[ESS] Yet another indentation question: indentation after parenthese

[Mathieu Basille]

Dear Vitalie,

This is brilliant! I just tested it on simple or more complex lines of
code, and it works perfectly and smoothly! I don't know yet what are the
limits or drawbacks of this function, but I really think I will use it a
lot on functions I write! And it works beautifully in both R and Rnw files,
as long as a R process is associated with the buffer. By the way, on my
system, 'indent-region' is bound to 'C-M-\' (not 'C-M-/').

One last questions, would it be possible to have the function actually
launch R if no process is available?

Thank you a lot! It would have taken me years to obtain something vaguely
similar!

Best,
Mathieu.


Le 17/12/2011 07:35, Vitalie Spinu a écrit :
> Mathieu Basille <basille at ase-research.org> writes:
> 
>> Le 16/12/2011 14:54, Rodney Sparapani a écrit :
>>> Mathieu Basille wrote:
>>>> Alright, I think I start to understand the problem, and what seemed easy to
>>>> me might not be possible. I thought that taking parentheses into account
>>>> would be easier than ignoring them, thus my question!
>>>>
>>>> I will stick to 'C++' or 'RRR' styles and manually adjust them when needed
>>>> (using e.g. 'source' with 'options(keep.source = FALSE)'). This makes,
>>>> however, the use of 'auto-fill-mode' virtually impossible to me, because
>>>> such statements:
>>>>
>>>> toto <- someFunction(arg1 = bli, arg2 = bla, arg3 = blo, arg2 = c("a",
>>>> "bla", c(1, 2, 3, 4), "blo", "blu", "bli"))
>>>>
>>>> will render awfully with a column width of 70. I can live with this (well,
>>>> I did for many years now!), I just hoped that there was a way to follow
>>>> what R uses internally to format R code. I find it a pity that such a great
>>>> editor like Emacs (which is, by far, the best tool I tried to edit R/Rnw
>>>> scripts, with the result that I couldn't work without it now) can't
>>>> reproduce R behaviour.
>>>>
>>>> Vitalie and Richard, many thanks go to you for your efforts to 1)
>>>> understand (!) and 2) trying to find a solution to my little problems!
>>>>
>>>> All the best,
>>>> Mathieu.
>>>
>>> Hi Mathieu:
>>>
>>> Making it into a style might be a challenge.  Mode grammar/syntax is not
>>> much fun to play with.  Look at the definition of parentheses to see if
>>> they are a word or punctuation.  Otherwise, you could probably create an
>>> elisp function that calls R to replace a function with its re-formmatted
>>> equivalent.  ESS has several functions that work like that.  For example,
>>> take a look at ess-ebcdic-to-ascii-search-and-replace which replaces EBCDIC
>>> passages in the buffer with their ASCII equivalents. You might also be able
>>> to use ESS primitives that find the function definitions and repeatedly
>>> replace them (just like you can search and replace multiple EBCDIC blurbs
>>> in one go).
> 
> Taking Rodney's idea further, you can actually call deparse(parse()) on
> the region:
> 
> (defun ess-indent-region-as-R-function (beg end)
>   (let ((string (replace-regexp-in-string
> 		 "\"" "\\\\\\&"
> 		 (replace-regexp-in-string ;; how to avoid this double matching?
> 		  "\\\\\"" "\\\\\\&" (buffer-substring-no-properties beg end))))
> 	(buf (get-buffer-create "*ess-command-output*")))
>     (ess-command (format "local({oo<-options(keep.source=FALSE);
> cat('\n',paste(deparse(parse(text=\"%s\")[[1L]]),collapse='\n'),'\n',sep='')
> options(oo)})\n"  string) buf)
>     (with-current-buffer buf
>       (goto-char (point-max))
>       ;; (skip-chars-backward "\n")
>       (let ((end (point)))
> 	(goto-char (point-min))
> 	(goto-char (1+ (point-at-eol)))
> 	(setq string (buffer-substring-no-properties (point) end))
> 	))
>     (delete-region beg end)
>     (insert string)
>     ))
> 
> 
> (add-hook 'ess-mode-hook '(lambda () (set (make-local-variable 'indent-region-function)
> 					  'ess-indent-region-as-R-function)))
>      
> Now you can select a region and pres C-M-/ to get it indented as R
> thinks it should. Pay attention that the code in the region should be
> syntactically correct.
> 
> 
> 
> Now a question for programmers out there. In the above code I have
> parse(text="%s") where "%s" is a string from the buffer. In order that
> to work one have at least to substitute all " and \" in %s (there might
> be other problems with escaped chars which I don't see now).
> 
> I assume this is a common task but I cannot find a function in emacs
> which would escape all problematic chars for me. Any ideas?
> 
> Vitalie.
> 
> 
> 
>> Dear Rodney,
>>
>> That looks interesting! I have no knowledge of elisp and probably not much
>> time to dig into it, but I'll keep this on my todo list, in case I find
>> time for it! An interactive function could actually be quite useful...
>>
>> Thanks!
>> Mathieu.
>>
>>>
>>> ______________________________________________
>>> ESS-help at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/ess-help

-- 

~$ whoami
Mathieu Basille, Post-Doc

~$ locate
Laboratoire d'Écologie Comportementale et de Conservation de la Faune
+ Centre d'Étude de la Forêt
Département de Biologie
Université Laval, Québec

~$ info
http://ase-research.org/basille

~$ fortune
``If you can't win by reason, go for volume.''
Calvin, by Bill Watterson.