Currently, we generate an Atom-style CSON grammar file. However, support for TextMate-style scopes is weak. So this has two tasks:
The last task requires some design thinking: The grammar file is automatically generated. Automatically generated files are best kept out of git. However, Atom language extensions must be published as an Atom package (e.g., language-emerald), which in turn requires that the package is backed by a GitHub repository.
It seems that there is no choice but to include the automatically generated file in modes-atom. There could however be some tooling to avoid common mistakes with having automatically generated code in your repository.
Currently, we generate an Atom-style CSON grammar file. However, support for TextMate-style scopes is weak. So this has two tasks:
The last task requires some design thinking: The grammar file is automatically generated. Automatically generated files are best kept out of git. However, Atom language extensions must be published as an Atom package (e.g., language-emerald), which in turn requires that the package is backed by a GitHub repository.
It seems that there is no choice but to include the automatically generated file in modes-atom. There could however be some tooling to avoid common mistakes with having automatically generated code in your repository.