R devel - Nov 2004 - Rcmd check for Windows fails (PR#7369)

Full_Name: Harald Weedon-Fekj?r
Version: 2.0
OS: Windows
Submission from: (NULL) (80.202.87.143)


Dear R core

Thanks again for a superb program. This is really one of the rare instances I
find a bug. The bug is easy to bypass, but can be quite confusing;

"Rcmd check" seems to fail without any good error report if the
package contains
a certain amount of R code and is run from a directory contain a space in the
name (strange, but ...very annoying as I tried to include/exclude new parts many
times).

For error report and the source that fail, see:
http://folk.uio.no/hfe/pub/slettes-dato/2005-12/addreg-check-log.txt
http://folk.uio.no/hfe/pub/slettes-dato/2005-12/addreg.zip

Feedback would be welcommed.

PS: R is really a wounder full program. S-PLUS used to be "best command
line
statistical package", no it is just "best money can buy". Use R
in nearly all my
work!

Best
Harald Weedon-Fekjaer, Oslo, Norway

harald.weedon-fekjar@medisin.uio.no writes:
> "Rcmd check" seems to fail without any good error report if the
package contains
> a certain amount of R code and is run from a directory contain a space in
the
> name (strange, but ...very annoying as I tried to include/exclude new parts
many
> times).
> 
> For error report and the source that fail, see:
> http://folk.uio.no/hfe/pub/slettes-dato/2005-12/addreg-check-log.txt
> http://folk.uio.no/hfe/pub/slettes-dato/2005-12/addreg.zip
> 
> Feedback would be welcommed.
If you're quick you could check and tell us whether this is fixed in
2.0.1, see

http://cran.r-project.org/bin/windows/base/rpatched.html

In particular, this stuff from the CHANGES file could be relevant:

"We have increased the chances that installing a source package from
an explicit path with spaces in will work, so e.g.

    R CMD INSTALL "/Documents and Settings/myname/Source Files/mypkg"

will probably work -- it is still not recommended.  You can also
get away with backslashes here, but forward slashes are safer."


-- 
   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

This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

--27464147-1359797818-1100466430=:26232
Content-Type: TEXT/PLAIN; charset=iso-8859-1; format=flowed
Content-Transfer-Encoding: QUOTED-PRINTABLE

On Sun, 14 Nov 2004 harald.weedon-fekjar@medisin.uio.no wrote:
> Full_Name: Harald Weedon-Fekj=E6r
> Version: 2.0
> OS: Windows
> Submission from: (NULL) (80.202.87.143)
>
>
> Dear R core
>
> Thanks again for a superb program. This is really one of the rare instances
I
> find a bug. The bug is easy to bypass, but can be quite confusing;
>
> "Rcmd check" seems to fail without any good error report if the
package=20
> contains a certain amount of R code and is run from a directory contain=20
> a space in the name (strange, but ...very annoying as I tried to=20
> include/exclude new parts many times).
Please *do* read README.packages, which does tell you that is not=20
supported.

It does in 2.0.1 beta, BTW, due for release tomorrow.  The CHANGES file=20
says

We have increased the chances that installing a source package from
an explicit path with spaces in will work, so e.g.

     R CMD INSTALL "/Documents and Settings/myname/Source Files/mypkg"

will probably work -- it is still not recommended.  You can also
get away with backslashes here, but forward slashes are safer.

--=20
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595
--27464147-1359797818-1100466430=:26232--