fetch_ci_vars_lists

#!/usr/bin/env bash

# SPDX-FileCopyrightText: 2021 Robin Vobruba <hoijui.quaero@gmail.com>
#
# SPDX-License-Identifier: Unlicense

# Exit immediately on each error and unset variable;
# see: https://vaneyckt.io/posts/safer_bash_scripts_with_set_euxo_pipefail/
set -Eeuo pipefail
#set -Eeu

script_dir=$(dirname "$(readlink -f "${BASH_SOURCE[0]}")")

function filter_tabular_html() {

  VARS_LIST_HTML_FILE="$1"
  COL_NUM_VAR_NAME="$2"
  COL_NUM_DESC="$3"
  cat "$VARS_LIST_HTML_FILE" \
    | awk -e '
  BEGIN {
    in_row = 0
    col_num = -99
    COL_NUM_VAR_NAME = '"$COL_NUM_VAR_NAME"'
    COL_NUM_DESC = '"$COL_NUM_DESC"'
  }

  /^<tr[^>]*>$/ {
    in_row = 1
    col_num = 0
  }

  /^<\/tr>$/ {
    in_row = 0
    col_num = -99
  }

  /^<td[^>]*>/ {
    if (col_num == COL_NUM_VAR_NAME) {
      var_name = $0
      gsub("^<td[^>]*>", "", var_name)
      gsub("</td>$", "", var_name)
      gsub("^<code[^>]*>", "", var_name)
      gsub("</code>$", "", var_name)
    }
    if (col_num == COL_NUM_DESC) {
      description = $0
      gsub("^<td[^>]*>", "", description)
      gsub("</td>$", "", description)
      printf("%s=\"%s\"\n" , var_name, description)
    }
    col_num = col_num + 1
  }
  ' | sort
}

function filter_list_html() {

  VARS_LIST_HTML_FILE="$1"
  cat "$VARS_LIST_HTML_FILE" \
    | awk -e '
  BEGIN {
    var_name_set = 0
    desc_set = 0
  }

  {
    if (var_name_set && !desc_set) {
      # We actually want the space inserted between them here!
      desc = desc " " $0
      if (match($0, "<li>")) {
        open_li_s = open_li_s + 1
      }
      if (match($0, "</li>")) {
        open_li_s = open_li_s - 1
      }
      if (open_li_s == 0) {
        desc_set = 1
      }
    }
    if (desc_set) {
      printf("%s=\"%s\"\n" , var_name, desc)
      var_name_set = 0
      desc_set = 0
    }
  }

  /^<li><code class="highlighter-rouge">.*<\/code>:/ {
      var_name = $0
      gsub("^<li[^>]*>", "", var_name)
      gsub("</li>$", "", var_name)
      gsub("^<code[^>]*>", "", var_name)
      gsub("</code>.*", "", var_name)
      var_name_set = 1
      if (match($0, "</li>$")) {
        desc = $0
        gsub(".*</code>: ", "", desc)
        gsub("</li>$", "", desc)
        desc_set = 1
      } else {
        desc = ""
        open_li_s = 1
      }
  }
  ' | sort
}

function cleanup_html_files() {

  tidy -wrap 4096 --show-warnings false \
    -modify \
    "$@" \
    2> /dev/null || true
}

VARS_LIST_HTML_GITLAB_URL="https://docs.gitlab.com/ee/ci/variables/predefined_variables.html"
VARS_LIST_HTML_GITHUB_URL="https://docs.github.com/en/actions/reference/environment-variables"
VARS_LIST_HTML_BITBUCKET_URL="https://support.atlassian.com/bitbucket-cloud/docs/variables-and-secrets/"
VARS_LIST_HTML_JENKINS_URL="https://jenkins-x.io/docs/resources/guides/using-jx/pipelines/envvars/"
VARS_LIST_HTML_TRAVISCI_URL="https://docs-staging.travis-ci.com/user/environment-variables/#default-environment-variables"
# NOTE CirceCI does not use environment variables. We suggest to manually transfer config values from there into environment variables.
# NOTE TeamCity uses (almost) no environment variables. We suggest to manually transfer config values from there into environment variables.

VARS_LIST_HTML_GITLAB_FILE="target/ci_vers_gitlab.html"
VARS_LIST_HTML_GITHUB_FILE="target/ci_vers_github.html"
VARS_LIST_HTML_BITBUCKET_FILE="target/ci_vers_bitbucket.html"
VARS_LIST_HTML_JENKINS_FILE="target/ci_vers_jenkins.html"
VARS_LIST_HTML_TRAVISCI_FILE="target/ci_vers_travisci.html"

GITLAB_FILE="target/gitlab.sh.env"
GITHUB_FILE="target/github.sh.env"
BITBUCKET_FILE="target/bitbucket.sh.env"
JENKINS_FILE="target/jenkins.sh.env"
TRAVISCI_FILE="target/travisci.sh.env"

download=false
if $download
then
  wget "$VARS_LIST_HTML_GITLAB_URL"    -O "$VARS_LIST_HTML_GITLAB_FILE"
  wget "$VARS_LIST_HTML_GITHUB_URL"    -O "$VARS_LIST_HTML_GITHUB_FILE"
  wget "$VARS_LIST_HTML_BITBUCKET_URL" -O "$VARS_LIST_HTML_BITBUCKET_FILE"
  wget "$VARS_LIST_HTML_JENKINS_URL"   -O "$VARS_LIST_HTML_JENKINS_FILE"
  wget "$VARS_LIST_HTML_TRAVISCI_URL"  -O "$VARS_LIST_HTML_TRAVISCI_FILE"
fi

# Cleanup the downloaded HTML
cleanup_html_files \
  "$VARS_LIST_HTML_GITLAB_FILE" \
  "$VARS_LIST_HTML_GITHUB_FILE" \
  "$VARS_LIST_HTML_BITBUCKET_FILE" \
  "$VARS_LIST_HTML_JENKINS_FILE" \
  "$VARS_LIST_HTML_TRAVISCI_FILE"

sed -i -e 's|<p[^>]*>||g' -e 's|</p>||g' "$VARS_LIST_HTML_BITBUCKET_FILE"
cleanup_html_files "$VARS_LIST_HTML_BITBUCKET_FILE"

filter_tabular_html "$VARS_LIST_HTML_GITLAB_FILE" 0 3 > "$GITLAB_FILE"
filter_tabular_html "$VARS_LIST_HTML_GITHUB_FILE" 0 1 > "$GITHUB_FILE"
filter_tabular_html "$VARS_LIST_HTML_BITBUCKET_FILE" 0 1 > "$BITBUCKET_FILE"
filter_tabular_html "$VARS_LIST_HTML_JENKINS_FILE" 0 1 > "$JENKINS_FILE"
filter_list_html "$VARS_LIST_HTML_TRAVISCI_FILE" > "$TRAVISCI_FILE"