On Tue, 19 Apr 2005, John Reese wrote:
On 4/15/05, Andi Vajda <[EMAIL PROTECTED]> wrote:
Well, PyLucene's Makefile is currently 924 lines long. I do not know (I assume not) that distutils can do everything that the Makefile does. It was suggested earlier that a distutils script could wrap the 'make' task. If that is indeed possible and if that distutils script could also pass the correct env vars values to the Makefile, then editing the Makefile would no longer be necessary and that would be progress indeed. Patches are welcome :)
The Makefile may be long, but it's all just conditionals and variable assignments. A distutils script doesn't have to worry about dependency information, which is most of what a Makefile does... setup.py can just build everything.
In any case, a distutils script can't wrap the Makefile right now because the Makefile conflates the concept of python's install directory and the target install directory of the package. I'll try to prise these concepts apart in the Makefile, defaulting of course to the current behavior, and submit a patch.
Great, this would get us one step closer to a 1.0 release !
As for the security provider path in the %init block of the .i file, just to be clear... are we talking about the value that gets stored in the gnu.classpath.home.url property in line 4263 of PyLucene.i in the subversion trunk, r188?
Yes, that very one.
Andi.. _______________________________________________ pylucene-dev mailing list [email protected] http://lists.osafoundation.org/mailman/listinfo/pylucene-dev
