We should consider using es6 modules to build QUnit, moving from source concatenation to a standard format. The biggest issue with transpiling to es6 are dependencies, but since QUnit bundles its dependencies anyway, we don't have to care about that. There shouldn't be a difference for users either, since the built files would look mostly the same as they do now. If that works out, we could also use other es6 features in the QUnit source.
Currently our tests don't need a build step, which we may want to keep. If compilation is fast enough, we could move those to es6 as well, if there are obvious benefits.
Some background and examples: https://medium.com/@brianleroux/es6-modules-amd-and-commonjs-c1acefbe6fc0 (which mostly uses a transpiler and browserify, not exactly what we'd want)
We should consider using es6 modules to build QUnit, moving from source concatenation to a standard format. The biggest issue with transpiling to es6 are dependencies, but since QUnit bundles its dependencies anyway, we don't have to care about that. There shouldn't be a difference for users either, since the built files would look mostly the same as they do now. If that works out, we could also use other es6 features in the QUnit source.
Currently our tests don't need a build step, which we may want to keep. If compilation is fast enough, we could move those to es6 as well, if there are obvious benefits.
Some background and examples: https://medium.com/@brianleroux/es6-modules-amd-and-commonjs-c1acefbe6fc0 (which mostly uses a transpiler and browserify, not exactly what we'd want)