classpath
[Top][All Lists]
Advanced

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

Re: Classpath building (was re: Classpath future?)


From: Brian Jones
Subject: Re: Classpath building (was re: Classpath future?)
Date: 12 Jul 2001 19:44:50 -0400
User-agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7

"Nic Ferrier" <address@hidden> writes:

> >>> Brian Jones <address@hidden> 12-Jul-01 11:05:22 PM >>>
> 
>    A number of projects do just put everything on 
>    the command line.  I'm unsure what exactly is 
>    "fixed" by making this change.
> 
> No. I don't see this as an issue either. 
> 
> I think Aaron was responding to my suggestion of clearly separating
> the build process into 3 steps:
> 
> 1. build java classes
> 2. build native header files
> 3. build native code

Okay, I guess my confusion is in that I already view the build in this
way.  I admit that we haven't provided separate targets to do just one
of these, but it is so close!  The build is thus (I'm overly
simplifying for clarity):

cd lib
build java classes
build native header files
cd native/*
build native code

Having targets which short-circuit this process are simple.  Have fun
and add them.  Just recognize that the user should be able to perform
the following (when we release a stable non-alpha version) and it
should work!

tar -xzvf classpath-1.00.tar.gz
cd classpath-1.00
./configure
make

That's the goal and we're not there.  That's the problem I see with
the build.

> When I suggested that it was based on the assumption that some VMs
> other than GCJ might want to implement their own native code systems.
> I was not aware that KNI was being removed (the message I got from the
> kaffe people [about 6 weeks ago] implied that KNI was still the
> preferred option for Kaffe specific native code).

I could be full of it; my take was based on a message I saw from Tim
on the pocketlinux mailing list about trying to get rid of all the KNI
code.

=-=-=-=-=-=
Tim Wilkinson

    * Spent a very happy week making all the old Kaffe Native
      Interface stuff go away and replacing it with JNI (now JNI is
      lots faster then it use to be).  Next stage is to remove all the
      support from Kaffe.

      This task seems to have broken stuff for lots of people and
      probably made me no friends at all :-(
=-=-=-=-=-=
-- 
Brian Jones <address@hidden>



reply via email to

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