[Rd] Segmentation fault when options(max.print = .Machine$integer.max)

Martin Maechler maechler at stat.math.ethz.ch
Tue Sep 25 17:29:59 CEST 2012


>>>>> Martin Morgan <mtmorgan at fhcrc.org>
>>>>>     on Tue, 25 Sep 2012 05:34:12 -0700 writes:

    > On 09/25/2012 05:26 AM, Martin Maechler wrote:
    >> 
    >>> Seemed like a good idea at the time,
    >> 
    >> I'm curious.  Why is it (setting max.print much too
    >> large) a good idea?

    > I usually set it considerably smaller (50) than default to
    > conserve screen real estate, but then occasionally need to
    > see more than my small setting (e.g.,
    > ls(getNamespace("Matrix"))) 

:-)  :-)   that's a good one...

but even for that and for the longer  
ls(getNamespace("Matrix")), all=TRUE) 

  options(max.print = 1000)

is sufficient.

    > and don't want to guess at how much I want to see.

I understand.   But really the reason we had introduced it,
*was* exactly equivalent to saying that setting it to
"practically Inf" is unreseasonable.
If you want it really large, use a million which is already more
than you want, and '1e6' is really faster typing than .Machine$..


    > Thanks for your fix.

de nada,
Martin



    > Martin

    >> 
    >>> but > options(max.print = .Machine$integer.max) > 1:10
    >>> [1] Program received signal SIGSEGV, Segmentation fault.
    >> 
    >>> because of an integer overflow at
    >>> src/main/printvector.c:176
    >> 
    >>> > sessionInfo() R Under development (unstable)
    >>> (2012-09-24 r60800) Platform: x86_64-unknown-linux-gnu
    >>> (64-bit) ...
    >> 
    >>> also R-patched, etc.
    >> 
    >> Thank you, Martin.  I'm about to commit fixes for this.
    >> "another" Martin.
    >> 
    >> 
    >>> Martin
    >>> --
    >>> Computational Biology / Fred Hutchinson Cancer Research
    >>> Center 1100 Fairview Ave. N.  PO Box 19024 Seattle, WA
    >>> 98109


    > -- 
    > Computational Biology / Fred Hutchinson Cancer Research
    > Center 1100 Fairview Ave. N.  PO Box 19024 Seattle, WA
    > 98109

    > Location: Arnold Building M1 B861 Phone: (206) 667-2793



More information about the R-devel mailing list