Newsgroups: comp.lang.clos
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!parc.xerox.COM!Gregor
From: Gregor@parc.xerox.COM (Gregor Kiczales)
Subject: Re: PCL troubles
Message-ID: <91Jun26.092812pdt.29186@tracer-bullet.parc.xerox.com>
Sender: Gregor Kiczales <gregor@parc.xerox.com>
Organization: The Ohio State University Department of Computer and Information Science
References: <9106260320.AA10456@churchy.gnu.ai.mit.edu>
Date: Wed, 26 Jun 1991 16:27:58 GMT
Lines: 24

   Organization: The Ohio State University Department of Computer and Information Science
   Date:	Tue, 25 Jun 1991 20:20:42 -0700
   From:	whatis@gnu.ai.mit.edu (....What Is?....)

   It works fine for a while, then it tries to compile "gazonk0.lsp"
   over and over. 

This might not actually be a problem.  Have you tried giving it more
time to finish?

PCL, when it is run, needs to compile a bunch of little pieces of code
"on the fly."  These pieces of code are what implement the method
dispatch for each "kind" of generic function.  So, PCL ends up calling
the COMPILE function a lot, particularly when it is started up.  Now, in
KCL, where the Lisp compiler is built on the C compiler, the COMPILE
function first writes the code out to a file called gazonkn, and then
compiles that file.  So, all you are seeing is the normal behavior of
PCL as it finishes compiling itself.

Once you get PCL compiled, you can eliminate some of this overhead the
next time you load it using the PCL-specific utility PRECOMPILE-RANDOM-
CODE-SEGMENTS.  See the notes.text file for information about this.

Gregor