Various things that make it easier for me personally to use:
Use an HTTPS pull URL (still the SSH push URL). This is because pulling over HTTPS doesn't need authentication but pushing needs me to tap my security key; cloning over HTTPS halves the number of taps I need.
Don't require libjsonnet.wasm to exist - we can just keep the existing one from the gh-pages branch if there isn't a newly built one to use.
Require running on a TTY and prompt the user for confirmation before the push. Probably not what we want long term but I was running the script a lot to check things and didn't want to actually push then.
Then update docs/ref/stdlib.html by running update_web_content.sh. We might want to get rid of that file; if not then we'll probably need to add an automated check that it's not out of sync with the source stdlib-content.jsonnet.
Various things that make it easier for me personally to use:
Use an HTTPS pull URL (still the SSH push URL). This is because pulling over HTTPS doesn't need authentication but pushing needs me to tap my security key; cloning over HTTPS halves the number of taps I need.
Don't require libjsonnet.wasm to exist - we can just keep the existing one from the gh-pages branch if there isn't a newly built one to use.
Require running on a TTY and prompt the user for confirmation before the push. Probably not what we want long term but I was running the script a lot to check things and didn't want to actually push then.
Then update docs/ref/stdlib.html by running update_web_content.sh. We might want to get rid of that file; if not then we'll probably need to add an automated check that it's not out of sync with the source stdlib-content.jsonnet.