On 10/5/15, 4:39 PM, "Justin Mclean" <justinmcl...@me.com> wrote:
>Hi, > >> So IMO, the question still remains: how do you develop a framework >>and a compiler at the same time and run automated tests? > >By using CI. CI doesn’t only have to run only checkin to the dev branch >or even a single repo, you can run it on checkins to your own branches >and even against your working development environment How useful that is >may depending little on how fast your tests run. I think you may not have picked up the full context of the question. If you have a repo of framework code, its automated tests need the compiler to compile them. Meanwhile repo of compiler code has an automated test to prove that any change to the compiler can still compile the framework. This creates a circularity: when starting from a clean computer, you can’t build the compiler and run all of its tests because the framework hasn’t been built and its build cannot complete if it fails because there is no compiler. It also creates a circularity when a bug fix or new feature requires both a change to the compiler and the framework. So far, the only suggestion I have is to make failing test not fail the build, but then you run the risk that a true failure isn’t noticed right away. Concrete or more detailed suggestions are welcome. We are already using CI. -Alex
