classpath
[Top][All Lists]
Advanced

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

Re: classpath ./ChangeLog gnu/classpath/Configurati...


From: Tom Tromey
Subject: Re: classpath ./ChangeLog gnu/classpath/Configurati...
Date: 19 Jan 2003 18:22:08 -0700
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2

>>>>> "Mark" == Mark Wielaard <address@hidden> writes:

Mark> I thought you meant it to be a way for the VM to select which
Mark> AWT implementation to select.

That's definitely part of my intention.  A patch like that one will go
in libgcj eventually.  We already have a couple peer implementations
(the xlib one seems to have been partially resurrected).  I'd like to
set things up so that the default one is chosen by configure, and
ideally I'd like to share AWT 100% with Classpath.  Maintaining one
little divergence is no problem though.

Mark> And if that is the case then I would like to see one level of
Mark> indirection (either a System property or a VM method) to select
Mark> the desired Toolkit class.

Why is that?

I suppose one idea would be that you could compile Classpath once and
use it for all (non-libgcj) VMs.  That does make sense as a goal.
OTOH there is already one bit of Configuration.java that is
VM-specific, namely INIT_LOAD_LIBRARY.  But perhaps since it is fairly
gcj-specific, it doesn't count.

Mark> Currently only specific VMs might have different AWT
Mark> implementations and in that case I don't think adding this to
Mark> Configure.java.in is a good idea.

I'd like to understand this better.  And ideally I'd like the
rationale to be documented.

Tom




reply via email to

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