ess bets mismanages help output?

[Martin Maechler]

>>>>> "Rodney" == Rodney Sparapani <rsparapa at post.its.mcw.edu>
>>>>>     on Fri, 19 Mar 2004 08:50:29 -0600 (CST) writes:

    >>> options(htmlhelp = TRUE)
    >>> 
    >>> or did this stop working ?
    >>> (for me, help.start() doesn't even work currently with a running
    >>> Mozilla, you see how much I use the 'only true' interface ..).
    >> 
    >> Martin, from Paul's description, here is what I thought happened:
    >> 
    >> If you do help.start(), then subsequent ?lm  calls appeared in the
    >> browser.  
    >> 
    >> I know that ESS grabs the ?lm and sends it to R, redirecting the
    >> output. 
    >> 
    >> But you are suggesting that R will "grab the help and shove it into
    >> the browser" if that option is set?  (I don't have a browser with Java
    >> currently running on this laptop, can't test).
    >> 
    >> best,
    >> -tony

    Rodney> Hmmm.  options(htmlhelp = TRUE) does nothing for me.
    Rodney> I still get ?lm in an emacs buffer.  
Did you *delete* the already existing  *help[R](lm)*  buffer
first ?

Existing help buffers are 'recycled' by default.
IIRC, you can also C-u C-c C-v which does "forces" not re-using
an existing help buffer.

    Rodney> I still get ?lm in an emacs buffer.  Of course, I'm
    Rodney> using help.start(), but Martin seems to be using
    Rodney> something else; what is that? 

did I allude to anything else? 
No, I don't. But I know about the "reuse existing help buffer"
magic...

    Rodney> By the way, is anyone else bothered by the sporadic
    Rodney> highlighting of the 'for's, 'in's, 'switch's,
    Rodney> etc. of their output.  Shouldn't just the input be
    Rodney> highlighted, i.e. lines that start with '>' or '+'?

yes, you are right (but it might be a bit of work to achieve
correctly in all cases).

Regards,
Martin