>You can ignore most of my previous message.  I just got the
>CVS update mail where I noticed that you have made the changes
>already (plus others).  I would like to ask you a favor though.
>With the rapid changes going on currenly, I always worry that
>your changes are a day or two behind the rest of the changes
>announced in the daily update (or retrived via sup).  Would
>it be possible for you to post a message to the pc532-port
>mailing list indicating the sup-date of the rest of your sources
>whenever you check-in a new set of pc532-specific changes?

Ok, that sounds resonable.  I wasn't sure how much people were
following what was going on with the kernel changes.  Until just
recently, I was so far behind, there was almost no way anyone
could compile a kernel.  

I am still running about 1 to 3 days behind the current changes.
That is still the case.  On any day, a sup version will have changes
that may cause the pc532 port to not compile.  This process is slowing
down.  Do report problems to me just in case I have not caught them.

I usually try to update *only* when I can actually compile a working
kernel.  If a change makes the pc532 port not work, an update is usually
longer in comming.  I hope there are very few of these left in the update
process.

To help people who are interested catch up a bit, I just put a new
(May 19-20 sup + pc532 changes) kernel on sun-lamp.  *SEVERAL* older
binaries don't run under the new kernel.  (Older is defined here as
the previous binary snapshot!)  If you upgrade, make sure you don't
make your machine "unrunable".  (You could always restart by downloading
the ram.root and kernel from a serial line :)

The binary snapshot has most binaries from the -current source tree.  The
exceptions are :

    programs that have source in the NetBSD source tree

	a) gcc - cc1 and cc1plus are from 2.5.8, cc and cpp are from 2.4.5
		in the source tree.
	b) awk - the current sources when compiled on the pc532 have a
		*REAL* bad bug that causes programs that use awk in their
		build process to have incorrect build sequences. (bin/sh)
		The binary in this shapshot is an April 2 binary. (Yes, it
		does run.)

		The problem ... pipe input:

		echo a | awk '{printf("xxx %s\n", $1)}'

		should print "xxx a".  It does so when input from a file
		or a terminal, but not from a pipe.

		I have not had time to find the problem.  If someone here
		could find it, that would be great!

   programs that do not have source in the NetBSD tree.

	c) pc (gpc) gpc1 libgpc.a -- are 2.5.7 gpc source linked with
		gcc-2.5.8.  It still has a few known bugs, but it does
		quite well with standard Pascal and has a lot of stuff
		from Extended Pascal.

	d) Various other programs *Not* in the NetBSD source tree.
		(kermit, mg to name a couple.)


Not much else has happened with the distribution.  It should be quite similar
to the April version.

--Phil


------------------------------------------------------------------------------