[Rd] followup to graphics bug #414

[Paul Murrell]

hi


> > I set this up so that one line (red) would clip at the bottom only, one
> > line (green) would clip at the bottom and the top, and one line (black)
> > wouldn't need clipping at all. In R 1.0.0 everything works fine under
the
> > X11 driver but under the Postscript driver, both of the clipped lines
are
> > wrong, but are different depending on how the clipping parameter is set.
> >
> >   I haven't yet been able to track the source of this down, but I hope
> > that this information helps if someone gets to it before I do.  If the
> > recent fix also fixes this bug, that would be very nice.
>
> I suspect it will. The clipping functions are broken with logged axes
> on all devices which cannot do their own clipping, and since Paul made
> devPS capable of clipping...


or it may have just made it a little more sinister.  devices which can do
their own clipping still use the R clipping to preclip to the device extent.
so if you specify a line which extends outside of the device extent (and you
have log axes) it will be (incorrectly) clipped by R before sending on to
the device.

also, Ross has pointed out that R's polygon clipping probably suffers from
the same problem (has anyone tried drawing polygons with log axes ?).  He
may take a look at this soon because he's mucking about with polygon
clipping for his maps stuff.


> We still need a fix for clipping for 1.0.1. I assume that the new
> devPS has been changed too much to qualify as a bugfix (?) and
> anyways, it is still a bug for any other device without clipping
> capabilities.


devPS has not changed that much, but graphics.c has.

paul


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._