without error after increasing the duration of the sleeps in that
script (I've noticed for the last week or so that connecting to this
system with ssh takes an unusually long time).

OSR6 does need BROKEN_UPDWTMP defined.  Otherwise wtmp gets trashed when
someone connects via ssh.  This can be seen by running /bin/last
(Continue reading)

Permalink | Reply | 5 comments |
headers
David Bronder | 2 Sep 2006 03:14
Picon
Favicon

Re: Testing for the 4.4p1 release

Damien Miller wrote:
> 
> The 4.4p1 release is approaching now, so we are now asking people to 
> actively test snapshots or CVS and report back to the mailing list.

[AIX 5.1 ML5, IBM VAC 6 compiler, openssh-SNAP-20060902]

Fails to build.  See below for details.  Same result with a stripped
down configure line (just adding --with-zlib=/usr/local).

-----

$ ./configure --libexecdir='${exec_prefix}/bin' --sysconfdir=/etc/ssh --with-pid-dir=/etc/ssh
--with-privsep-path=/var/empty/sshd --with-tcp-wrappers=/local/admin --with-zlib=/usr/local
--with-xauth=/usr/bin/X11/xauth --with-cflags="-O3 -qstrict"

OpenSSH has been configured with the following options:
                     User binaries: /usr/local/bin
                   System binaries: /usr/local/sbin
               Configuration files: /etc/ssh
                   Askpass program: /usr/local/bin/ssh-askpass
                      Manual pages: /usr/local/share/man/manX
                          PID file: /etc/ssh
  Privilege separation chroot path: /var/empty/sshd
            sshd default user PATH: /usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin
                    Manpage format: man
                       PAM support: no
                   OSF SIA support: no
                 KerberosV support: no
                   SELinux support: no
(Continue reading)

Permalink | Reply | 8 comments |

Gmane