On Tue, Jul 22, 2014 at 7:10 PM, Andrzej OleÅ <andrzej.o...@gmail.com> wrote:
> Hi Dan, Michael, Julian, > > Thank's for keeping the links to the tarballs! > > I don't argue that mixing release and devel is a good idea in general. > Rather, that for some users this might be the best compromise between > the following two objectives: > 1. a stable working environment > 2. the possibility to use or just quickly check a specific new feature > available in the devel version of package X > > IMHO this is the road to ruin. > Switching entirely to devel is quite often a no-no for them because of > the unstable nature of the devel branch. And maintaining both release > I personally have never done the homework required to have both branches on hand with convenient updating. It is surely feasible and I would imagine a number of folks reading this have done it for themselves. It is probably not trivial to do it portably but a doc with suggestions for key platforms would be a nice contribution. I used to have Rrel and Rdev scripts that made it work and that is easily worked out but it is not very elegant and transitioning to new releases of R is laborious. > and devel only adds to their frustration. As a developer I would like > to have the freedom to advise people on using the latest devel version > of my package regardless of whether they are running release or devel > if I think that this is safe for them, which is typically the case for > many upstream packages without (many) reverse dependencies. I don't > see the point of unnecessary obstructing this approach and I'm not > sure I understand why there is such an outrage about mixing release > and devel. In contrary, quite often this can be much safer than > Depends on your definition of safety. I think that we have gained much from the clean separation, in terms of user support effort and freedom to experiment in devel. "Experts" can do what they like and deal with the consequences themselves, but the general approach to the user community should be principled and it should be: 1) in general, use release branch and report bugs and see that they are fixed in a timely way; if they are not, let the core know -- 2) if you expect help, do all package installations via biocLite() -- 3) we like seeing mileage on the devel branch and encourage its use for novel features, but it needs to be used with the appropriate version of R and the devel package set. > switching between BioC branches. I personally do not want to find > myself in a position when I advice a user to switch to BioC devel > because of some new function from my package he/she would like to give > This is an engineering commitment to a stable release branch. No new features, only bug fixes. We have benefited from this immensely. > a try only to learn that this broke his/hers scripts (due to the > changes in some other packages). > > To sum up, I believe that mixing release and devel might be beneficial > in some specific cases similar to the above described one and it's > important that the infrastructure allows leveraging this approach, > e.g. by providing direct access to devel tarballs. > > Cheers, > Andrzej > > > On Tue, Jul 22, 2014 at 11:12 PM, Julian Gehring <julian.gehr...@embl.de> > wrote: > > Hi Andzrej, > > > > thank you, I see your point but I'm afraid I must disagree with you. > > I've had this situation numerous times that I have added/fixed > > something in the devel branch of a package and had to advice the users > > to use this latest version. Needless to say, they were typically using > > the release branch, and it was a relatively painless procedure for > > them to pick the tarball from the devel landing page and proceed with > > manual installation. Of course, this could be also achieved by > > installing from the svn, however, this is not very welcome from the > > user's perspective. > > > > > > I'm not sure I understand to need to mix devel and release. If there is > a > > bug in the release branch, it should be also patched there. And if users > > need the features of the devel branch, they would have to switch to devel > > anyway. > > > > > > Best > > Julian > > _______________________________________________ > Bioc-devel@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/bioc-devel > [[alternative HTML version deleted]]
_______________________________________________ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel
