Re: VMMaker( Re: KCP & 3.6)

On Sat, Jun 21, 2003 at 12:46:26PM +0200, Andreas Raab wrote:
> 
> Neither in the current build process - if you use VMMaker (and after all
> that's what it's for ;-) See my comment above - I got so totally used to
> controlling everything from VMMaker that the only thing that I'm still
> missing is a big green "build" button in it (which invokes the build
> process, saves the image and restarts with the new VM). Heh, heh ;-)

Actually, I did that a while ago with OSProcess, and it worked fine.
You could click one button on the VMMakerTool, VMMaker did its thing,
the VM would be rebuilt, and if the make was successful Squeak would
restart itself with the new VM and/or plugins. All one click, and rather
entertaining to watch it go.

That was just for Unix/Linux at the time, but I think it would work on
Windows with a bit more hacking. I don't have time right now, but maybe
I'll get this working again one of these days.

Dave
 
From [email protected] Sat Jun 21 17:01:13 2003
Return-Path: <[email protected]>
Delivered-To: [email protected]
Received: (qmail 27148 invoked from network); 21 Jun 2003 17:01:12 -0000
Received: from mail531.nifty.com (202.248.37.220)
  by mail.theinternetone.net with SMTP; 21 Jun 2003 17:01:12 -0000
Received: from YOSHIKI-T2.mb.infoweb.ne.jp
	(66-214-136-65.gln-eres.charterpipeline.net [66.214.136.65])by
	mail531.nifty.com with ESMTP id h5LH15dw009365;
	Sun, 22 Jun 2003 02:01:06 +0900
Date: Sat, 21 Jun 2003 10:00:56 -0700
Message-ID: <[email protected]>
From: [email protected]
To: Discussing the Squeak Foundation
	<[email protected]>
Subject: Re: [Squeakfoundation]re: TrueType font support and 3.6
In-Reply-To: <002101c33801$df1bce10$8ef7fea9@atlantis>
References: <[email protected]>
	<002101c33801$df1bce10$8ef7fea9@atlantis>
User-Agent: Wanderlust/2.10.0 (Venus) SEMI/1.14.5 (Awara-Onsen) FLIM/1.14.5
	(Demachiyanagi) APEL/10.5 Emacs/20.7 (i386-msvc-nt5.1.2600) MULE/4.1 (AOI)
	Meadow/1.15 (SHOUBU:63)
MIME-Version: 1.0 (generated by SEMI 1.14.5 - "Awara-Onsen")
Content-Type: text/plain; charset=US-ASCII
X-BeenThere: [email protected]
X-Mailman-Version: 2.1
Precedence: list
Reply-To: Discussing the Squeak Foundation
	<[email protected]>
List-Id: Discussing the Squeak Foundation
 <squeakfoundation.lists.squeakfoundation.org>
List-Unsubscribe: <http://lists.squeakfoundation.org/listinfo/squeakfoundation>,
	<mailto:[email protected]?subject=unsubscribe>
List-Archive: <http://lists.squeakfoundation.org/pipermail/squeakfoundation>
List-Post: <mailto:[email protected]>
List-Help: <mailto:[email protected]?subject=help>
List-Subscribe: <http://lists.squeakfoundation.org/listinfo/squeakfoundation>,
	<mailto:[email protected]?subject=subscribe>
X-List-Received-Date: Sat, 21 Jun 2003 17:01:13 -0000

  Andreas,

> >   If you get the MessageTally, 70% of time goes to the
> > BalloonEngine>>copyLoopFaster.  This is why I think Arjen's new BB
> > combinationRule will help here a lot.
> 
> How so? The time spent in #copyLoopFaster exclusively goes into the
> anti-aliased rasterization of the vector shape in which BitBlt isn't
> involved at all. I don't see how some new combination rule would help.

  I was plain wrong...  What I was thinking?  This test was to create
the glyphs in cache.

  In more realistic examples, where character glyphs are reused, this
combination rule:

    sourceForm: 8 bit alpha mask
    halftone: color
    destForm: destination

will help the BB part because (I think) it reduce the memory traffic
and unnecessary cache flushing at text color change.

-- Yoshiki

P.S.
  This is now more squeak-dev topic.  If it continues, let's move to
there.