[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Axiom-developer] clisp and sbcl

From: C Y
Subject: Re: [Axiom-developer] clisp and sbcl
Date: Tue, 22 May 2007 21:53:33 -0700 (PDT)

--- Bill Page <address@hidden> wrote:

> Cliff,
> Just a quick note: I would strongly recommend that you do an
> out-of-source build - especially if you are building multiple
> targets. 

Heh - whoops.  Good idea, but I'm already a ways into the sbcl build
after doing several attempts at sbcl, clisp, and gcl in the same tree. 
Something to try if something goes wrong though, thanks!

> At least that way you wont be stepping on the toes of the
> deterious of a previous build. I am always suspicious about
> how well 'make clean' might or might not work. 

Actually, if wh-sandbox is put into git it should be able to spot any
un-cleaned-up files for us.  It would mean a manual cleanup until make
clean does work properly, but from that standpoint it's a very handy
way to debug make clean :-).

>From what I understand thus far about git we wouldn't want to stick all
of wh-sandbox's changes into a git branch and try to merge whole hog
with silver, as we couldn't "pick and choose" from such a merge.  I
think this is one reason Tim is advocating more "community" development
- with smaller, compartmentalized divergences from the "silver" or
"gold" trees we could use git branches to very, very efficiently manage
this process.  I know at this early stage it's not really workable
since some of the changes simply can't be minor and leave the build in
a workable state (e.g. autoconf), but I hope later on we can make use
of this, especially for Algebra development.  It really is a very, very
interesting way of doing work and I can see why Linus (who, IIRC,
advocates small changes that are easily understood as opposed to sudden
large scale changes) would want it like this.

Tim, as an aside, how would you suggest handling the ANSI changes?  1st
merge enough make changes to be able to try building on sbcl (even
though we know it will flop) and then the source code changes that make
it work?


 Yahoo! in your own image.  Join our Network Research Panel today! 

reply via email to

[Prev in Thread] Current Thread [Next in Thread]