Re: configure + make fails on AIX due to missing -lpthread

Using rcs - I was actually thinking about that last night. Now that I have
found you again I will probably do that.
I only stopped using it because I was not programming much myself. Now I
have some things I would rather not have in svn - I knew rcs too well and
that still gets in the way!

Any - test 1 - different system, not as root.

I believe the key difference is that I do not have coreutils installed on
this system and, at least the 'date' command is complaining.

michael@x071:[/data/prj/gnu/rcs/rcs-5.9.4]buildaix --disable-threads
+ CPPFLAGS="-I/opt/include" CFLAGS="-O2 -qlanglvl=extc99" ./configure \
                --prefix=/opt  \
                --sysconfdir=/var/rcs/etc \
                --sharedstatedir=/var/rcs/com \
                --localstatedir=/var/rcs \
                --mandir=/usr/share/man \
                --infodir=/opt/share/info/rcs --disable-threads \
                        > build/aix/configure.out
date: illegal option -- r
Usage: date [-u] [+Field Descriptors]
date: illegal option -- r
Usage: date [-u] [+Field Descriptors]
+ /opt/bin/make > build/aix/make.out
date: 0551-402 Invalid character in date/time specification.
Usage: date [-u] [+"Field Descriptors"]
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
-I.: A file or directory in the path name does not exist.
pic: Not a recognized flag: n
Usage: pic [-D] [-T Term] [--] [-] [File] ...
make[2]: *** [rcsfile.5] Error 1
make[1]: *** [all] Error 2
make: *** [all-recursive] Error 1
/opt/bin/make returned an error
michael@x071:[/data/prj/gnu/rcs/rcs-5.9.4]lslpp -L | grep coreutil
michael@x071:[/data/prj/gnu/rcs/rcs-5.9.4]


On Fri, Jan 23, 2015 at 9:07 AM, Thien-Thi Nguyen <[email protected]> wrote:

> () Michael Felt <[email protected]>
> () Thu, 22 Jan 2015 22:55:01 +0100
>
>    I am off to bed - but what files would you like besides
>    config.log, Makefile for starts here.
>
> Thanks.
>
>    The simple bit is configure determines that it is
>    pthread:posix, adds the -lpthread to some Makefile arguments,
>    but never includes those in the cc commands to create the
>    programs.
>
> Right.  That's because such functionality is not needed.  So the
> problem really is why the configure script generates such vars
> in the first place.
>
>    And I shall rerun everything tomorrow. I shall add the
>    -bnoquiet option and it shall reveal where the calls to
>    pthread are coming from.
>
> After some digging, i see that "./configure --help" output
> includes the lines:
>
>    --enable-threads={posix|solaris|pth|windows}
>                            specify multithreading API
>    --disable-threads       build without multithread safety
>
> so i would ask you to include ‘--disable-threads’ in the
> configure invocation.  I will arrange for that to be the default
> in the next release.
>
>    Check in the Makefile - there are variables defined that are
>    never used by cc to link the programs.
>
> Right.  That part of the flow is correct (RCS doesn't care about
> threads).
>
>    As to previous builds - I have not built it before. I saw it
>    in the announcements and remember rcs as a program I used
>    over 25 years ago. Been forced to use cvs and svm pretty much
>    the last 10-15 years - and they have yet newer systems
>    thought up it seems.
>
> Well, i hope we can resolve these issues so that you might find
> RCS useful (or at least fun to play with) henceforth.
>
> --
> Thien-Thi Nguyen
>    GPG key: 4C807502
>    (if you're human and you know it)
>       read my lisp: (responsep (questions 'technical)
>                                (not (via 'mailing-list)))
>                      => nil
>