On 4 October 2005 at 08:23, Matt Kraai wrote:
| On Tue, Oct 04, 2005 at 06:17:27AM -0500, Dirk Eddelbuettel wrote:
| >
| > On 3 October 2005 at 21:59, Matt Kraai wrote:
| > | > | > I disabled a few of these tests on other packages formerly provided
by
| > | > | > r-recommended, and I guess I need to disable it here too.
| > | > |
| > | > | This was on i386.
| > | >
| > | > Odd, so how does that square with my pbuilder logs showing that it
built just
| > | > fine?
| > |
| > | I don't know. I'll make sure that I can reproduce the problem
| > | tomorrow and, if I can, send you the list of packages that were used
| > | in the build environment.
| >
| > After sending the mail I realized that it may just be the difference between
| > the previous R [ my logs show I used R 2.1.1 ] whereas you will have gotten
a
| > R 2.2.0.beta.* for the R 2.2.0 release due this week. If I get a chance I'll
| > try boot as well here.
|
| I was able to reproduce the failure. It used version
| 2.2.0.beta.20050929-1 of both r-base-core and r-base-dev.
Thanks for doing that. I wonder, though, how it then survives the three tests
(against R stable, patched and devel -- roughly equivalent to our stable,
testing and unstable) on the CRAN tests, but looking at lines 167 and 168 in
http://cran.r-project.org/src/contrib/checkSummary.html
reveals that CRAN uses boot 1.2.23 against R-devel and 1.2.22 against
R-released (aka 2.1.1) and R-patched (I presume that's the release
candidate).
So it looks like I am guilty by having inserted boot 1.2.23 "too soon". Oh
well.
Dirk
--
Statistics: The (futile) attempt to offer certainty about uncertainty.
-- Roger Koenker, 'Dictionary of Received Ideas of Statistics'
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
