mirror of
https://github.com/discourse/discourse.git
synced 2026-08-08 17:53:55 +08:00
GitHub oneboxes and the discourse-github plugin talked to GitHub's REST
and
GraphQL API with no rate-limit awareness. On busy instances this
exhausted
GitHub's limits (60 requests/hour unauthenticated, 5000 authenticated),
and
because there was no backoff every render kept hitting GitHub and
re-failing
-- which GitHub's docs warn can get an integration banned. The recently
added PR-status onebox multiplied the number of calls and made it far
worse.
GitHub access was also fragmented: the core onebox engines used OpenURI,
the
discourse-github plugin used Octokit, and the discourse-ai bot tools
used
FinalDestination::HTTP -- three HTTP stacks, three tokens, and
inconsistent
(or entirely missing) error and rate-limit handling.
This introduces a single client, Discourse::GithubApi, that every GitHub
data-API request now flows through. It is built on Faraday with the
SSRF-safe
FinalDestination adapter and:
- authenticates per token (Bearer) and returns plain string-keyed Hashes
(get/post) or raw bodies (raw_get) -- one response shape, no
Octokit/Sawyer
- only ever sends the access token to api.github.com and
raw.githubusercontent.com, rejecting any other absolute URL, so a
user-derived path can never leak a token to an arbitrary host
- backs off on rate limits both reactively (403/429) and proactively
(when
X-RateLimit-Remaining hits 0), honouring Retry-After /
X-RateLimit-Reset,
via a shared Redis flag (GithubRateLimit) keyed per token so each
token's
budget and the shared unauthenticated/IP budget back off independently
- short-circuits while backing off without ever sleeping, so onebox
rendering
and post baking degrade to a plain link instead of blocking a request
- caches ETags and sends If-None-Match, so unchanged resources return
304s
that do not count against the rate limit
Every caller was moved onto it:
- the 6 core GitHub onebox engines, via a slimmed
Onebox::Mixins::GithubApi
adapter that keeps their public methods and translates client errors
back
to the OpenURI::HTTPError vocabulary they already rescue (engines
unchanged)
- the github_blob raw.githubusercontent.com fetch
- the discourse-github plugin (badges, linkback, permalinks, token
validator),
which no longer uses the octokit and sawyer gems (they stay in the
Gemfile for
the discourse-code-review official plugin, which still depends on them)
- the discourse-ai bot's GitHub tools (search code, diff, file content,
search files)
Also adds a GithubOneboxBackoff admin problem check that surfaces while
one of
the onebox token identities is backing off -- scoped to the tokens
resolved by
Onebox::GithubAccess (each configured github_onebox_access_tokens entry
plus the
unauthenticated client) so a backoff on the AI bot or linkback token is
not
misattributed to onebox. Its message points admins at the relevant
setting with
the {{setting:...}} link marker, which problem-check messages now expand
too.
Onebox token resolution is centralised in Onebox::GithubAccess, and the
onebox
cache TTL for transient GitHub failures is shortened so they recover
quickly.
GitHub OAuth login, theme git-clone, the inbound webhook, and the
Oneboxer
FinalDestination URL-resolution special-cases for github.com are
intentionally
out of scope -- they are different concerns, not the rate-limited data
API.
295 lines
9.4 KiB
Ruby
Vendored
295 lines
9.4 KiB
Ruby
Vendored
# frozen_string_literal: true
|
|
|
|
module DiscourseGithubPlugin
|
|
class CommitsPopulator
|
|
MERGE_COMMIT_REGEX = /^Merge pull request/
|
|
HISTORY_COMPLETE = "history-complete"
|
|
class GraphQLError < StandardError
|
|
end
|
|
|
|
ROLES = { committer: 0, contributor: 1 }
|
|
|
|
class PaginatedCommits
|
|
def initialize(client, repo, cursor: nil, page_size: 100)
|
|
@client = client
|
|
@repo = repo
|
|
@cursor = cursor
|
|
@page_size = page_size
|
|
raise ArgumentError, "page_size arg must be <= 100" if page_size > 100
|
|
if cursor && !cursor.match?(/^\h{40}\s(\d+)$/)
|
|
raise ArgumentError,
|
|
"cursor must be a 40-characters hex string followed by a space and a number"
|
|
end
|
|
fetch_commits
|
|
end
|
|
|
|
def next
|
|
cursor = next_cursor
|
|
return unless cursor
|
|
PaginatedCommits.new(@client, @repo, cursor: cursor, page_size: @page_size)
|
|
end
|
|
|
|
def commits
|
|
history["nodes"]
|
|
end
|
|
|
|
def next_cursor
|
|
info = history["pageInfo"]
|
|
return unless info["hasNextPage"]
|
|
info["endCursor"]
|
|
end
|
|
|
|
private
|
|
|
|
def history
|
|
@data.dig("repository", "defaultBranchRef", "target", "history")
|
|
end
|
|
|
|
def fetch_commits
|
|
owner, name = @repo.name.split("/", 2)
|
|
history_args = "first: #{@page_size}"
|
|
history_args += ", after: #{@cursor.inspect}" if @cursor
|
|
|
|
query = <<~QUERY
|
|
query {
|
|
repository(name: #{name.inspect}, owner: #{owner.inspect}) {
|
|
defaultBranchRef {
|
|
target {
|
|
... on Commit {
|
|
history(#{history_args}) {
|
|
pageInfo {
|
|
endCursor
|
|
hasNextPage
|
|
}
|
|
nodes {
|
|
oid
|
|
message
|
|
committedDate
|
|
associatedPullRequests(first: 1) {
|
|
nodes {
|
|
author {
|
|
login
|
|
}
|
|
mergedBy {
|
|
login
|
|
}
|
|
}
|
|
}
|
|
author {
|
|
email
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
QUERY
|
|
response = @client.post("/graphql", { query: query })
|
|
raise GraphQLError, "Empty GraphQL response" if response.nil?
|
|
raise GraphQLError, response["errors"].inspect if response["errors"]
|
|
raise GraphQLError, response.inspect if !response["data"]
|
|
@data = response["data"]
|
|
end
|
|
end
|
|
|
|
def initialize(repo)
|
|
@repo = repo
|
|
@client = Discourse::GithubApi.for(token: SiteSetting.github_linkback_access_token)
|
|
end
|
|
|
|
def populate!
|
|
return unless SiteSetting.github_badges_enabled?
|
|
return if @client.backing_off?
|
|
return if @client.get("/repos/#{@repo.name}/branches").blank?
|
|
|
|
if @repo.commits.size == 0
|
|
build_history!
|
|
else
|
|
front_sha = Discourse.redis.get(front_commit_redis_key)
|
|
if front_sha.present? && removed?(front_sha)
|
|
# there has been a force push, next run will rebuild history
|
|
@repo.commits.delete_all
|
|
Discourse.redis.del(back_cursor_redis_key)
|
|
Discourse.redis.del(front_commit_redis_key)
|
|
return
|
|
end
|
|
fetch_new_commits!(front_sha)
|
|
front_sha = Discourse.redis.get(front_commit_redis_key)
|
|
@repo.reload
|
|
|
|
back_cursor = Discourse.redis.get(back_cursor_redis_key)
|
|
return if back_cursor == HISTORY_COMPLETE
|
|
if back_cursor.present?
|
|
build_history!(cursor: back_cursor)
|
|
elsif front_sha.present?
|
|
count = @repo.commits.count
|
|
build_history!(cursor: "#{front_sha} #{count - 1}")
|
|
else
|
|
# this is a bad state that we should never be in
|
|
# But in case it happens, easiest way to recover
|
|
# is to start from scratch.
|
|
@repo.commits.delete_all
|
|
Discourse.redis.del(back_cursor_redis_key)
|
|
Discourse.redis.del(front_commit_redis_key)
|
|
end
|
|
end
|
|
rescue Discourse::GithubApi::NotFound
|
|
disable_github_badges_and_inform_admin(
|
|
title: I18n.t("github_commits_populator.errors.repository_not_found_pm_title"),
|
|
raw:
|
|
I18n.t(
|
|
"github_commits_populator.errors.repository_not_found_pm",
|
|
repo_name: @repo.name,
|
|
base_path: Discourse.base_path,
|
|
),
|
|
)
|
|
Rails.logger.warn(
|
|
"Disabled github_badges_enabled site setting due to repository Not Found error ",
|
|
)
|
|
rescue Discourse::GithubApi::Unauthorized
|
|
disable_github_badges_and_inform_admin(
|
|
title: I18n.t("github_commits_populator.errors.invalid_octokit_credentials_pm_title"),
|
|
raw:
|
|
I18n.t(
|
|
"github_commits_populator.errors.invalid_octokit_credentials_pm",
|
|
base_path: Discourse.base_path,
|
|
),
|
|
)
|
|
Rails.logger.warn(
|
|
"Disabled github_badges_enabled site setting due to invalid GitHub authentication credentials via github_linkback_access_token.",
|
|
)
|
|
rescue Discourse::GithubApi::Error => err
|
|
Rails.logger.warn("#{err.class}: #{err.message}")
|
|
end
|
|
|
|
private
|
|
|
|
def is_contribution?(commit)
|
|
pr = commit.dig("associatedPullRequests", "nodes")&.first
|
|
pr && pr["author"] && pr["mergedBy"] &&
|
|
pr.dig("author", "login") != pr.dig("mergedBy", "login")
|
|
end
|
|
|
|
def fetch_new_commits!(stop_at)
|
|
paginator = PaginatedCommits.new(@client, @repo, page_size: 10)
|
|
batch = paginator.commits
|
|
done = false
|
|
commits = []
|
|
recent_commits =
|
|
stop_at.present? ? [] : @repo.commits.order("committed_at DESC").first(100).pluck(:sha)
|
|
while !done
|
|
batch.each do |c|
|
|
if c["oid"] == stop_at || recent_commits.include?(c["oid"])
|
|
done = true
|
|
break
|
|
end
|
|
commits << c
|
|
end
|
|
break if done
|
|
paginator = paginator.next
|
|
batch = paginator&.commits || []
|
|
break if batch.empty?
|
|
end
|
|
return if commits.size == 0
|
|
existing_shas = @repo.commits.pluck(:sha)
|
|
commits.reject! { |c| existing_shas.include?(c["oid"]) }
|
|
batch_to_db(commits)
|
|
set_front_commit(commits.first["oid"])
|
|
end
|
|
|
|
# detect if a force push happened and commit is lost
|
|
def removed?(sha)
|
|
commit = @client.get("/repos/#{@repo.name}/commits/#{sha}")
|
|
return false if commit.nil?
|
|
return true if commit["commit"].nil?
|
|
found =
|
|
@client.get(
|
|
"/repos/#{@repo.name}/commits",
|
|
until: commit.dig("commit", "committer", "date"),
|
|
page: 1,
|
|
per_page: 1,
|
|
).first
|
|
return false if found.nil?
|
|
commit["sha"] != found["sha"]
|
|
end
|
|
|
|
def build_history!(cursor: nil)
|
|
paginator = PaginatedCommits.new(@client, @repo, cursor: cursor, page_size: 100)
|
|
batch = paginator.commits
|
|
return if batch.empty?
|
|
set_front_commit(batch.first["oid"]) if cursor.blank?
|
|
|
|
while batch.size > 0
|
|
batch_to_db(batch)
|
|
next_cursor = paginator.next_cursor
|
|
set_back_cursor(next_cursor) if next_cursor
|
|
|
|
paginator = paginator.next
|
|
batch = paginator&.commits || []
|
|
end
|
|
set_back_cursor(HISTORY_COMPLETE)
|
|
end
|
|
|
|
def batch_to_db(batch)
|
|
fragments = []
|
|
batch.each do |c|
|
|
hash = commit_to_hash(c)
|
|
fragments << DB.sql_fragment(<<~SQL, hash)
|
|
(:repo_id, :sha, :email, :committed_at, :role_id, :merge_commit, CURRENT_TIMESTAMP, CURRENT_TIMESTAMP)
|
|
SQL
|
|
end
|
|
DB.exec(<<~SQL)
|
|
INSERT INTO github_commits
|
|
(repo_id, sha, email, committed_at, role_id, merge_commit, created_at, updated_at) VALUES #{fragments.join(",")}
|
|
SQL
|
|
end
|
|
|
|
def commit_to_hash(commit)
|
|
{
|
|
sha: commit["oid"],
|
|
email: commit.dig("author", "email"),
|
|
repo_id: @repo.id,
|
|
committed_at: commit["committedDate"],
|
|
merge_commit: commit["message"].match?(MERGE_COMMIT_REGEX),
|
|
role_id: is_contribution?(commit) ? ROLES[:contributor] : ROLES[:committer],
|
|
}
|
|
end
|
|
|
|
def set_front_commit(sha)
|
|
Discourse.redis.set(front_commit_redis_key, sha)
|
|
end
|
|
|
|
def set_back_cursor(cursor)
|
|
Discourse.redis.set(back_cursor_redis_key, cursor)
|
|
end
|
|
|
|
def front_commit_redis_key
|
|
# this key should refer to the MOST RECENT commit we have in the db
|
|
"discourse-github-front-commit-#{@repo.name}"
|
|
end
|
|
|
|
def back_cursor_redis_key
|
|
# this key should refer to the cursor that lets us continue
|
|
# building history from the point we reached in the previous run
|
|
# that couldn't continue for whatever reasons
|
|
# e.g., if we got rate-limited by github
|
|
"discourse-github-back-cursor-#{@repo.name}"
|
|
end
|
|
|
|
def disable_github_badges_and_inform_admin(title:, raw:)
|
|
SiteSetting.github_badges_enabled = false
|
|
site_admin_usernames =
|
|
User.where(admin: true).human_users.order("last_seen_at DESC").limit(10).pluck(:username)
|
|
PostCreator.create!(
|
|
Discourse.system_user,
|
|
title: title,
|
|
raw: raw,
|
|
archetype: Archetype.private_message,
|
|
target_usernames: site_admin_usernames,
|
|
skip_validations: true,
|
|
)
|
|
end
|
|
end
|
|
end
|