On Dec 23, 2010, at 2.14 , Alexander Hansen wrote: > The package families in question carry a BuildDepends hdf5|hdf5-18 . > This doesn't seem right, since (1) hdf5 and hdf5-18 have different APIs > and (2) the libraries have different install_names so they are not at > all drop-in compatible. Also, since hdf5-18 currently uses a private > install location for its libs and headers (%p/lib/hdf5-18/lib), to use > them one needs to specify this directory, and this isn't done in grass64. > > However, grass64* doesn't seem actually to link to hdf5-shlibs or > hdf5-18-shlibs. > > If the package just wants the HDF5 runtime executables, then it would be > appropriate to to change the BuildDependends to hdf5-bin|hdf518-bin. > Otherwise, if it actually _does_ build against hdf5*, then the > BuildDepends should pick one of hdf5 and hdf5-18 . Currently, however, > it doesn't appear really to be using hdf5* at all. > > [unrelated: Does it really need to use the ancient ncurses, rather than > the less ancient (and essential, so everybody has it) libncurses5? :-) ]
I was thinking to switch from python26 to python 27. With all these, in the next revision, I will change: BuildDepends: hdf5-18-bin | hdf5-bin libncurses5-dev, python27, (%type_raw[gui] = -x11) wxgtk2.8-py27-dev (>=2.8.1.1-1) Depends: << libncurses5-shlibs, python27-shlibs, (%type_raw[gui] = -x11) wxgtk2.8-py27 Thanks -- BABA Yoshihiko Fink Maintainer for GRASS ------------------------------------------------------------------------------ Learn how Oracle Real Application Clusters (RAC) One Node allows customers to consolidate database storage, standardize their database environment, and, should the need arise, upgrade to a full multi-node Oracle RAC database without downtime or disruption http://p.sf.net/sfu/oracle-sfdevnl _______________________________________________ Fink-devel mailing list Fink-devel@lists.sourceforge.net http://news.gmane.org/gmane.os.apple.fink.devel Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-devel