--- Bill Page <[EMAIL PROTECTED]> wrote: > > Indeed, an ASDF based version of Axiom will completely enclose > > the whole issue into a proper lisp build framework and the > > whole build process can disappear. > > I simply don't believe this statement at all.
Other issues aside Bill, asdf is in fact quite powerful. Obviously the build process won't disappear entirely - what Tim means is that the dependence of Axiom's build on machinery external to the lisp environment will disappear. This is probably true to a fairly large extent, and I think will be a Good Thing for portability, simplicity, etc. Maxima is actually an excellent example of this in practice. Maxima builds using the configure make make-install routine, but most of the core logic of things like build order is contained in defsystem, which is a competitor/predecessor to asdf. It is possible to build Maxima entirely within a lisp environment - configure and friends act basically as convenience interfaces to the lisp build process, while also picking up a few non-lisp details like xmaxima. Now of course, the self-containment aspect is practical only for the parts of the system that are written in lisp, or a language implemented in lisp. Building other components in other languages like C are not going to be as straightforward, but hopefully we will be able to operate almost entirely in a lisp environment in the future and in that situation the potential to simplify matters is considerable (think the old cmucl port, for example). However, asdf will most likely require that Axiom work in ANSI lisp. So it's a long term measure. The beauty of Gaby's autoconf work is that it can handle the current situation, but can also gradually and gracefully be scaled back as more build logic gets added into lisp without loss of functionality. Eventually, what will hopefully happen is that configure and friends will become the "public face" to the lisp build process, as in Maxima today. If Windows has other "standard" build methods than autoconf we could trigger the lisp build process using those as well - making maintaining them a minimal amount of work. I think using asdf is a good idea and one of the exciting possibilities opened up by ANSI porting. When I played around with Axiom on sbcl, I tried using asdf and found it quite interesting. Unfortunately, it seems to result in a more finicky situation where code that would normally (e.g during a load) just give out warnings cause failures instead, which halted my progress at bookvol5.lisp. (Or maybe this is good, since it forces one to confront the issues - dunno.) Of course, that leaves organizing the code using asdf, which in turn involves the little detail of understanding what the code is doing... ;-) Cheers, CY ____________________________________________________________________________________ Low, Low, Low Rates! Check out Yahoo! Messenger's cheap PC-to-Phone call rates (http://voice.yahoo.com) _______________________________________________ Axiom-developer mailing list Axiom-developer@nongnu.org http://lists.nongnu.org/mailman/listinfo/axiom-developer
