Nico Kadel-Garcia wrote on Sun, Jun 30, 2013 at 19:53:59 -0400:
> On Tue, Jun 25, 2013 at 9:40 AM, Daniel Shahaf <danielsh_at_apache.org> wrote:
> > On Mon, Jun 24, 2013 at 03:40:44PM -0400, Nico Kadel-Garcia wrote:
> >> The normal way to handle changing deployment environments is with an
> >> actual configuration or deployment tool, such as autoconf, that sets
> >> the path to the locally detected python and uses *that*.
> >
> > The detected Python (if any) is available in Makefile as $(PYTHON).
> >
> > (I have nothing to say to the rest that wasn't already said somewhere on this
> > thread (perhaps on the dev@ part of it).)
>
> Good point.
>
> Setting it in svnpubsub would then mean *generating* the files to be
> installed form the existing *.py files. Ideall for that, they'd be
> named "script.py.in" instead of "script.py" and built during the
> "make" process.
>
Or make could just change the first line during installation. That way,
people working on svnpubsub won't have to face the overhead of having to
change the .py.in file and generate the .py from it during development.
> But for now, unless the author of the svnpubsub tools cares to rewrite
> them, I think the wrapper script in the SRPM I've written will cover
> the RHEL setups.
Cool.
Received on 2013-07-01 16:59:51 CEST