[torqueusers] [torquedev] Movement to Newer Libraries - Boost
"Mgr. Šimon Tóth"
toth at fi.muni.cz
Tue Oct 22 14:24:44 MDT 2013
> For our next major release (not coming out until next year but starting
> development in the near to immediate feature). As part of this release,
> we are planning some major optimizations for TORQUE. In an ideal world,
> we'd like to be able to use either Boost 1.4x or 1.5x and g++ 4.6, or a
> more radical option might be to compile using Clang 3.0 and llvm.
> The reason we want to use these is they offer well-tested, built-in code
> that can replace some existing code and code we'd have to write
> otherwise. It is strongly preferred not to write our own implementations
> of things when there are existing, functional implementations available
> to us.
> Of course, we understand that for some of you all, moving to these
> versions might create some work. So please give me your feedback. How
> much work would it be? If you believe it would not be possible to
> upgrade to these, please let me know the technical details of why.
Torque could really use a dual release mode. Maintain a
long-term-support version that will really be stable and will only add
bug fixes and parallel to that, do active development and release new
versions with short-term-support.
We are covering various users with our Torque fork, some are mainly
interested in just running the software as it is, without any further
requests for new features, others want continuous improvements.
Maintaining a long-term-support version would also be great for us,
since one of the main reasons why we are still stuck on the 2.4 version
as the base of our fork is that there is no bugfix-only branch. All the
branches include many new features that were breaking our fork. With a
long-term-support version, we could actually maintain a fork of that
version while staying up-to-date.
More information about the torqueusers