Using a second monitor, running R under Unix

Thanks.  I have of course been aware of the event loop issue.  But
does this also explain why the quartz window appears off screen?  I'd
not have expected that the event loop issue ought to generate the
"CGGStackRestore: gstack underflow" message.

John Maindonald             email: john.maindonald at anu.edu.au
phone : +61 2 (6125)3473    fax  : +61 2(6125)5549
Centre for Mathematics & Its Applications, Room 1194,
John Dedman Mathematical Sciences Building (Building 27)
Australian National University, Canberra ACT 0200.


On 17 Jan 2007, at 6:54 AM, Byron Ellis wrote:

Hi James,

This is because of the different between the Quartz and X11 Window
Server notion of How The World Works. Quartz event processing is an
in-process thing such that ALL events, including window dragging, must
be handled by the in-process event loop. When you start a Quartz
device when running in ESS there effectively isn't an event loop to do
the processing so you don't get to drag windows around. To top it off,
I believe the server may also differentiate between interactive and
non-interactive applications so despite getting access to the server
they can't receive events (but my memory here is hazy).

In Mojave I had a function that actually gave you an event loop by
putting the console reader onto a separate thread and pumping the
event loop on the main thread, which seemed to work pretty well. It
also used an undocumented call to give you an interactive application,
which had the unfortunate side effect of a weird (unusable) dock icon,
but let you use a graphics device. I'll try to dig up a copy of that
(I took it out of the svn Mojave for reasons I've forgotten :-) ) and
post it somewhere. IIRC you might also get a bonus quartz graphics
device I was experimenting with. :-)


On 1/14/07, John Maindonald <john.maindonald at anu.edu.au> wrote:

I am using a MacBook Pro with Core 2 Duo, with an external screen
connected that is set to appear above the MacBook Pro's screen.  The
laptop's menu appear bar is set to appear at the top of the external
monitor.

When I run R under ESS, or from the Unix command line, and type
quartz
(), the graphics window appears to the "left" of the external
monitor.  Hitting the F9 key brings it into view, but there seems no
way to grab it with the mouse and move it onto less ephemeral real
estate.  [Under R.app, all is well.]

[With x11() half the window is off to the left, and I can drag it to
anywhere I want (either screen).]

A further issue is that when I plot anything on the quartz screen, I
get an "error: message, thus:

 > quartz()

Warning message:
quartz() device interactivity reduced without an event loop manager
in: quartz()

 > plot(1:5, 1:5)

CGGStackRestore: gstack underflow.

The "CGGStackRestore: gstack underflow" message looks ominous, but it
does not cause evident problems.

Is there something I can do so that the positioning of the second
screen is recognized?  At some time in the past, I am sure that
quartz
() [ and x11()] used to correctly recognize the positioning of the
screens.  However it is a while since I used quartz(), and I would
not have paid too much attention to the x11() behaviour.
John Maindonald

 > sessionInfo()

R version 2.4.1 (2006-12-18)
i386-apple-darwin8.8.1

locale:
C

attached base packages:
[1] "splines"   "stats"     "graphics"  "grDevices" "utils"
"datasets"
[7] "methods"   "base"


John Maindonald             email: john.maindonald at anu.edu.au
phone : +61 2 (6125)3473    fax  : +61 2(6125)5549
Centre for Mathematics & Its Applications, Room 1194,
John Dedman Mathematical Sciences Building (Building 27)
Australian National University, Canberra ACT 0200.

--
Byron Ellis (byron.ellis at gmail.com)
"Oook" -- The Librarian