[ESS] Problems after Debian update

[Vitalie Spinu]

I got a copy of stata, so got playing with ess-sta-d.el.

There is one ugly, historical thing over there -
inferior-STA-program-name is set to "env" and the final call is
hardcoded as "env TERM=emacs stata". It looks like this workaround is no
longer necessary, comint takes care of setting
TERM=dumb. inferior-ess-make-comint is also setting STATATERM=emacs. 

Is there any difference in setting TERM=emacs vs TERM=dumb for stata?

In my test it doesn't matter at all. So, if there are no objections I am
changing inferior-STA-program-name to "stata" with all ancillary
corrections.


>>>> Vitalie Spinu <spinuvit at gmail.com>
>>>> on Sat, 05 May 2012 00:45:26 +0200 wrote:

  > Indeed, the right thing would be to set the secondary prompt to nil as
  > it's meaningless. As I don't have stata, your help is really, really
  > appreciated here. Please try

  > (setq STA-customize-alist
  >   '((ess-local-customize-alist     . 'STA-customize-alist)
  >     (ess-language                  . "STA")
  >     (ess-dialect                   . STA-dialect-name)
  >     (ess-suffix                    . "ado")
  >     (ess-mode-editing-alist        . STA-editing-alist)
  >     (ess-mode-syntax-table         . STA-syntax-table)
  >     (ess-mode-edit                 . 'STA-mode)
  >     (ess-help-sec-regex            . ess-help-STA-sec-regex)
  >     (ess-help-sec-keys-alist       . ess-help-STA-sec-keys-alist)
  >     (ess-loop-timeout              . 500000 )
  >     (ess-object-name-db-file       . "ess-sta-namedb.el" )
  >     (inferior-ess-font-lock-keywords . ess-STA-mode-font-lock-keywords)
  >     (inferior-ess-program          . inferior-STA-program-name)
  >     (inferior-ess-objects-command  . "description\n")
  >     (inferior-ess-help-command     . "set more off\n help %s\n")
  >     (inferior-ess-exit-command     . "exit\n")
  >     (inferior-ess-primary-prompt   . ". ")
  >     (inferior-ess-secondary-prompt . nil) 
  >     (comint-use-prompt-regexp      . t)
  >     (inferior-ess-start-file       . nil) 
  >     (inferior-ess-start-args       . "")) 
  > )

  > (defun inferior-ess--goto-input-start:regexp ()
  >   "Move point to the begining of input skiping all continuation lines.
  > If in the output field, goes to the begining of previous input.
  > "
  >   (beginning-of-line)
  >   (unless (looking-at inferior-ess-prompt)
  >     (re-search-backward (concat "^" inferior-ess-prompt)))
  >   (when inferior-ess-secondary-prompt
  >     (while (and (looking-at inferior-ess-secondary-prompt)
  >                 (not (eq (point) (point-min))))
  >       (forward-line -1)))
  >   (if (looking-at inferior-ess-prompt)
  >       (comint-skip-prompt)
  >     (ess-error "Beggining of input not found"))
  >   )

  > I hope it solves both problems.

  > Vitalie.

>>>> brendan.halpin at ul.ie (Brendan Halpin)
>>>> on Fri, 04 May 2012 23:16:21 +0100 wrote:

  >> Thanks, that works nicely for the first problem: I can now start Stata. 
  >> However, the previous-input problem remains. It seems to be in this
  >> function from ess-inf.el:

  >> (defun inferior-ess--goto-input-start:regexp ()
  >> "Move point to the begining of input skiping all continuation lines.
  >> If in the output field, goes to the begining of previous input.
  >> "
  >> (beginning-of-line)
  >> (unless (looking-at inferior-ess-prompt)
  >> (re-search-backward (concat "^" inferior-ess-prompt)))
  >> (while (and (looking-at inferior-ess-secondary-prompt)
  >> (not (eq (point) (point-min))))
  >> (forward-line -1))
  >> (if (looking-at inferior-ess-prompt)
  >> (comint-skip-prompt)
  >> (ess-error "Beggining of input not found"))
  >> )

  >> This seems to be looking for the secondary prompt, to backtrack from it
  >> to the initial prompt. But it is defined the same as the primary prompt,
  >> so if we're looking-at the primary prompt, (looking-at
  >> inferior-ess-secondary-prompt) returns t. 

  >> In interactive mode, Stata only accepts single lines, so there isn't a
  >> secondary prompt. Perhaps this should be defined to nil?

  >> Regards,

  >> Brendan