Files
coder/scripts/release/docs_update_experiments.sh

184 lines
5.6 KiB
Bash
Executable File

#!/usr/bin/env bash
# Usage: ./docs_update_experiments.sh
#
# This script updates the available experimental features in the documentation.
# It fetches the latest mainline and stable releases to extract the available
# experiments and their descriptions. The script will update the
# feature-stages.md file with a table of the latest experimental features.
set -euo pipefail
# shellcheck source=scripts/lib.sh
source "$(dirname "${BASH_SOURCE[0]}")/../lib.sh"
cdroot
if isdarwin; then
dependencies gsed gawk
sed() { gsed "$@"; }
awk() { gawk "$@"; }
fi
# From install.sh
echo_latest_stable_version() {
# https://gist.github.com/lukechilds/a83e1d7127b78fef38c2914c4ececc3c#gistcomment-2758860
version="$(curl -fsSLI -o /dev/null -w "%{url_effective}" https://github.com/coder/coder/releases/latest)"
version="${version#https://github.com/coder/coder/releases/tag/v}"
echo "v${version}"
}
echo_latest_mainline_version() {
# Fetch the releases from the GitHub API, sort by version number,
# and take the first result. Note that we're sorting by space-
# separated numbers and without utilizing the sort -V flag for the
# best compatibility.
echo "v$(
curl -fsSL https://api.github.com/repos/coder/coder/releases |
awk -F'"' '/"tag_name"/ {print $4}' |
tr -d v |
tr . ' ' |
sort -k1,1nr -k2,2nr -k3,3nr |
head -n1 |
tr ' ' .
)"
}
# For testing or including experiments from `main`.
echo_latest_main_version() {
echo origin/main
}
sparse_clone_codersdk() {
mkdir -p "${1}"
cd "${1}"
rm -rf "${2}"
git clone --quiet --no-checkout "${PROJECT_ROOT}" "${2}"
cd "${2}"
git sparse-checkout set --no-cone codersdk
git checkout "${3}" -- codersdk
echo "${1}/${2}"
}
parse_all_experiments() {
# Go doc doesn't include inline array comments, so this parsing should be
# good enough. We remove all whitespaces so that we can extract a plain
# string that looks like {}, {ExpA}, or {ExpA,ExpB,}.
#
# Example: ExperimentsAll=Experiments{ExperimentNotifications,ExperimentAutoFillParameters,}
go doc -all -C "${dir}" ./codersdk ExperimentsAll |
tr -d $'\n\t ' |
grep -E -o 'ExperimentsAll=Experiments\{[^}]*\}' |
sed -e 's/.*{\(.*\)}.*/\1/' |
tr ',' '\n'
}
parse_experiments() {
# Extracts the experiment name and description from the Go doc output.
# The output is in the format:
#
# ||Add new experiments here!
# ExperimentExample|example|This isn't used for anything.
# ExperimentAutoFillParameters|auto-fill-parameters|This should not be taken out of experiments until we have redesigned the feature.
# ExperimentMultiOrganization|multi-organization|Requires organization context for interactions, default org is assumed.
# ExperimentCustomRoles|custom-roles|Allows creating runtime custom roles.
# ExperimentNotifications|notifications|Sends notifications via SMTP and webhooks following certain events.
# ExperimentWorkspaceUsage|workspace-usage|Enables the new workspace usage tracking.
# ||ExperimentTest is an experiment with
# ||a preceding multi line comment!?
# ExperimentTest|test|
#
go doc -all -C "${1}" ./codersdk Experiment |
sed \
-e 's/\t\(Experiment[^ ]*\)\ \ *Experiment = "\([^"]*\)"\(.*\/\/ \(.*\)\)\?/\1|\2|\4/' \
-e 's/\t\/\/ \(.*\)/||\1/' |
grep '|'
}
workdir=build/docs/experiments
dest=docs/contributing/feature-stages.md
log "Updating available experimental features in ${dest}"
declare -A experiments=() experiment_tags=()
for channel in mainline stable; do
log "Fetching experiments from ${channel}"
tag=$(echo_latest_"${channel}"_version)
dir="$(sparse_clone_codersdk "${workdir}" "${channel}" "${tag}")"
declare -A all_experiments=()
all_experiments_out="$(parse_all_experiments "${dir}")"
if [[ -n "${all_experiments_out}" ]]; then
readarray -t all_experiments_tmp <<<"${all_experiments_out}"
for exp in "${all_experiments_tmp[@]}"; do
all_experiments[$exp]=1
done
fi
# Track preceding/multiline comments.
maybe_desc=
while read -r line; do
line=${line//$'\n'/}
readarray -d '|' -t parts <<<"$line"
# Missing var/key, this is a comment or description.
if [[ -z ${parts[0]} ]]; then
maybe_desc+="${parts[2]//$'\n'/ }"
continue
fi
var="${parts[0]}"
key="${parts[1]}"
desc="${parts[2]}"
desc=${desc//$'\n'/}
# If desc (trailing comment) is empty, use the preceding/multiline comment.
if [[ -z "${desc}" ]]; then
desc="${maybe_desc% }"
fi
maybe_desc=
# Skip experiments not listed in ExperimentsAll.
if [[ ! -v all_experiments[$var] ]]; then
log "Skipping ${var}, not listed in ExperimentsAll"
continue
fi
# Don't overwrite desc, prefer first come, first served (i.e. mainline > stable).
if [[ ! -v experiments[$key] ]]; then
experiments[$key]="$desc"
fi
# Track the release channels where the experiment is available.
experiment_tags[$key]+="${channel}, "
done < <(parse_experiments "${dir}")
done
table="$(
if [[ "${#experiments[@]}" -eq 0 ]]; then
echo "Currently no experimental features are available in the latest mainline or stable release."
exit 0
fi
echo "| Feature | Description | Available in |"
echo "|---------|-------------|--------------|"
for key in "${!experiments[@]}"; do
desc=${experiments[$key]}
tags=${experiment_tags[$key]%, }
echo "| \`$key\` | $desc | ${tags} |"
done
)"
# Use awk to print everything outside the BEING/END block and insert the
# table in between.
awk \
-v table="${table}" \
'BEGIN{include=1} /BEGIN: available-experimental-features/{print; print table; include=0} /END: available-experimental-features/{include=1} include' \
"${dest}" \
>"${dest}".tmp
mv "${dest}".tmp "${dest}"
# Format the file for a pretty table (target single file for speed).
(cd site && pnpm exec prettier --cache --write ../"${dest}")