R devel - Oct 2004 - Bug with png() and plot(type="n") (PR#7270)

Full_Name: Allen McIntosh
Version: 2.0.0
OS: Linux
Submission from: (NULL) (67.80.175.118)


Observed the bug with the following combinations: (RedHat 7.3, R 1.6.0) (RedHat
7.3, R 1.9.1) (Redhat 9, R2.0.0).

In the second plot (.png file) produced by the following sequence:

png(file="x%d.png")
plot(1:10,1:10,main="ooo",type="n")
segments(1:9,1:9,2:10,1:9)
plot(1:10,10:1,main="XXXX",type="n")

the axis labels, the tick marks and the title from the first plot have not been
erased.  The call to segments() seems to be necessary, as does
type="n".  The
example I extracted this from also called abline(), which was sufficient to
trigger the bug.  Inserting a call to frame() after the call to segments()
results in an extra file that ought to be blank, but instead contains the
offending text and tick marks.

Workarounds short of multiple calls to png() appreciated.

mcintosh@research.telcordia.com writes:
> Full_Name: Allen McIntosh
> Version: 2.0.0
> OS: Linux
> Submission from: (NULL) (67.80.175.118)
> 
> 
> Observed the bug with the following combinations: (RedHat 7.3, R 1.6.0)
(RedHat
> 7.3, R 1.9.1) (Redhat 9, R2.0.0).
> 
> In the second plot (.png file) produced by the following sequence:
> 
> png(file="x%d.png")
> plot(1:10,1:10,main="ooo",type="n")
> segments(1:9,1:9,2:10,1:9)
> plot(1:10,10:1,main="XXXX",type="n")
> 
> the axis labels, the tick marks and the title from the first plot have not
been
> erased.  The call to segments() seems to be necessary, as does
type="n".  The
> example I extracted this from also called abline(), which was sufficient to
> trigger the bug.  Inserting a call to frame() after the call to segments()
> results in an extra file that ought to be blank, but instead contains the
> offending text and tick marks.
> 
> Workarounds short of multiple calls to png() appreciated.
Apparently, the clipping of segments is still active when the drawing
surface is erased.

Inserting a call to box() before the 2nd plot seems to help. 

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk)             FAX: (+45) 35327907

On 8 Oct 2004, Peter Dalgaard wrote:
> mcintosh@research.telcordia.com writes:
> 
> > Full_Name: Allen McIntosh
> > Version: 2.0.0
> > OS: Linux
> > Submission from: (NULL) (67.80.175.118)
> > 
> > 
> > Observed the bug with the following combinations: (RedHat 7.3, R
1.6.0) (RedHat
> > 7.3, R 1.9.1) (Redhat 9, R2.0.0).
> > 
> > In the second plot (.png file) produced by the following sequence:
> > 
> > png(file="x%d.png")
> > plot(1:10,1:10,main="ooo",type="n")
> > segments(1:9,1:9,2:10,1:9)
> > plot(1:10,10:1,main="XXXX",type="n")
> > 
> > the axis labels, the tick marks and the title from the first plot have
not been
> > erased.  The call to segments() seems to be necessary, as does
type="n".  The
> > example I extracted this from also called abline(), which was
sufficient to
> > trigger the bug.  Inserting a call to frame() after the call to
segments()
> > results in an extra file that ought to be blank, but instead contains
the
> > offending text and tick marks.
> > 
> > Workarounds short of multiple calls to png() appreciated.
> 
> Apparently, the clipping of segments is still active when the drawing
> surface is erased.
Yes, and a fix has now been committed to R-patched/R-devel.

-- 
Brian D. Ripley,                  ripley@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

On 8 Oct 2004, Peter Dalgaard wrote:
> mcintosh@research.telcordia.com writes:
> 
> > Full_Name: Allen McIntosh
> > Version: 2.0.0
> > OS: Linux
> > Submission from: (NULL) (67.80.175.118)
> > 
> > 
> > Observed the bug with the following combinations: (RedHat 7.3, R
1.6.0) (RedHat
> > 7.3, R 1.9.1) (Redhat 9, R2.0.0).
> > 
> > In the second plot (.png file) produced by the following sequence:
> > 
> > png(file="x%d.png")
> > plot(1:10,1:10,main="ooo",type="n")
> > segments(1:9,1:9,2:10,1:9)
> > plot(1:10,10:1,main="XXXX",type="n")
> > 
> > the axis labels, the tick marks and the title from the first plot have
not been
> > erased.  The call to segments() seems to be necessary, as does
type="n".  The
> > example I extracted this from also called abline(), which was
sufficient to
> > trigger the bug.  Inserting a call to frame() after the call to
segments()
> > results in an extra file that ought to be blank, but instead contains
the
> > offending text and tick marks.
> > 
> > Workarounds short of multiple calls to png() appreciated.
> 
> Apparently, the clipping of segments is still active when the drawing
> surface is erased.
Yes, and a fix has now been committed to R-patched/R-devel.

-- 
Brian D. Ripley,                  ripley@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