I agree here. I've been looking at what we do now in Xerces and Xalan and it seems that think's are simple enough to get away with pure make.
Which would remove yet another dependency. Having seen that, I _am_ impressed by the way Ant works. See the JSP and Tomcat dist's. And given that it is small, and thus can be part of the source distribution, with a small sh-script to kick it off. This would be my second choise. Autoconf style ./Configure or gmake rank a lot lower. Dw On Wed, 10 Nov 1999, Steve Suehring wrote: > I would much rather stick with make. From an implementation standpoint > it's much easier to debug make on the end-user level than it is to work > with another tool. > > Steve > > > On Wed, 10 Nov 1999, Sander van Zoest wrote: > > > On Wed, 10 Nov 1999, Dirk-Willem van Gulik wrote: > > > > > > Stefano pointed out the existance of a Make tool implemented in > > > > Java, which would be more than fine with me. I don't know how big the > > > > download for this is, but it's worth investigating. As far as jar > > > > files > > > > go, that would be fine with me as well. > > > Another option would be to stick to pure 'make'. As in a subset of 'gmake' > > > :-). > > > > I would vote for a pure 'make' install mechanism, so the FreeBSD version > > could simply use the native pmake rather then requiring people to install > > gmake or any other build tool. > > > > If we do require another build tool, it would be nice if it came with the > > tarball sort of like ralfs' APACI for the apache http server. > > > > -- > > Sander van Zoest [EMAIL > > PROTECTED] > > High Geek (858) 623-7442 > > MP3.com, Inc. http://www.mp3.com/ > > > >
