On Wed, Jul 25, 2012 at 21:50:12 -0500, Dirk Eddelbuettel wrote: > > On 25 July 2012 at 21:27, Dirk Eddelbuettel wrote: > | > | close 680903 > | thanks > | > | On 26 July 2012 at 03:12, Andreas Beckmann wrote: > | | found 680903 1.0.3-22 > | | thanks > | | > | | On 2012-07-25 17:46, Dirk Eddelbuettel wrote: > | | > * debian/control: Set (Build-)Depends: to current R version > | | > | | I disagree that this is fixed properly. Assuming this will break again > | | for R 2.15.2, there should be a > | | > | | Depends: r-base-core (>= 2.15.1), r-base-core (<< 2.15.2) > | > | Not so. R never breaks at minor revisions. > | > | You have argued for adding it at major revisions, I argue against. I am the > | maintainer and have stood behind this for many years. I have no interest to > | play games here. > | > | It "simply" needs to be rebuild when new R comes out -- this is due to a bad > | design choice by upstream which should change as rpy is not maintained > | anymore. > > There was a "not" missing here -- no change or improvement anticipated as rpy > has not been maintained for years (but it being kept around as rpy2 is a > little different). > > Now: having R itself (which is used by (tens or hundreds) of thousands of > people) blocked just because rpy (which at best a few dozen people use) may > be hang up is just wrong. > > Hence my opposition to something like the above which would effectively stop > R migration. That is _not_ a service to our users. > Then let's remove rpy from wheezy. As far as I'm concerned this bug is still unfixed and RC for wheezy.
Cheers, Julien
signature.asc
Description: Digital signature
