Amiga 3000UX first impressions

[Keith Gabryelski]

In article <NKKw6uca at visix.com> brett at visix.com (Brett Bourbin) writes:
>Ok, if the machine and OS is now public, I guess I can comment for the
>software developers point-of-view.

The machine is not public.

>For a CISC processor version of UNIX which are suppose to be smaller
>than their RISC counterparts, AMIX seems to be alittle bloated.  On a
>9Meg system (8 fast, 1 chip), the UNIX when booted only leaves me
>with alittle over 5Megs.  This is before the X Window System.

This is with the kernel fully loaded with NFS, RFS, RPC, and several
other packages you may or may not use.  A later release has kernel
linking options where you can add/remove packages at will.

>I don't know if I agree not to put color into the console X server.
>The reason I heard was about speed and I really don't fully believe
>it.

The fact is the console X server does not yet have color (well,
atleast not on the distribution).  We are working on it.  It will also
be slower.  The U of Lowell board offers a nice alternative to native
color X server.

>The OPEN LOOK window manager has some problems with window placements.  I
>don't think they are handling the USPosition and PPosition hints.

I've seen no bug reports from you on this.  Please report this bug
however you have been asked to report such things or send them it to
bugs at amix.commodore.com or me directly.

>The X server has some major drawing problems if you interrupt it when it
>is at work using the window manager (dragging a window while it is being
>drawn).

This is fixed in the lastest release.  Again I saw no bug report.

>I don't know what they are sync'ing the mouse pointer with for X, but it
>is very unstable.  I will venture a guess that it is not a sprite.  8^)

The speed of the pointer is faster in the current release.  An Amiga
sprite would not make a good pointer because it is low res.  Sure,
there are tricks to make it look high res but you lose precision.  I
agree the pointer could atleast be sync'd with beam.  You are also able
to make suggestions through the bug report channels.  I've seen none
on the mouse pointer.

>The good points are that it seems pretty complete as SVR4 goes.  Besides
>a major bug in the socket library (recvfrom () call) and the X problems,
>it is pretty stable.  NFS works like a charm, along with TCP/IP (this was
>not the case for the earlier beta versions back 2-3 months ago with kernel
>panics on NFS writes).

Again, I've seen no bug report of this recvfrom() problem.  If it is
`major' problem I would think people would want us to fix it by
the release.

Pax, Keith