John Maddock wrote:
Robert Jones wrote:
I am unclear of what the vault contains, and of its relation to the sandbox.
The sandbox is a versioned repository, whereas the the vault appears not to be.
What is the workflow here, if there is one? By what path do libraries arrive at Boost proper (ie. part of some release), and where in the process are the review points?
Both the vault and the sandbox provide alternative means developers can use to make code still in development available: the vault provides zip file downloads while the sandbox provides full version control when that's more applicable. Developers producing code for Boost are free to use either, both or neither depending on preference :-)
I honestly do not see the point of having two different locations for putting nascent implementations for others who are interested to try out, which may eventually be submitted to Boost. I believe it would be much cleaner and clearer if there were a single path, which all end users could peruse, and that path had a regularized way for access both by the developer and the end user. Having two different places, and having implementations spread out in those places where they are hard to find and difficult to try out, does not help Boost in its incubation of new ideas and libraries based on those ideas. As someone who has at least thought of putting code somewhere 'out there' for other interested Boost end users to try, this shotgun approach of getting end users interested and uploading implementations 'somewhere' does not suit my approach to such a situation. For me a more logically rigorous "this is what you need to do" suits me much better.