Skip Menu |
 

To: krb5-bugs@mit.edu
From: "Henry B. Hotz" <hotz@jpl.nasa.gov>
Subject: Problems with 1.5 on Solaris 9 w gcc
Date: Thu, 3 Aug 2006 16:27:02 -0700
Download (untitled) / with headers
text/plain 2.1KiB
No obvious build problems (though I can't seem to build against the
native ndbm any more). No obvious operation problems for the
ordinary stuff like kinit, kvno, klist, etc. Could build openssh 4.3
(with Simon's patches) and it does everything it should.

In fact the only problem I actually see is that a "make check" fails
as follows.


Show quoted text
> gcc --version
gcc (GCC) 3.4.2
Copyright (C) 2004 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There
is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR
PURPOSE.

Show quoted text
> uname -a
SunOS redhotz 5.9 Generic_118558-25 sun4u sparc SUNW,Sun-Blade-1500
Show quoted text
> ./configure --prefix=/usr/demo CPPFLAGS="-I/usr/local/include"
CFLAGS="-m32 -mcpu=ultrasparc" LDFLAGS="-R/usr/demo/lib -R/usr/local/
lib -L/usr/demo/lib -L/usr/local/lib" --with-tcl=/usr/local --without-
krb4

make
sudo make install
setenv KRB5_CONFIG /etc/krb5/krb5.conf
make check

=== krb tests ===

Schedule of variations:
unix

Running target unix
Using /usr/ets/share/dejagnu/baseboards/unix.exp as board description
file for target.
Using /usr/ets/share/dejagnu/config/unix.exp as generic interface
file for target.
Using ./config/default.exp as tool-and-target-specific interface file.
Running pass `des' ...
Running ./krb-root/rlogin.exp ...
NOTE: rlogin: rlogin doesn't like command-line flags
Running ./krb-root/telnet.exp ...
NOTE: telnet test requires ability to log in as root
Running ./krb-standalone/gssapi.exp ...
Running ./krb-standalone/gssftp.exp ...
ERROR: error in gssftp.exp: can not find channel named "exp10"
Running ./krb-standalone/kadmin.exp ...
ERROR: error in kadmin.exp
can not find channel named "exp12"
make[2]: *** [check-runtest-yes] Error 1
make[2]: Leaving directory `/usr/local/dist/krb5-1.5/src/tests/dejagnu'
make[1]: *** [check-recurse] Error 1
make[1]: Leaving directory `/usr/local/dist/krb5-1.5/src/tests'
make: *** [check-recurse] Error 1
Show quoted text
>


channel "exp12"
whazzat?
------------------------------------------------------------------------
----
The opinions expressed in this message are mine,
not those of Caltech, JPL, NASA, or the US Government.
Henry.B.Hotz@jpl.nasa.gov, or hbhotz@oxy.edu
This sounds like an old problem with the Solaris 9 pty driver. Do you
have patch 117171-11 (or later rev of that patch) applied? Is
PRICNTL_HACK=1 in the arguments to runtest when you run "make check"?
From: "Henry B. Hotz" <hotz@jpl.nasa.gov>
Subject: Re: [krbdev.mit.edu #4108] Problems with 1.5 on Solaris 9 w gcc
Date: Fri, 1 Dec 2006 12:19:26 -0800
To: rt-comment@krbdev.mit.edu
RT-Send-Cc:

On Dec 1, 2006, at 9:27 AM, Tom Yu via RT wrote:

Show quoted text
> This sounds like an old problem with the Solaris 9 pty driver. Do you
> have patch 117171-11 (or later rev of that patch) applied? Is
> PRICNTL_HACK=1 in the arguments to runtest when you run "make check"?

I have patch 117171-17 installed. (Also -07, and a patch which
obsoletes -10, not that they matter.)

I don't have a record handy, but I'm pretty sure I didn't give "make
check" any special arguments to cause that. I suppose you'd like to
recheck this, right?

------------------------------------------------------------------------
The opinions expressed in this message are mine,
not those of Caltech, JPL, NASA, or the US Government.
Henry.B.Hotz@jpl.nasa.gov, or hbhotz@oxy.edu
To: rt@krbdev.mit.edu
Subject: Re: [krbdev.mit.edu #4108] Problems with 1.5 on Solaris 9 w gcc
From: Tom Yu <tlyu@MIT.EDU>
Date: Fri, 01 Dec 2006 15:28:08 -0500
RT-Send-Cc:
Show quoted text
>>>>> "Henry" == \"\"Henry B Hotz\" via RT\" <Henry> writes:

Show quoted text
Henry> On Dec 1, 2006, at 9:27 AM, Tom Yu via RT wrote:

Show quoted text
>> This sounds like an old problem with the Solaris 9 pty driver. Do you
>> have patch 117171-11 (or later rev of that patch) applied? Is
>> PRICNTL_HACK=1 in the arguments to runtest when you run "make check"?

Show quoted text
Henry> I have patch 117171-17 installed. (Also -07, and a patch which
Henry> obsoletes -10, not that they matter.)

Ok, that patch should have fixed the pty problem, and has in our
experience here, so something different might be going on.

Show quoted text
Henry> I don't have a record handy, but I'm pretty sure I didn't give "make
Henry> check" any special arguments to cause that. I suppose you'd like to
Henry> recheck this, right?

PRIOCNTL_HACK is set by the Makefile when "make check" is run in
tests/dejagnu, based on whether configure detects the patch 117171-11
or later. Is the "make check" failure consistent between runs? Do
you have the dbg.log file saved?

---Tom