[Rd] strange behaviour in R 2.0.0 (PR#7275)

Fan xiao.gang.fan1 at libertysurf.fr
Sun Oct 10 19:45:40 CEST 2004


Dear Prof Ripley,

After looking at my original C code, which is different in 2.0.0
should be (sorry, not tested as I've desinstalled 2.0.0,
tired of the time wasted):
 > a=list(x=1:2)
 > a$y = strptime(c("20010101","20020101"),"%Y%m%d")
 > a = data.frame(a)
 > rbind(a,a)

For the point 1 in PR#7269 (failure in installing packages from source),
I've encountered that for several packages (so "toto" means nothing).
I guess that should come from changes somewhere related to the "index"
check in the perl script for making process. I agree with you it'd be
difficult to claim that a bug, but anyway, it's not expected as
I've checkec the different files (man, INDEX, DESCRIPTION) and can't
figure out what's wrong, and it WORKS with 1.9.1.

ps: sorry for having posted 2 times (PR#7269, PR#7272), I've used
the Web interface of R-Bugs to send the report, and after posting
the first one, I've figured out that there was an error in the email
address (xiao.gang.fan at libertysurf.fr instead of xiao...1 at libertysurf.fr).
--
Fan

Prof Brian Ripley wrote:
> This is the fourth bug report today from Xiao Gang FAN, and we have had no
> response to replies to the first three.  Please do not send bug reports 
> unless you are prepared to take responsibility for them and answer 
> requests for further information.
> 
> Please read the posting guide and the FAQ and use a meaningful subject 
> line.
> 
> Now, where is the requested code to reproduce PR#7272?
> 
> 
> On Sun, 10 Oct 2004 xiao.gang.fan1 at libertysurf.fr wrote:
> 
> 
>>Full_Name: Xiao Gang FAN
>>Version: R 2.0.0
>>OS: Windows
>>Submission from: (NULL) (82.226.247.142)
>>
>>
>>The following codes worked for R 1.9.1
>>
>>
>>>a = data.frame(x=1:2)
>>>a$y = strptime(c("20010101","20020101"),"%Y%m%d")
>>
>>this no longer works for R 2.0.0:
>>
>>Error in "$<-.data.frame"(`*tmp*`, "y", value = list(sec = c(0, 0), min = c(0, 
>>: 
>>        replacement has 9 rows, data has 2
> 
> 
> That code is indeed incorrect, and did fail in 1.9.1:
> 
> R : Copyright 2004, The R Foundation for Statistical Computing
> Version 1.9.1  (2004-06-21), ISBN 3-900051-00-3
> 
> 
>>a = data.frame(x=1:2)
>>a$y = strptime(c("20010101","20020101"),"%Y%m%d")
> 
> Error in "$<-.data.frame"(`*tmp*`, "y", value = strptime(c("20010101",  :
>         replacement has 9 rows, data has 2
> 
> It even failed in 1.8.0.  It `worked' in 1.7.0, but did not do what you
> most likely wanted it to do (see the NEWS for 1.8.0).  You cannot sensibly
> add "POSIXlt" objects to a data frame: you can add "POSIXct" objects.
>



More information about the R-devel mailing list