Brian Jones <[EMAIL PROTECTED]> writes:
> "Aaron M. Renn" <[EMAIL PROTECTED]> writes:
>
> > Well, as promised I have fixed most of the manual hacks in the build
> > process. You can now actually build it from CVS with little or no problems,
> > at least on my computer...
> >
> > Here are updated instructions for building:
> >
> > -- You don't need JavaDeps at all.
> >
> > -- Get the latest source from CVS, do "aclocal; autoheader; automake;
> > autoconf"
>
> I could be wrong, but I don't think we use autoheader for anything.
>
> > Additionally, "make dist" now works. I want to fix one minor problem, then
> > I'll load a sample dist file to my ftp site for people to test. (The
> > problems is that make rebuilds all the Java classes even though I include
> > glibj.zip already in the package).
>
> Cool. Should we make compiling glibj.jar part of a dist-hook and
> remove it from the general 'make'?
Actually have two better thought out ideas.
1. Use dist-hook to cp -r java and gnu directories into the
distribution. This should keep 'make' from compiling anything in lib.
2. Change the if MAINTAINER_BUILD section to encompass compiling the
Java files and creating glibj.zip.
Anyone have a preference? The first option is affected by timestamps
after the end user untars the package.
Brian
--
|-------------------------------|Software Engineer
|Brian Jones |[EMAIL PROTECTED]
|[EMAIL PROTECTED] |http://www.nortel.net
|http://www.classpath.org/ |------------------------------
