Subject: Re: [Boost-mpi] [EXTERNAL] master vs develop
From: Matthias Troyer (troyer_at_[hidden])
Date: 2014-11-19 12:54:24
Sorry for the late reply, I've been traveling to France and California over the past two days. Did Noel's reply answer your question? I generally tested it on one or two local machines before committing to the release branch back when we used svn.
> On 18 Nov 2014, at 21:11, Belcourt, Kenneth <kbelco_at_[hidden]> wrote:
> Hi Alain,
> On Nov 18, 2014, at 3:48 AM, Alain Miniussi <alain.miniussi_at_oca.eu> wrote:
>> I have a question regarding:
>> I guess that all the commit, specially the pull requests, listed are those that made their way to develop but not to master.
>> What is the protocol/usage for having the stuff on the develop branch merge onto the master branch ?
> Generally Ive been waiting for testing to cycle for a while before merging changes to master.
>> I guess the stuff on develop is already considered as being validated in some way, so what is needed for the next step ? does some testing report need to be produced ? on a specific set of platform ?
> Since very few people test Boost MPI on develop or master, I try to be very cautious with commits to develop and merges to master. Its kinda easy to break things and, given the scarcity of MPI and Graph Parallel testers, I prefer to test the PRs, develop and master branches on internal resources to ensure I dont hit any problems before I push them to github.
> Are you familiar with the Boost testing pages? Can you contribute any testing resouces to specifically test MPI and Graph Parallel?
>> Also, does everything that is in master supposed to go automagically in the next release ?
> Generally yes, everything in master goes into the release, thats why were careful about what ends up in master.
> Boost-mpi mailing list
Boost-Commit list run by troyer at boostpro.com