[Rd] Depends: R(>= ?)

Peter Dalgaard BSA p.dalgaard@biostat.ku.dk
12 Jun 2000 19:41:33 +0200

Paul Gilbert <pgilbert@bank-banque-canada.ca> writes:

> I am just working through my set of tests for R 1.1.0 and everything
> seems to be going fairly well other than a couple of bugs already
> reported and fixed.
> I previously specified Depends: R(>= 1.0.1). I have been making small
> changes to my own code and since I have been doing this with R 1.0.1  I
> am fairly sure everything still works with it, but I don't intend to
> test this as carefully as R 1.1.0. Should I change "Depends" to indicate
> 1.1.0?
> Now I realize what the "safest" answer would be, but I can see this will
> be an ongoing problem and the implication of changing each time is that
> my packages always depend on the latest version of R. I can see the
> posssibilty of testing 1.0.1 and 1.1.0, but when R 1.99 arrives I don't
> see any possibility for testing every version since R 1.0.1. Should
> "Depends" indicate what is "known to be at least necessary" or should it
> be something stronger than that?

Hmm. We try pretty hard not to change the API and otherwise break
package code after 1.0. I think it should be reasonable if you took a
laid-back attitude and assume that things work as they used to unless
you knowingly access new features. If things turn out to be broken
with earlier versions, then blame the core team and update the version
number. If you do use new features and try to blame us, we'll blame
you back...

It would of course be nice if you kept a version of 1.0.1 around and
checked against it, but checking all intermediate versions would be
overdoing it.

   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
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch