On Wed, 13 Mar 2013, Peter W. Draper wrote:
> On Wed, 13 Mar 2013, Tim Jenness wrote:
>
> > On Wed, Mar 13, 2013 at 8:05 AM, Malcolm J. Currie <[log in to unmask]>
> > wrote:
> > PWD> So cd into JNIAST and try again.
> >
> > That worked fine on Snow Leopard and 64-bit CentOS with a
> > graphic displayed during the test phase. Thanks.
> >
> > So should the native-build be a regular part of the release
> > build steps,
> > if AST has new functions or APIs?
> >
> >
> > Well, we should do it for each release to track any bug fixes that David has
> > put into AST. Tracking new API is something we haven't done for ages.
> >
> > I can't actually remember whether we are meant to copy the latest AST source
> > files into the JNIAST tree or whether it always works by pulling apart the
> > installed library.
>
> You only do that if the JNIAST API support changes (right Mark?).
Yes, that's about right. The AST C source files are there to support
the ant build-source target, which uses the comments in the C
source files to contruct the javadocs (and I think some of the
java API code itself, can't remember) for the JNIAST java source
files. In principle updating the AST C source files in JNIAST and
re-running ant build-source in JNIAST could pull in improvements
made to the AST documentation, but the improvements are not likely
to be that important and the whole system is somewhat complicated,
so the chances of it breaking something in either obvious or
non-obvious way are not negligable - I don't recommend it.
--
Mark Taylor Astronomical Programmer Physics, Bristol University, UK
[log in to unmask] +44-117-9288776 http://www.star.bris.ac.uk/~mbt/
|