Python bindings not regenerated in 1.8.0RC1

[Even Rouault]

The Driver.CopyFiles() interface has been exposed to swig, but the python bindings in 1.8.0RC1 have not been regenerated and don't include it consequently, causing the new HFA test to fail.

I'm also wondering about what should be done for the Perl bindings. Ari removed the generated files from SVN, but shouldn't they be generated at the time of building the distributed .zip / .tar.gz ?

[warmerdam]

Doh!

The nightly source preparing script does all this, and I mistakenly thought it was part of mkgdaldist.sh but it is not.

I will need to generate an RC2 shortly. Perhaps I'll wait to the end of the day in case anyone notices other major botches.

[warmerdam]

I have determined that mkgdaldist.sh *does* regenerate the perl and C# bindings, but not the Python which are also not regenerated for nightly snapshots. I have commited regenerated bindings in trunk (r21493) and 1.8 (r21494).

I have also found that the perl bindings generation has not worked for some time since it depends on running "make generate" which only works if GDALmake.opt exists which is never true for mkgdaldist.sh since it is always working in a fresh svn extract.

I have modified the mkgdaldist.sh script to very temporarily create a GDALmake.opt (r21495).

I have prepared and will announce an RC2 shortly with the updated python bindings and included perl bindings.