[R-SIG-Mac] plot() does not give labels under quartz

[Prof Brian Ripley]

On Wed, 19 Jan 2011, Kasper Daniel Hansen wrote:

> This _may_ be unrelated to a corrupt font cache.  Note that the OP
> says the issue is fixed by resizing the quartz screen.
>
> To see this in action (explanation below) I have posted
>  http://www.biostat.jhsph.edu/~khansen/screen1.png
> (the initial look of the quartz window)
>  http://www.biostat.jhsph.edu/~khansen/screen2.png
> (after resize)

But the OP said (s)he had 'no numbers under the tics', whereas what 
you show is explicable as the effect of clipping and the axis 
annotation is there.  I don't think it is the same issue, and I see no 
reference to resizing in the original post (the solution was to use 
X11()).

> I have always had a (somewhat) similar experience: the first quartz
> window does not show the entire plotting area.  If one looks closely
> (in my case) you can see the top of the xlab in the plotting window,
> but there are no scroll bars indicating that the plot is bigger than
> the screen estate, despite the fact that the little icon in the lower
> left corner indicates that this is really the lower left corner (I am
> referring to the "resize triangle" - no idea what the technical term
> is).
>
> When I click the green button (maximize) twice I now get a window
> where I can see everything (screen2).
>
> I have also assumed that this is because the default size of the
> quartz device was bigger my laptop's screen (physical size, I am on a
> Macbook right now).  This is not an issue on an external monitor.
> Specifically I have assumed that the window manager has problems
> initializing a window that is physically bigger than the  screen (note
> that I say initializing - of course the window can be bigger than the
> physical screen if I resize it to be so...).  This is pure
> speculation, I have no basis for these claims.  I have always assumed
> that this was a problem with Apple's SDK and not R.  And while I have
> the issue I have always felt it was not worth spending time fixing (if
> it is even fixable by someone else than Apple).
>
> Now, the OP can see even less than I can.  Perhaps she is using a new
> Macbook Air with even less vertical physical size?  Please respond to
> this Tena.
>
> Finally, this was using R-2.12.1, compiled by myself, running from the
> terminal but using the quartz device.
>
> Kasper
>
>
> On Wed, Jan 19, 2011 at 6:22 AM, Prof Brian Ripley
> <ripley at stats.ox.ac.uk> wrote:
>> Missing labels has been reported several times, and AFAIK has always
>> indicated a corrupt font cache (or other problems in the font book). There
>> are various ways to repair that, so I'll leave you to Google for one.
>>
>> On Wed, 19 Jan 2011, Phillip Jardine wrote:
>>
>>> Tena,
>>>
>>> I had the same problem myself a few months ago, when I upgraded to OS X
>>> 10.6.5 (from OS X 10.4) I don't know why this happened, but by trial and
>>> error I found that putting family = "sans" into the plot command brought
>>> back the labels.
>>>
>>> Bizarrely though things recently started working normally again, with no
>>> updates/upgrades on my part. Strange, eh?
>>>
>>> Anyway, try the family = "sans" thing, it might help.
>>>
>>> Best wishes,
>>>
>>> Phil
>>> --
>>> Phil Jardine
>>> Geosystems Research Group
>>> School of Geography, Earth, and Environmental Sciences
>>> University of Birmingham
>>> Edgbaston
>>> Birmingham
>>> B15 2TT
>>>
>>> pej083 at bham.ac.uk
>>> Tel. 0121 414 6146
>>> http://www.gees.bham.ac.uk/staff/pgpej083.shtml
>>> ________________________________________
>>> From: r-sig-mac-bounces at r-project.org [r-sig-mac-bounces at r-project.org] On
>>> Behalf Of Tena Sakai [tsakai at gallo.ucsf.edu]
>>> Sent: 19 January 2011 04:29
>>> To: r-sig-mac at r-project.org
>>> Subject: [R-SIG-Mac] plot() does not give labels under quarts.
>>>
>>> Greetings,
>>>
>>> I have noticed a bit odd behavior of plot() on OS X 10.6.5.
>>> The R I downloaded from cran is R-2.12.1.pkg.  Below is
>>> what I did:
>>>
>>>       R version 2.12.1 (2010-12-16)
>>>              .
>>>              .
>>>       [R.app GUI 1.35 (5665) x86_64-apple-darwin9.8.0]
>>>
>>>       > foo <- c( 4, 6, 9 ,12 )
>>>       > moo <- c( 34, 37, 41, 49 )
>>>       > plot( foo ~ moo )
>>>
>>> This gives me a plot with no labels.  No "moo", "foo", or numbers under
>>> tics.  But,
>>>
>>>       > X11()
>>>       > plot( foo ~ moo )
>>>
>>> This gives me xlabel ("moo"), ylabel ("foo"), and numbers under tics.
>>> In the first case, by default, I am using quartz, while the second,
>>> obviously, I am using X11.
>>>
>>> Does anybody know why this happens under quartz and not under X11?
>>>
>>> Thank you.
>>>
>>> Tena Sakai
>>> tsakai at gallo.ucsf.edu

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595