From: Godmar Back (gback@cs.utah.edu)
Date: Wed Sep 09 1998 - 20:26:12 EDT
>
> Now I understand the implicit assumptions you're referring to.
> However, does it make sense to *not* do it this way? That is,
> if your purpose is to regenerate derived files so you can
>
> (a) Check them in, or
> (b) Create a tarball
>
> then you would want to do this inside the source tree. The main
> motivation for "make derived-files" is for a developer to update
> all of the *source* files that are derived from other source files.
I never configure in my source tree, and I don't see why I should
be required to start doing that for derived files.
I think we got feedback from the mailing list that was in favor of
having a complete environment available via CVS, so let's respect that.
For instance, I checked out japhar. They don't even have a configure
file in their CVS. The result: it doesn't work (aclocal fails), despite
the fact that I have the required versions of automake and autoconf
installed. So much for that piece of GNU wonderware. I sent them
mail asking about it.
Actually, if we assume for a moment that we do have a compatible
Klasses.jar file in the CVS repository, then we should be able to first
build kaffeh, and then create the derived header files. This would avoid
the need of having them in CVS.
- Godmar
This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:56:56 EDT