Re: [Boost-mpi] Develop VS master

Subject: Re: [Boost-mpi] Develop VS master
From: MM (finjulhich_at_[hidden])
Date: 2017-07-14 22:01:24

On 14 July 2017 at 20:40, Richard via Boost-mpi
<boost-mpi_at_[hidden]> wrote:
>
> Hi Alain,
>
> please don't get me wrong.
> I would certainly appreciate merging the develop branch to the master to
> get improvements and new features in boost.mpi.
> And if develop is well tested already, I do not see any reason not to
> merge it.
> But I'm not in charge...
>
> I just think the merge deserves to be independent of merely fixing the
> master branch. The changes in develop are worth way more than just
> fixing a regression.
>
> I also do understand that you would like to avoid cherry-picking from
> develop to master, but we are taking about a regression fix, not new
> developments.
> In fact I was just about to submit my own merge request of the patch I
> wrote for this bug, when I found a similar fix in the develop branch.
>
> Best,
> Richard

I believe we are not supposed to do top-posting if that's what it's called.

I am a long term user of boost-mpi with openmpi and I would love to
carry on doing that
+1 for whoever reported build issues with what's in develop to put
logs and compiler version to help with that.
As a occasional but long time user, I have a very far view of boost
collection, but the well know present of lack of maintainers is
forever present.
The fact that Alain is actively working on boost/mpi is a strong
argument for me to keep using boost-mpi.
I would love that complications should not discourage him from
carrying on the maintenance work he's doing especially that the
original implementors have not been seen on these mailing lists for
years, most likely they moved on to other pastures (it's been 12/3
years or so...)
Dream wishes for boost mpi are extension to support MPI3-1, for
c++11/14 if applicable, clearer documentation in terms of optimization
of mpi datatypes, and usage of serialization lib.

rds,