[ITP] astrometry.net-0.38-1

[Jussi Kantola]

On Thu, Nov 3, 2011 at 2:02 PM, Corinna Vinschen wrote:
> Is it really necessary to have all the binaries as user-accessible
> binaries in /usr/bin?  Or are many of the binaries just called from
> another (or other) binaries which serve as the primary UI?

The latter was true.  I updated the packages as per your instructions:
the actual binaries are now in /usr/lib/astrometry/bin, whereas two
user interface commands are PATH-setting shell scripts,
/usr/bin/solve-field and /usr/bin/wcsinfo.  Solve-field is the actual
astrometry.net solver, whereas wcsinfo is used to extract World
Coordinate System data from the output generated by solve-field.

There may be more commands that should be available in the default
PATH; I'm sure the astrometry.net community will notify me if this is
the case.

> There's also still the issue with usr/lib/astrometry/libbackend.so.
> If that's really a shared lib, it should be called cygbackend.dll
> and should reside in /usr/bin.  If the binaries are using it, they
> should also be linked against it, rather than being linked statically.

I dug a little deeper into it, and to my best knowledge the .so was
being used only by two Python-scripts, which weren't even installed by
the Makefile.  So I'm assuming it's either a WIP or obsolete code.  I
modified the package source (and Makefile) to reflect Cygwin's naming
scheme, however, the shared lib is _not_ installed anymore (neither in
the binary package, nor by the Makefile in the source package).  I
haven't noticed anything breaking up due to this.

Package filenames are the same, but the contents have been updated.

wget \
http://rubor.org/astrometrytortilla/astrometry.net/astrometry.net-0.38-1.tar.bz2
\
http://rubor.org/astrometrytortilla/astrometry.net/astrometry.net-0.38-1-src.tar.bz2
\
http://rubor.org/astrometrytortilla/astrometry.net/setup.hint

-- 
jussi