[ESS-bugs] Variable name affects indentation

[Martin Maechler]

It's clearly from  (ess-calculate-indent ... ) in  ess-mode.el
and its very old code that special cases "expression"....
Code that goes back to when Tony took started using the revision control system.
The code in question is

	       ;; modified by shiba at isac 7.3.1992
	       (cond ((and (numberp ess-expression-offset)
			   (re-search-backward "[ \t]*expression[ \t]*" bol t))
		      ;; This regexp match every "expression".
		      ;; modified by shiba
		      ;;(forward-sexp -1)
		      (beginning-of-line)
		      (skip-chars-forward " \t")
		      ;; End
		      (+ (current-column) ess-expression-offset))

and it is clearly why it leads to the wrong indentation here.

The corresponding indentation variable is documented in ess-utils.el :

(defvar ess-expression-offset 4
  "Extra indent for internal substatements of `expression' that specified
in `obj <- expression(...)' form.
If not number, the statements are indented at open-parenthesis following
`expression'.")

and that I think explains the intent, i.e., what 'shiba' *did* solve
in March 1992,
 by replacing the  (forward-sexp -1)
by the the two   (beginning-of-line)   (skip-chars-forward " \t")

Tony, would you remember  more ?

It seems it's "easy" to solve if
"[ \t]*expression[ \t]*"    is replaced by a smarter one.

I'll try a bit.
Martin



On Sat, Aug 18, 2012 at 10:51 AM, Martin Maechler
<maechler at stat.math.ethz.ch> wrote:
> Dear Hendrik,
>
> On Wed, Aug 15, 2012 at 3:15 AM, Hendrik Weisser
> <weisser at imsb.biol.ethz.ch> wrote:
>> Hi!
>>
>> Working with protein expression data, I've noticed that using the variable
>> name "expression" breaks the indentation in my R code:
>>
>> What I expect to see:
>> expression <- somefunction(longarg1,
>>                            longarg2)
>>
>> What happens:
>> expression <- somefunction(longarg1,
>>     longarg2)
>>
>> I.e. the indentation at the level of the opening brackets after a line break
>> doesn't work as usual (same for square brackets).
>
> I can confirm.   This is a buglet.
>
>> I assume this has
>> something to do with "expression" being a pre-defined function in R.
>
> Well, not really directly:  if I use  "data.frame" or "list" instead
> of "expression",
> the indentation is correct.
>
>>
>> It's just a little annoyance, but maybe it could be fixed nonetheless.
>
> Yes, I hope we'll find a cure before the next release of ESS.
> Martin
>
>>
>> Best regards
>>
>> Hendrik Weisser
>>
>> _______________________________________________
>> ESS-bugs ESS-bugs at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/ess-bugs
>>
>> _______________________________________________
>> ESS-core list: https://stat.ethz.ch/mailman/listinfo/ess-core
>>