Re: Bug#819343: ITP: r-cran-dplyr -- A Grammar of Data Manipulation for GNU R
On 30 March 2016 at 19:01, Andreas Tille wrote:
| On Wed, Mar 30, 2016 at 07:02:04AM -0500, Dirk Eddelbuettel wrote:
| >
| > | Could you please explain what size exactly is doubled? As far as I
| > | understood r-cran-bh would be just a wrapper for the Debian packaged
| > | libboost.
| >
| > Nope. As I explained to you before and in this thread, there is downside in
| > differing from what is on CRAN.
|
| OK. At a second look I noticed how simple it is to replace BH by the
| Debian packaged boost. I think I've got the downside but to my opinion
Yes, it is a simple sed call on DESCRIPTION, coupled with a Build-Depends.
But -- you risk creating different packages, creating different behaviour and
possibly very different to track bugs.
Having thought about this for a bit, I came to the conclusion that I'd rather
package r-cran-bh.
| its way more sensible to avoid code duplication as a general rule and
| run the accompanying test suite of the packages to ensure that
| everything works as expected.
|
| In other words I do not need the r-cran-bh package as a predependency
| of my packages any more.
Entirely your call, but as I state above, one that I would NOT make.
One of the strongest things about R is the consistent reliability across
installation. You are starting to differ here. It may not matter most of
the time, only to all of sudden become an issue -- that may be hard to track
form someone not familiar with these details.
In my view saving 5mb in the archive is not worth it.
Dirk
|
| Kind regards
|
| Andreas.
|
| --
| http://fam-tille.de
--
http://dirk.eddelbuettel.com | @eddelbuettel | edd@debian.org
Reply to: