ess-eval-buffer is omitting text about ess HOT 5 CLOSED

Conor Nash [email protected]
on Mon, 31 Dec 2012 14:30:51 -0800 wrote:

[...]

R> R> + Error: unexpected symbol in:

"ggplot(uniquefull[Widgets.for.Session < 500]) + geom_freqpoly(aes(x=Widgets.for.Session, y=..count..,
group=ktTestCggplot(uniquefull[Widgets.for.Sessi
print"`

I don't see ktTestCggplot anywhere in the code you have provided. That
can mean either you are sending some other code than the one you have
pasted, or ESS inserts arbitrary text into the text, which is very
unlikely.

So what text is omitted? Can you please find a reproducible example? I
am pretty sure ESS cannot omit any text, especially with your settings.

[...]

(require 'ess-eldoc)

^^^^^^ this is not needed since long :) It is activated by default.

[...]

   (auto-complete-mode t)

Also not needed ^^^^^^^^^^^^^^^. In 12.09 auto-complete is activated by
default.

   (define-key inferior-ess-mode-map "\C-ct" 'ess-R-object-tooltip)

In 12.09-1 you have C-c C-d C-e - a cycling command for object
description.

from ess.

Hey vitoshka,

Thanks for your fast response. Apologies for the confusion - I left a typo in my initial message that I have since fixed. I have also removed the redundant ELisp code from my setup, as you suggested.

The error appears as TestCggplot. From what I can see the second line of code (first line beginning with ggplot) is being truncated for some reason and the next line appended immediately afterward. This line is also truncated at the point of ggplot(uniquefull[Widgets.for.Sessi and the line terminated.

Do you know why this truncation and appending would be happening in ESS?

from ess.

Conor Nash [email protected]
on Mon, 31 Dec 2012 15:39:24 -0800 wrote:

The error appears as TestCggplot. From what I can see the second line of code
(first line beginning with ggplot) is being truncated for some reason and the
next line appended immediately afterward. This line is also truncated at the
point of ggplot(uniquefull[Widgets.for.Sessi and the line terminated.

ESS is not truncating anything, that is for sure. The whole buffer text
is sent to process as it is.

Your problem is that you have some errors in the code before the one you
have reported. Particularly what is :=? There is no such thing in R.

Do you have more errors above the error you have reported? That could
happen. Because you send the whole buffer, only the first error is
relevant.

On local machine, if you activate ess-tracebug (from ESS menu) you will
get a nice arrow indicating your last input. It is very handy in such
cases. Unfortunately it is not available for ess-remote.

In such cases the best thing is to step through the code with C-c C-c or
C-RET and see where the first error occurred.

from ess.

Your problem is that you have some errors in the code before the one you
have reported. Particularly what is :=? There is no such thing in R.

I'm using the data.table package which uses the := operator (explanation here)

This is the first error reported in the buffer. The error did not occur when I ran the code line by line.

I ran the code on my local machine and I did not see the error. This would indicate that TRAMP may be at fault so I'll continue down that path of investigation. Thank you for your help!

from ess.

Conor Nash [email protected]
on Tue, 01 Jan 2013 12:19:05 -0800 wrote:

I'm using the data.table package which uses the := operator
(explanation [here])

Interesting, I didn't know that. Thanks.

This is the first error reported in the buffer. The error did not
occur when I ran the code line by line.

Hm, that makes me worried. I am quite sure ESS does nothing than
process-send-string with the buffer content. But I might be wrong.

Could you please make a small debug for me? Please go to ess-inf.el to
this line (assuming you are using the recent git head or svn trunk):

╭──────── #1141 ─ /home/vitoshka/ESS/git/lisp/ess-inf.el ──
│ (t
│ (process-send-string process (ess--concat-new-line-maybe string)))))
╰──────── #1142 ─

Just before (process-send-string ... ) insert something like (message
(ess--concat-new-line-maybe string)). And eval ess-send-string function
(C-M-x)

Then execute ess-eval-buffer and go to message buffer (C-h e). Is the
string that is sent to process is indeed your buffer content? If not
then it is a problem with ESS, otherwise it's emacs or ssh problem.

For the above check to work, please make sure you have ess-eval-visibly
set to nil. You have this in your config, but just to double check. See
the ESS>Process>Eval_visibly sub-menu.

I ran the code on my local machine and I did not see the error. This
would indicate that TRAMP may be at fault so I'll continue down that
path of investigation.

Tramp? You said that you used shell -> ssh -> ess-remote.

Thanks,
Vitalie

from ess.