[coyotos-dev] New binutils, gcc, crtfiles, newlib

[Sam Mason]

On Fri, Apr 11, 2008 at 07:02:13PM -0400, Jonathan S. Shapiro wrote:
> I think I know what is happening, but I'm not going to be able to
> address this tonight. The server rotation needs to be completed by
> tomorrow AM, and until the domain name changes propagate at verio there
> won't be a reliable way to get an update into the tree that you can
> retrieve.

OK, have fun with the upgrade! they tend to be much more involved than
I ever expect.  I've also got another question, when you get another
minute in your ample free time :)

When I've tried to boot Coyotos recently, I've wondered why the
HelloWorld test doesn't actually print out "hello world".  I've just
(not sure why I didn't think of this earlier) looked through the code
and found a console_detach() procedure that is run before the "first
user mode instruction".  For some reason I'd always been looking for
something explicit in the user-mode code so far.  That explains where
the message goes, I'll just have to wait until it starts running again
to have a look for the message_buffer and watch things appended to it.
Is it planned to expose read access to this?  I've noticed the KernLog
interface, but don't see how to get anything back out again.

I'm tempted to try and hack something together that gets the current
contents of the kernel message buffer and dumps it off to the
TextConsole.  If that works I could maybe arrange for the request to
block until the buffer gets half full and then returns.  Any comments
on whether that could be a good way to get a bit more understanding of
the kernel would be appreciated.  I'm thinking I could do something like
the interval stuff.  Just to confirm, I don't expect this functionality
to end up in the kernel proper, I just think it could be an interesting
exercise.


Thanks,
  Sam