uh, doing this properly is turning out to be pretty complicated:
[ ] adapt my (container-based, to be safe) action to push to gh_pages to:
merge in subdirectories but also deleting the history (so as to avoid the repo getting too fat)
write out a JSON of all the current branches and tags available on gh-pages and paths, because there's no way to do this directly in JS.
Needs a data object first.
This data object must sit at the root, and be overwritten by every push to gh-pages, updating it by reading the file structure anew.
set a robots.txt or whatever to make sure the other branches and tags aren't being indexed by search engines
deploy the latest release (or master or whatever) also to root
ensure that no directory name clashes with any of the refs
uh, doing this properly is turning out to be pretty complicated:
gh_pages
to:gh-pages
and paths, because there's no way to do this directly in JS. Needs a data object first. This data object must sit at the root, and be overwritten by every push togh-pages
, updating it by reading the file structure anew.refs.json
as a dropdown