Re: [Plplot-general] help building ocaml bindings

[Alan W. I. <ir...@be...>]

Hi Ed:

On 2010-01-12 16:56+0800 Ed Zaron wrote:

>
> Hi All,
>
> I'm having trouble building the ocaml bindings from the latest plplot in
> the svn repository.
>
> This is on a recent Ubuntu/Debian linux. And it seems to be a makefile
> problem.
>
> Has anyone seen errors like this from "make build"?
>
> Scanning dependencies of target target_lib_plplot_stubs
> make[2]: Circular bindings/ocaml/plplot_core.idl <-
> bindings/ocaml/plplot_core.idl dependency dropped.
> [ 21%] Generating plplot_core.idl, plplot_core.h, plplot_core.ml,
> plplot_core.mli, plplot_core_stubs.c
> Error copying file (if different) from
> "/home/ezaron/plplot/bindings/ocaml/plplot_core.idl" to
> "/home/ezaron/plplot/bindings/ocaml/plplot_core.idl".
> make[2]: *** [bindings/ocaml/plplot_core.idl] Error 1
> make[1]: *** [bindings/ocaml/CMakeFiles/target_lib_plplot_stubs.dir/all]
> Error 2
> make: *** [all] Error 2

Thanks for your report.  I think you have caught a pervasive bug in our
build system for the case where the build tree is identical to the source
tree.

For some time now, I have assumed CMake's copy-if-different would work fine
(i.e., would take no action) if the build tree was identical to the
source tree (as in your case) so that no copying from source tree
to build tree would be required.  But from your error report it looks
like copy-if-different does not work the way I had assumed, and instead
tries to do something circular for this case.

The workaround I suggest for you now is to use a separate build tree, i.e.,
do _not_ use "cmake .".  "cmake some_other_directory" should be fine.

Note, however, that once you have done "cmake ." it permanently messes up
the source tree and does not allow you to use a separate build tree. So you
must remove the contaminated source tree completely and start over with a
svn checkout before you attempt further builds using a separate build tree.

Frankly, separate build trees are really convenient since all you have
to do to get a fresh start is to remove the tree without worrying that
you are also removing source.  So separate build trees are mostly all
that we test.  But now you have drawn my attention to the issue, I will
try to solve it.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________