[ESS] Displaying plots in Emacs

[Vitalie Spinu]

 >>> Michael Lawrence on Tue, 19 May 2015 08:04:28 -0700 wrote:

 > What about the httpuv R package?

Their implementation is hardwired towards http parsing and TCP sockets
are not exposed. I think it's a bit of a shortsighted engineering on
their side. The implementation of httpuv could have been considerably
more modular. I plan to use it for websockets transport though.

 >  My workaround idea is to have one R master server and one R slave
 >  fork for each new client instance, The master will take care of
 >  routing client requests to slave forks and will also be responsible
 >  for tooling requests which are different from eval requests
 >  (completions, help etc). That would allow a pretty good amount of
 >  non-blocking communication and should be more than enough for most
 >  needs.

 > What about just making an Emacs client for Rserve?

I haven't investigated that option thoroughly but Rserve seems to be too
low level, too bulky, too hard. There is no way you can work with bits
and pieces of that level from emacs. Nor you need to access all the R
object structures for the purpose of IDE. The last time I looked into
its java client it cost me some hair. Correct me if I am wrong, but it
also works on local machines only. And it seems to allow only one
connection per R instance. So it's out of question AFAIC.

I also have personal reasons for nREPL. I do Clojure and I plan to
intensively develop for CIDER. So having two similar systems to deal
with is a big win for me personally.

  Vitalie