Following is the discussion extracted from the FedoraHosted's ticket following up the FESCo submission and process
- Updated by kevin 20:58, 28 June 2010 (UTC)
- Keywords set to meeting.
- Updated by mclasen 16:35, 29 June 2010 (UTC). Leaving my comments here, since I am likely going to miss FESCo today:
- Shipping upstream boost instead of a forked build system seems to be the right thing to do. Working with upstream on moving to more standard build tools is also a good idea, but shipping a fork is probably not the best way to do that...
- I'm not that familiar with boost. Will this update require a rebuild of all dependencies? If so, can we plan to do this in a separate tag?
- Updated by denis 13:30, 30 June 2010 (UTC):
- You are right, mclasen, it is a priori better to work with upstream shipped Boost. However, some developers continue working on the CMake build system for Boost. The latest Boost release, integrated with the CMake build system, is available on Gitorious. For instance, latest contributions have been made today (http://gitorious.org/~zeuner/boost/zeuners-boost-cmake/commits/1.44.0). The strategy of packaging Boost with a CMake-based build system has proved to be successful in Fedora 13, and has allowed to deliver new sub-packages such as Boost.MPI, which were absent from Fedora releases until now. Going back to a BJam-based build system (upstream choice) would mean giving up those new sub-packages, without speaking of a lot work to have BJam back to work on Fedora Rawhide/14. I do not underestimate the amount of work to be done for the CMake-based Boost 1.44, but it may well be worth the additional efforts.
- Yes, this update will require a rebuild of all the dependencies. That is the procedure each time Boost is updated. For Fedora 13, it went very smoothly, and no issue was raised.
- Updated by kevin 22:24, 29 June 2010 (UTC)
- Status changed from new to closed.
- Resolution set to fixed.
- This feature was approved at the 2010-06-29 meeting.