On 8 August 2015 at 09:03, Dirk Eddelbuettel wrote: | | On 8 August 2015 at 14:34, Jonathan Wiltshire wrote: | | On 8 August 2015 at 13:08, Julien Cristau wrote: | | > However, instead of being properly split out as required by | | > policy, the libraries are shipped as part of the 'jags' binary package, | | > which makes this a bit of a pain. | | | | On Sat, Aug 08, 2015 at 06:50:44AM -0500, Dirk Eddelbuettel wrote: | | > I don't see the need for splitting the libraries out of the jags | | > package. What am I missing? | | | | Making it not be "a bit of a pain" in the future. | | I hear you "in theory", but "in practice" I know the upstream author very | well, and have a reasonable idea of how this is used (in Bayesian stats | around R in the large sense). | | This isn't really a library to link against -- all statistical analysis uses | jags the binary. | | But I look into it. The debian/rules is (apart from some overrides) just | | %: | dh $@ | | so I should get by with some simple-enough debian/*.files magic.
Abort. I just wasted a few hours on this (luckily while doing other stuff). I am not going to split libjags, libjags-dev off but simply rebuild depending on g++ (>= 4:5.2). Chris can then depend on this version, and that is all that there is to the jags part of the g++5 transition. Dirk -- http://dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org