R devel - Oct 2000 - This mail was found after problem with mailserver Oct 3rd - postmaster@hh.umu.se

>>>>> "MM" == Martin Maechler
<maechler@stat.math.ethz.ch> writes:
>>>>> "david" == david 
<david@orion-10.cs.byu.edu> writes:
    david> When I issue the command: if ( c(2.8,3)[1] < pretty
    david> (c(2.8,3))[1]) print ("problem")

    david> the result is "problem" prints.  I have traced it to
bits and
    david> the 2.8 of the returned vector from pretty is slightly larger.
    david> This seems to be due to the conversion to base 10 with the ns
    david> and unit variables in src/appl/pretty.c.

    david> Anyway if I do something like: hist(c(2.8,3)) it fails as well
    david> because it uses pretty.

    david> Thanks for you attention, DavidMcLaughlin@byu.edu

    MM> Thank you for the bug report.  This bug is ``new'' (a bit
more than
    MM> a year!) and was introduced when someone fixed another problem in
    MM> pretty.c by adding "1e-5" fuzz a bit too generously
hmm,.......
    MM> I'll have a look.

    david> Version: .... R 1.1.1 ...

Ok, I've fixed pretty() for the development version
such that
     p <- pretty(x)
     all(p[1] <= x & x <= p[2])

should be guaranteed (again).

Note however, that this really also changes some "results",
that may be unexpected,
e.g.

    R 1.1.1 :
    ---------

    > data(women); (hT <- hist(women$weight, nc= 7, plot = FALSE))[2]

    $counts
     [1] 3 1 2 2 1 1 2 1 1 1

    --------------

    R-devel [to be 1.2.0]:
    ---------------------

    > data(women); (hT <- hist(women$weight, nc= 7, plot = FALSE))[2]

    $counts
     [1] 2 2 2 1 2 1 2 1 1 1

    > 

------

Martin Maechler <maechler@stat.math.ethz.ch>
http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum  LEO D10	Leonhardstr. 27
ETH (Federal Inst. Technology)	8092 Zurich	SWITZERLAND
phone: x-41-1-632-3408		fax: ...-1228			<><

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

maechler@stat.math.ethz.ch writes:
> Ok, I've fixed pretty() for the development version
> such that
>      p <- pretty(x)
>      all(p[1] <= x & x <= p[2])
> 
> should be guaranteed (again).
Gulp. I'm pretty(!) sure there was a reason for changing it in the
other direction... [I have a vague recollection of pretty(pretty(x))
!= pretty(x) or something like that] Are we quite sure that it is not
the caller that needs to be a little more lenient in the bounds
checking?

OTOH, it's in code you wrote yourself...

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk)             FAX: (+45) 35327907
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._