Testing for the 4.4p1 release

Darren Tucker dtucker at zip.com.au
Thu Sep 7 10:53:29 EST 2006

David Bronder wrote:
> Yes, if I manually change the "langlvl" value from "ansi" to "extc89" in
> the relevant Makefiles, it builds and tests cleanly (with the expected
> skipped tests).  Looking at the config.log, configure didn't use "extc89"
> because of this error:
> configure:2734: cc -qlanglvl=extc89 -c -g  conftest.c >&5
> "conftest.c", line 43.19: 1506-195 (S) Integral constant expression with a value
>  greater than zero is required.
> The line in question is in a section of the test program specifically
> checking some IBM C 6 for AIX behavior.

OK, I see that in configure (it's put there by autoconf not explicitly 
by our configuration).

Could you please send me the output from "./configure && make -k" and 
plus a copy of the config.log file (just to me please, they'll be big 
and there's no need to fill the list members' inboxes :-)

>>> "/usr/include/usersec.h", line 625.32: 1506-310 (I) The type "struct aud_rec" was introduced in a parameter list, and will go out of scope at the end of the function declaration or definition.
>> The patch below should fix those.
> Switching "langlvl" to "extc89" fixed these without applying the patch
> to port-aix.h.

OK, I think this means that the configure test for AIX_LOGINFAILED_4ARG 
fails with langlvl=ansi for some reason.

You can use extc89 at configure time with:

CC="cc -qlanglvl=extc89" ./configure

but it would be nice to figure out how to make VAC6 work out of the box.

>> [...]
>>> "packet.c", line 162.12: 1506-010 (E) Macro TAILQ_HEAD invoked with a null argument for parameter name.
>> Please see my query above about compiler flags.
> Yeah, the "langlvl" change fixed this, too.
>>> "/usr/include/paths.h", line 50.9: 1506-213 (S) Macro name _PATH_BSHELL cannot be redefined.
>>> "/usr/include/paths.h", line 50.9: 1506-358 (I) "_PATH_BSHELL" is defined on line 322 of defines.h.
>> Hmm, the "cannot be redefined" sounds like we need to fix those before the
>> release.
> Still had these with openssh-SNAP-20060906 and "langlvl=extc89".

Was that both warnings or just the "is defined on line xxx"?

Darren Tucker (dtucker at zip.com.au)
GPG key 8FF4FA69 / D9A3 86E9 7EEE AF4B B2D4  37C9 C982 80C7 8FF4 FA69
     Good judgement comes with experience. Unfortunately, the experience
usually comes from bad judgement.

More information about the openssh-unix-dev mailing list