[ESS] RE: Filename completion becomes object completion

Martin Maechler maechler at stat.math.ethz.ch
Wed May 4 14:57:35 CEST 2005


>>>>> "tony" == A J Rossini <blindglobe at gmail.com>
>>>>>     on Mon, 2 May 2005 21:31:42 +0200 writes:

    tony> Not sure what is happening -- C-TAB is the correct
    tony> object completion, and TAB is a funky "tries to do the
    tony> right completion".

yes, and it doesn't work anymore, at least under certain
circumstances.  I can confirm this, and I agree that this was
not the intention.
We should find how exactly the
   `` funky "tries to do the right completion" ''
works and where it fails to see that it should do file name
rather than object name completion.

At the moment I have no time to delve into this problem though....

    tony> On 5/2/05, Brahm, David <David.Brahm at geodecapital.com>
    tony> wrote:
    >>
    >> I should have mentioned I am using Emacs 21.3.1 on RedHat
    >> Linux.

yes, thank you, David, for your persistent (IIRC) report of this
problem.


    >> > -----Original Message----- 
    >> From: Brahm, David 
    >> Sent: Monday, May 02, 2005 11:46 AM 
    >> To: ESS-help at stat.math.ethz.ch 
    >> Subject: Filename completion becomes object completion

    >> We just upgraded to R-2.1.0 and ESS-5.2.7.  In an *R*
    >> buffer, inside quotes, we used to get filename
    >> completion with TAB, but now ESS is attempting 
    >> object completion instead.  E.g. if I type 
    >> 
    >>  source("/tmp/R 
    >>      and hit TAB, it tells me about R.Version, etc. instead of
    >> directories in /tmp.

{yes, I can confirm the current behavior, see above}

    >> > ESC TAB correctly does filename completion.

aka M-Tab  {and that's what I have been using all the time and
	    hence never noticed the TAB -> smart-completion problem}

    >> > I've seen some discussion on ESS-help on this, but I
    >> didn't find a resolution.  Is there a way to recover
    >> the old (correct) behavior?  Thanks.
    >> >
    >> > -- David Brahm (brahm at alum.mit.edu)

Martin Maechler, ETH Zurich




More information about the ESS-help mailing list