[ESS] Command completion (redux)

[Stephen Eglen]

James W. MacDonald <jmacdon at med.umich.edu> wrote:

> Hi Stephen,
> 
> Here is my backtrace:
> 
> Debugger entered--Lisp error: (wrong-type-argument stringp nil)
>   all-completions(nil nil)
>   comint-dynamic-simple-complete(nil nil)
>   (or (comint-dynamic-simple-complete token-string
> possible-completions) (quote none))
>   (let* ((comint-completion-addsuffix nil) (beg-of-line ...)
> (end-of-line ...) (line-buffer ...) (token-string ...)
> (possible-completions ...)) (or (comint-dynamic-simple-complete
> token-string possible-completions) (quote none)))
>   alt-ess-complete-object-name()
>   call-interactively(alt-ess-complete-object-name)
> 

Thanks Jim; the backtrace is yet again helpful... that bottom line tells
me that you called alt-ess-complete-object-name -- is that still in your
.emacs somewhere?

Since 5.3.7, Martin put this completion code into the official ESS,
as:

;; From Jim (James W.) MacDonald, based on code by Deepayan Sarkar,
;; originally named  'alt-ess-complete-object-name'.
;; Use rcompgen in ESS
;; Can be activated by something like
;; (define-key inferior-ess-mode-map "\t" 'ess-R-complete-object-name)
(defun ess-R-complete-object-name ()  [code deleted]  )

and that completion code currently seems to be on the "C-c TAB" key,
rather than the TAB key in *R* buffers.  (Martin: is there a reason it
should be on C-c TAB, rather than TAB?).  Also, see that the value of
ess-use-R-completion is non-nil.

Jim: tell us what "C-h k TAB" and "C-h k C-c TAB" report in an *R*
buffer.

Stephen