Open Farof opened 8 years ago
Should the reference stay in this project and in what form: files, wiki or both? Or make a dedicated project?
Would be nice to team up with people working in other languages if possible.
Because Javascript is interpreted, I'm not opposed to make node-gyp modules (using c) for these, especially the mpq functions. Most of the time wasted in extracting the data is there.
Current work visible on the feature-automation
branch.
It's a work in progress, currently porting from a local clone of the heroprotocol
repository, but I want for the script to do it's own clone/pull.
The porting script reads the Python lines and transforms them into Javascript object representation of the data structure and events, thus paving the way for the automatically generated reference.
Considering the feature-automation
branch, a conundrum has occurred. Should the porting script and the distributables be in the same repo?
The following two questions are not mutually exclusive:
heroprotocoljs-build
repo?dist/
folder to copy into higher-level application; or, automatic github uploader to personal repo; or, automatic npm uploader to personal namespace)Two cases:
-build
repo automatically generates this repo.-build
repo would let them do so. Example: I have a minimally-working fork using storm-replay
rather than mpyqjs
. If there was a -build
repo, I could use it's script to grab the Blizzard/heroprotocol
repo and update the templates.Food for thought...
Writing an automated reference builder should be fairly easy using the protocol data structure and events description.
It should be able to track what was added or removed in each build and be backward compatible.
It could either output directly a reference or some intermediate files to be consumed by existing automatic documentation builder.