[Ecls-list] cl-launch and ecl, again

[Faré]

2008/10/31 Juan Jose Garcia-Ripoll <juanjose.garciaripoll at googlemail.com>:
> On Thu, Oct 30, 2008 at 10:06 PM, Faré <fahree at gmail.com> wrote:
>>> Thanks for the other files and the detailed description of the
>>> problem. I will download them and investigate it during the following
>>> days -- cannot promise full commitment, though, as I will be
>>> travelling.
>>>
>> Take your time. I believe that this is a bug in ECL and that when this
>> bug is fixed, cl-launch will work fine with ECL.
>
> I just read your blog entry. The actual problem has not even been
> identified and you are stating this is a bug of the implementation. Is
> there any need on stating as true something that has not been yet
> confirmed? You do not even know what is causing the problem, whether
> it is ECL or just that your implementation of cl-launch got outdated
> by current changes.
>
I did update cl-launch. cl-launch is not the software that tries to
use linker symbols that have not been generated or otherwise fails to
generate or register initialization routines. If I'm using the
available interface incorrectly (which I don't think I am, but hey,
that's possible), at the very least ECL is coming up with a much
lower-level error than should be. In other words, a bug in cl-launch
is quite possible, but a bug in ECL (or possibly the underlying C
compiler or linker toolchain, or the asdf-ecl contrib - but it looks
like ECL to me) is quite certain.

That said, I've been known for jumping to conclusions and opening my
big fat mouth. And sometimes publicly saying mea culpa for it.

/home/fare/.cache/lisp-fasl/ecl-8.10.0__cvs_2008-07-12_18_54_-linux-pentium3/dev/shm/tmp/foo/libclt-asd.a(ECLINITu45oWA.o):
In function `init_lib_CLT_ASD':
/dev/shm/tmp/fare/ECLINITu45oWA.c:40: undefined reference to
`_eclI7Z7omJ7_BAUArSy'
collect2: ld returned 1 exit status
An error occurred during initialization:
(SYSTEM "gcc -o
\"/home/fare/.cache/lisp-fasl/ecl-8.10.0__cvs_2008-07-12_18_54_-linux-pentium3/dev/shm/tmp/foo/cl-launch-10110-program\"
-L\"/usr/local/stow/ecl/lib/\" \"/dev/shm/tmp/fare/ECLINITz2jMy9.o\"
\"/home/fare/.cache/lisp-fasl/ecl-8.10.0__cvs_2008-07-12_18_54_-linux-pentium3/dev/shm/tmp/foo/libcl-launch-10110-prefix.a\"
\"/home/fare/.cache/lisp-fasl/ecl-8.10.0__cvs_2008-07-12_18_54_-linux-pentium3/dev/shm/tmp/foo/libclt-asd.a\"
\"/home/fare/.cache/lisp-fasl/ecl-8.10.0__cvs_2008-07-12_18_54_-linux-pentium3/dev/shm/tmp/foo/libcl-launch-10110-program.a\"
   -lecl -lpthread -ldl  -lm   -lgc -lgmp") returned non-zero value 1.

I swear the code inserting a dangling reference to ECLINITu45oWA isn't
on cl-launch's side! Maybe either asdf-ecl or build-static-library is
failing to link in the master object for the .asd files themselves, or
something? I admit I don't understand this part of the code.

In any case, it looks pretty reproducible whenever dumping asd systems
is involved, as in
   cl-launch -l ecl -B tests 76

[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
We don't want tradition. We want to live in the present and the only history
that is worth a tinker's dam is the history we make today. -- Henry Ford