On 28 May 2013 at 14:13, Julian Gilbey wrote: | On Wed, Apr 10, 2013 at 01:01:53PM +0900, Charles Plessy wrote: | > -- the version numbers already solve this | > | > -> At build time, it is not possible to guess what the "lower than" relation should be, | > and therefore the approach with a pseudopackage is more effective. | > | > -- this was needed three times in ten years | > | > -> I will be very happy to benefit from the R API pseudopackage in 2016 :) Time files | > really fast ! | | And I have one little piece of evidence that this will be helpful even | during transitions. I am running Debian testing and have hit a | breakage. The r-base-* packages have now transitioned to testing | (jessie has 3.0.0-2 at the time of writing), and I'm trying to compile | a local R package, only to get the following error: | | ** byte-compile and prepare package for lazy loading | Error : package ‘mgcv’ was built before R 3.0.0: please re-install it | | And sure enough: | | polya:~ $ rmadison r-cran-mgcv | r-cran-mgcv | 1.6-2-1 | squeeze | amd64, armel, i386, ia64, kfreebsd-amd64, kfreebsd-i386, mips, mipsel, powerpc, s390, sparc | r-cran-mgcv | 1.7-13-1 | wheezy | amd64, armel, armhf, i386, ia64, kfreebsd-amd64, kfreebsd-i386, mips, mipsel, powerpc, s390, s390x, sparc | r-cran-mgcv | 1.7-13-1 | jessie | amd64, armel, armhf, i386, ia64, kfreebsd-amd64, kfreebsd-i386, mips, mipsel, powerpc, s390, s390x, sparc | r-cran-mgcv | 1.7-22-1 | sid | ia64 | r-cran-mgcv | 1.7-22-2 | sid | kfreebsd-amd64, kfreebsd-i386 | r-cran-mgcv | 1.7-23-1 | sid | amd64, armel, armhf, hurd-i386, i386, mips, mipsel, powerpc, s390, s390x, sparc | | So because of the new version in sid, which didn't wait for the R | 3.0.0 migration to testing, the jessie version is still an R 2.7 | version :-( Had there been a "Provides: r-api-3.0" in r-base-core and | "Depends: r-api-2.7" in r-cran-mgcv, the transition would either have | been delayed or knowingly pushed though in spite of this, but the | breakage would have been caught at the time. (Then again, this might | have been the case anyway.)
There is no such thing as an "R 2.7" API, and as long as upstream does not define one I do not plan to jump into the frey and meddle with it. The status quo works well enough, but I will acknowledge the breakage you got with that r-cran-mgcv package. But I think we have bigger fish to fry. Dirk -- Dirk Eddelbuettel | e...@debian.org | http://dirk.eddelbuettel.com -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
