On 2018-10-22 03:08 PM, Robert Ramey via Boost wrote:
On 10/22/18 11:45 AM, Stefan Seefeld via Boost wrote:
OK... do we want CMake config files installed by 1.69 or not, then?
I'd like to know this as well. If so, I presume that we can consider the Boost/CMake issues resolved and we can just move on from this?
What issues do you think are resolved ? As far as I understand, the current endeavor is to auto-generate config files to make it easier to consume (individual) Boost libraries in CMake projects. That's entirely orthogonal to the question of whether and how Boost libraries switch to CMake.
So, as far as I'm concerned, your effort to agree on a Request For Proposals is still very much relevant.
The reason I ask is that working to implement something presumes that there's consensus on what should be implemented. I don't want to find myself in the position of committing serious effort to a task which has been eclipsed by events.
Historically, Boost Tooling has been developed in just this way. Someone with access just implements something and drops on the rest of us. I think that this practice is one of the reasons that we've been frustrated with boost tooling in general.
I understand, and agree. In fact, my main argument against the "wholesale move to CMake" is precisely that we (the maintainers) are left to deal with users who will expect support when something doesn't quite work the way *they* expect. So even if the initial infrastructure generation is done by someone else (or is automated entirely), this still has an impact on the projects' maintenance. In that sense I think the inclusion of those config files, or at least, any "official support" that may come with it (implied or not), is something that needs to be agreed upon. And *that* would definitely be part of your RFP. Stefan -- ...ich hab' noch einen Koffer in Berlin...