[ESS] Graphics in emacs window: how to copy R screen devices in non-interactive sessions (remotes)

[Yihui Xie]

Are you trying to reinvent the evaluate package? :)
http://cran.r-project.org/package=evaluate I think what you described
are basically what we did in evaluate:
https://github.com/hadley/evaluate/blob/master/R/eval.r

Regards,
Yihui
--
Yihui Xie <xieyihui at gmail.com>
Web: http://yihui.name
Department of Statistics, Iowa State University
2215 Snedecor Hall, Ames, IA


On Mon, Nov 18, 2013 at 7:40 PM, Vitalie Spinu <spinuvit at gmail.com> wrote:
>
> I was hoping to make it at least partially work with plot.new and
> before.plot.new hooks. But I cannot make it work even for completed
> plots.
>
> What about qtbase + qtutils? Any hope there for capturing incremental
> graphics? I installed it but cannot see anything useful.
>
>  > For incremental changes, it's easy to catch them at the device layer. Assuming
>  > implementing a custom graphics device is outside the scope, you would need to do
>  > something like take a snapshot each time R becomes idle, which would be a lot of
>  > overhead.
>
> Not really feasible. We cannot reliably know when R is idle.
>
>  > So you might need your own device that would print to stdout, with
>  > some prefix that hides it from the buffer. Not completely infeasible.
>
> It should be a modification of an existing device. That to say, instead
> of buffering the output, the device should flush the content into a file
> each time a major change has been made. The problem is how you define a
> "major change".
>
> Another idea is to change print methods for grid and a couple of
> internal plotting routines for base graphics to re-save the device each
> time a modification has been made.
>
>    Vitalie