[ESS] bug in ess-remote?

[Vitalie Spinu]

Why is ESSR.rda fails to download? Sending files to remote shell is very
unreliable, and, quite frankly, I don't know the reason for that and
have no clue how to solve it. So placing ESSR.rda on the web was an easy
solution. You need to instrument ess--R-load-ESSR and see why ESSR.rda
fails. 

Tramp is stuck because of an error on process loading. C-g and then go
to the *R* buffer and see what is going wrong. Most likely R has not
been found.

Thanks for the reports, 

   Vitalie

 >>> chris warth on Tue, 26 Nov 2013 14:58:40 -0800 wrote:

 > Sorry to be contrary, but are you sure this is fixed in the case when download
 > of ESSR.rda fails? I’ve upgraded to what I believe is the latest development
 > version of ESS (svn checkout https://svn.r-project.org/ESS/trunk) and I still
 > see this problem with ess-remote. 

 > The code in ess-rutils.ell and ess-utils.el has clearly been updated, but it
 > still encounters what looks like a buffer overflow when sending files to a
 > remote R, even one that is running under a local shell. In the 'remote' R
 > process I see a bunch of garbage inserted that looks like various lines were
 > truncated (see the start of this thread for an example).

 > I instrumented ‘ess—inject-code-from-file’ to print the name of the file it is
 > injecting, and this is what I see after running ess-remote: 

 > Failed to download ESSR.rda. Injecting ESSR code from local machine
 > injecting /Users/foo/.emacs.d/ESS/etc/ESSR/R/.basic.R
 > injecting /Users/foo/.emacs.d/ESS/etc/ESSR/R/.load.R
 > injecting /Users/foo/.emacs.d/ESS/etc/ESSR/R/completion.R
 > injecting /Users/foo/.emacs.d/ESS/etc/ESSR/R/debug.R
 > injecting /Users/foo/.emacs.d/ESS/etc/ESSR/R/developer.R
 > injecting /Users/foo/.emacs.d/ESS/etc/ESSR/R/misc.R
 > Warning: Bug in inferior-ess-mode: it forgets to call `run-mode-hooks'

 > That looks fine, the problem is trash in the R process. Putting a (sit-for 1)
 > into ‘ess—inject-code-from-file’ after injecting each file seems to help. There
 > is still the bug of inferior-ess-mode not calling run-mode-hooks, but at least
 > ess-remote works for me now.

 > BTW, I tried using Tramp, but it just hangs for me when starting a remote R
 > session in aquamacs on Mavericks. I’m hopeful that debugging ess-remote will be
 > easier than debugging tramp. Besides which I don’t need a generalized solution
 > to remote files and processes in emacs, I just need a specific solution for R
 > that works when R is running on a remote machine.