Closed mortenvp closed 10 years ago
@mortenvp
Great! I have been taking a look and it seems nice. Actually this is what I was thinking about to do at the very end to clean up the code, but after the "all green" builds which is what I'm about to finish. It is better because it helps to keep modularity and readability as we have discussed.
About the failed tests, I'll post an answer on the mailing list with details, but I have to go to the university first. I think the better here is just to set pre-defined false
values as test parameters in each CPU when you don't have a capability. Jeppe gave me a hand here, so it's making some changes in the buildbot.
ok, so I will merge this one
@nestorjhernandezm just did some final changes - review if you like then merge. The buildbot builds the project on all platforms only the tests fail - because of wrong or missing arguments.
@mortenvp
Its great, definitively the way it should be. If just in the other approach there would had been few differences between each implementation then that approach would have been better. At the very end, since there were several variants, a single source file was hard to read. Good, let's merge this one. I like it :+1:
@nestorjhernandezm I tried an alternative way of splitting the implementation.. I think it works pretty well.