On 1/13/07, Sylvain Le Gall <[EMAIL PROTECTED]> wrote:
tags 406711 +upstream
thanks
Hello,
On Wed, Jan 03, 2007 at 03:18:44PM +0000, Hans Fangohr wrote:
> Package: libocamlgsl-ocaml-dev
> Version: 0.5.1-3
> Severity: important
>
> Hi,
>
> When building any ocaml-toplevel executables that use gsl, we obtain the
following error (we show the
> output of an example below, the source files for that example are attached):
>
> [EMAIL PROTECTED]:~/tmp/gslbug/gslbug$ make top
> make[1]: Entering directory `/home/fangohr/tmp/gslbug/gslbug'
> making ._bcdi/gsltest.di from gsltest.mli
> making ._d/gsltest.d from gsltest.ml
> make[1]: Leaving directory `/home/fangohr/tmp/gslbug/gslbug'
> make[1]: Entering directory `/home/fangohr/tmp/gslbug/gslbug'
> ocamlfind ocamlc -package gsl,threads -c -thread gsltest.mli
> ocamlfind ocamlc -package gsl,threads -c -thread gsltest.ml
> ocamlfind ocamlmktop \
> -package gsl,threads -linkpkg \
> -thread -custom -o
gsltest.top \
> gsltest.cmo
> /tmp/ccZU8img.o:(.data+0x12b4): undefined reference to
`ml_gsl_sf_legendre_array_size_e'
> collect2: ld returned 1 exit status
> Error while building custom runtime system
> make[1]: *** [gsltest.top] Error 2
> make[1]: Leaving directory `/home/fangohr/tmp/gslbug/gslbug'
> make: *** [top] Error 2
>
> This has come up in a larger project, and means we can't compile this anymore
on debian etch. (Rather
> annoying from our point of view.)
>
> To make it easier to track this down, I attach a gzipped tar file, which can
be untarred. It contains the
> files 'gsltest.ml gsltest.mli Makefile META OCamlMakefile' in a
subdirectory 'gslbug'. If you cd into
> that subdirectory and run 'make top', you should get the error above.
>
> I have tested this with older versions of etch (based on libocaml 0.4, and it
works fine).
>
I think you mean : "based on libocamlgsl-ocaml-dev".
I think it is an upstream bug, and thus forwarding it to upstream.
Salut Sylvain,
indeed, it is. I've uploaded ocamlgsl 0.5.2
http://oandrieu.nerim.net/ocaml/gsl/ocamlgsl-0.5.2.tar.gz which should
fix the problem.
Cheers,
--
Olivier
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]