Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!elroy.jpl.nasa.gov!decwrl!claris!outpost.UUCP!peirce
From: peirce@outpost.UUCP (Michael Peirce)
Newsgroups: comp.sys.mac.programmer
Subject: Re: INIT/cdev communication question
Message-ID: <0B010004.moaxva@outpost.UUCP>
Date: 24 Mar 91 23:55:42 GMT
Reply-To: peirce@outpost.UUCP
Organization: Peirce Software
Lines: 42
X-Mailer: uAccess - Mac Release: 1.0.3


In article <91083.154101CXT105@psuvm.psu.edu>, CXT105@psuvm.psu.edu (Christopher Tate) writes:
> 
> Is there anything wrong with this scheme:
> 
> My INIT opens up a resource of a custom type and ID (chosen to minimize
> possible conflicts) which contains the data which needs to be shared by
> the INIT and the cdev.  Since the INIT file is first in the resource
> search chain, the INIT will always open the resource out of its own
> file.
> 
> The resource is MoveHHi'd into the top of the system heap and locked
> down, but *not* DetachResource'd.  The INIT gets various information
> from this resource when needed.
> 
> Now, the cdev calls GetResource() on that same resource type and ID.
> Since the cdev is in the same file as the INIT, it gets a handle to
> the resource image that's already stashed away in the system heap.
> By altering the proper fields and calling WriteResource(), the cdev
> can alter these persistent parameters both permanently and dynamically;
> the INIT will encounter the new values in the memory image the next
> time it looks.
> 
> This scheme maintains the distinction between code and data; nothing
> needs to be stored directly into the cdev or INIT resource images
> themselves.  Anyone see anything wrong with this approach?

Well, when the INIT is done running, its resource fork will be closed.
INITs don't run forever, they get a quick chance to run at startup
time and if they expect to be around later must stash parts of themselves
in the system somewhere.

What's wrong with creating a block in the system heap and using Gestalt
to register the pointer to it?  Then the INIT's resident code 
(patch/TimeManager task/whatever) and the cdev can ask Gestalt for
it.

-- michael

--  Michael Peirce         --   outpost!peirce@claris.com
--  Peirce Software        --   Suite 301, 719 Hibiscus Place
--  Macintosh Programming  --   San Jose, California 95117
--           & Consulting  --   (408) 244-6554, AppleLink: PEIRCE