Re: [DISCUSS] Changing many geode-core dependencies from compile to runtime

2019-03-16 Thread pulkit chandra
+1 this is a great idea for downstream systems like PCC. +1 for bom. It's just so much easier On Fri, Mar 15, 2019, 7:35 PM Jacob Barrett wrote: > +1 for the change and +1 for BOMs. > > We currently have an “all” BOM and a client BOM. Defining server and other > usecase derived BOMs should be e

Re: [DISCUSS] Changing many geode-core dependencies from compile to runtime

2019-03-15 Thread Jacob Barrett
+1 for the change and +1 for BOMs. We currently have an “all” BOM and a client BOM. Defining server and other usecase derived BOMs should be easy. -jake > On Mar 15, 2019, at 4:16 PM, John Blum wrote: > > If users will be explicitly declaring such dependencies in their > applications, then I

Re: [DISCUSS] Changing many geode-core dependencies from compile to runtime

2019-03-15 Thread John Blum
If users will be explicitly declaring such dependencies in their applications, then I might also suggest declaring/generating a Maven section in the POM to ensure that the user is getting and using the right version of these dependencies, especially when they don't care about the version (i.e. the

Re: [DISCUSS] Changing many geode-core dependencies from compile to runtime

2019-03-15 Thread Bruce Schuchardt
That seems like a great idea On 3/15/19 11:53 AM, Dan Smith wrote: Hi, We would like to start using gradle's new implementation dependency notation in our build files. This will affect downstream consumers of geode-core, hopefully in a good way, in that many of our dependencies will now be mar

[DISCUSS] Changing many geode-core dependencies from compile to runtime

2019-03-15 Thread Dan Smith
Hi, We would like to start using gradle's new implementation dependency notation in our build files. This will affect downstream consumers of geode-core, hopefully in a good way, in that many of our dependencies will now be marked runtime dependencies in the pom instead of compile. That means it