This contains a couple of core modules written in TypeScript to be used with Screeps. It also contains a Gulpfile to compile these into a complete package. Finally there is a sync tool that can be used to sync the generated scripts with the browser.
Currently I develop this using JetBrains WebStorm, take a look at their site, if you're interested.
Run npm install in the root to get all the required packages.
gulp build
Compiles the typescript files in the src directory to the output.
gulp deploy
Simply copies the contents of the output to the deploy.
This is to allow you to compile your scripts and deploy them once you're satified.
gulp sync/syncDeploy
Runs a full proxy server on http://localhost:9090/ that sits between the browser and the game. This allows us to sync your scripts by injecting a client-side helper, but also enables us to provide additional features, such as sourcemaps and engine source stepping.
The sync task deploys using the deploy directory.
The syncDeploy task deploys directly from the output directory and thus can be used to update your browser session immediately after saving source changes.
Please note that the proxy uses a blacklist/whitelist to control what gets communicated with the server. I don't want scores and replays posted for stuff done via a debugging proxy.
gulp syncDebug/syncDeployDebug
These Debug variants do a bit more trickery client-side, it's purely intended to allow remote debugging from WebStorm, do not use it otherwise.
- We inject a FakeWorker to replace the multi-threaded WebWorker model with a single-thread model. This is required since WebStorm cannot break in a webworker context.
- We disable the engine try-catch mechanism which would swallow uncaught exceptions, allowing you to break on uncaught exceptions.
Initially I opted to compile all scripts into one file, but that made it somewhat harder to debug properly. So instead I've separated it into modules.
The important bits:
lib/Screeps is the module that contains the definitions of the core Screeps api.
world/Game contains wrapped instances of that api exposing additional functionality as well as caching.
lib/* contains various other utilities and helpers, such as automatic name generation, although you might not like my pattern.
engines/* intended to contain global decision logic.
behaviors/* intended to contain specific mixins/behaviors that can be applied to for example Creeps.
mission/* intended to contain persistable missions and assignments.
Lemme know if you're working on an interesting feature. I would love to see some collaboration, so if you have something you want to contribute, make an issue or pull request.