Norman,
POLPACK cannot be built without tcl since polsub/polpack.c
includes <tcl.h> and invokes various Tcl_... functions (needed to
interpret the tcl scripts used to communicate with the GAIA polarimetry
toolbox). The configure.ac contains:
dnl Look for tck/tk system
STAR_PATH_TCLTK(8.2, [tk])
dnl POLPACK source files depend on include files from the following components
STAR_DECLARE_DEPENDENCIES([build], [ast cat chr cnf hds fio grp mers ndf
ndg par prm sae tcl trn ifd])
dnl The POLPACK link script links against the following components
STAR_DECLARE_DEPENDENCIES([build], [sla nbs kaplibs shl tcl], [link])
dnl Using POLPACK requires the following components to be available at run-time.
STAR_DECLARE_DEPENDENCIES([use], [kappa ccdpack tcl tk])
(although I've only just added tcl to the list of link dependencies).
Any advice gratefully received!
David
> Looking at the log file, I see a TCL_MISSING=1 definition -- presumably
> polpack is written to handle Tcl not being there, which is why it
> hasn't failed before (gross speculation). In which case I don't know
> why it's suddenly failed. Whose is polpack? Ah,
> <http://cvs.starlink.ac.uk/~nxg/componentset.html> says it's DavidB's
> -- what have you been up to, then, David?
>
> > I thought we had a deterministic
> > build system now?
>
> It won't be any more deterministic than `make'. And success won't
> generally imply that the network of dependencies is correct. It's
> Peter's four-day effort to build each component in isolation that
> proved that the network of dependencies was at least not too small (it
> doesn't indicate that it's minimal). That does suggest that polpack
> could handle Tcl being absent, which comes right back to the original
> problem.
>
> See you,
>
> Norman
>
>
> --
> ----------------------------------------------------------------------
> Norman Gray : Physics & Astronomy, Glasgow University, UK
> http://www.astro.gla.ac.uk/users/norman/ : www.starlink.ac.uk
>
|