Files
archie/internal/forgerpc/forgerpc.go
T
sam e0b6aa6105 feat(workflow): post line-anchored review comments with suggested fixes (#891)
* feat(workflow): post line-anchored review comments with suggested fixes

Bead archie-core-q9au (docs/prds/inline-review.md). A review's findings reached
the operator only as a flat list in the PR body, so the author had to map every
finding back to its line by hand and apply mechanical fixes by hand. Now the
line-anchored findings are also posted as inline review comments, carrying a
one-click suggestion when the reviewer could state the fix.

Domain (internal/domain/workflow):
- ReviewFinding gains Suggestion; Validate rejects one on anything but a
  confirmed finding, so the agent's record_finding call is refused with
  something it can act on. Validate splits into validateAnchor /
  validateClassification / validateDisposition, which also brings it back under
  the cyclop limit the gate enforces.
- StagePostReviewComments runs after StageOpenPR in the implement workflow. It
  posts only line-anchored findings (Line > 0); a whole-file finding has no line
  to attach to and stays in the body list, which remains the complete record.
  It is best-effort: the PR is open and the body already lists every finding, so
  a failed comment is logged, never a park -- the same reasoning as OpenPR's
  existing best-effort LinkBranch call.
- The suggestion fence is gated on Verdict == confirmed where the fence is
  written, because it is a one-click apply: a wrong suggestion breaks the code
  and the author has to notice.

Forge (internal/forge): ReviewCommentWriter + GitHub/Gitea implementations,
type-asserted like PullRequestReader so a forge that cannot is skipped rather
than faked (NoopForge is pinned as not satisfying it). GitHub loops one
CreateComment per comment and joins the failures, so one refused anchor cannot
drop the rest; Gitea carries them on a single COMMENT-state review.

Two deliberate departures from the PRD's sketch, both recorded in it:
- The interface is batch (CreateReviewComments), not one call per comment:
  Gitea's only inline shape is one submitted review, and the set then crosses
  NATS once instead of once per finding.
- The head SHA is resolved inside each forge implementation rather than by the
  caller. The stage runs in archie-agent, which holds no forge credentials and
  has no PullRequestReader; the daemon-side forgerpc handler is the only place
  that does.

Transport (internal/forgerpc): workflow.Forger gains the method, so the agent
reaches the daemon over the new create_review_comments subject; the server
type-asserts the writer and answers an incapable forge with an error the stage
logs. This is the one proxied method whose absence on an image skew does not
park a task.

The reviewer prompt and record_finding schema ask for a suggestion on a
confirmed, mechanically fixable defect only.

Regenerating the wfextract symbol table -- the documented way a new exported
workflow symbol reaches interpreted stage plugins -- also picked up symbols that
were missing from the committed table (the Status* and Source* constants,
ReviewCheck, Store, Task, and the _Store wrapper). Nil-guards were re-applied to
every wrapper afterwards, as that file's header requires.

Evidence: task check exit 0 (incl. 229 UI tests). Three mutations of the new
stage logic each fail a test that names the behaviour -- dropping the
confirmed-only gate, dropping the Line > 0 filter, and letting a posting failure
escape -- so the tests discriminate rather than passing vacuously.

* fix(build): drop the Dockerfile pin assertion

The guard fetched golangci-lint's go.mod at image build time to fail the build
when GOLANGCI_LINT_VERSION and GOFUMPT_VERSION disagreed. It only ever tripped
on self-inflicted pin drift, but it added a network fetch and a new failure
mode to the release build, which is not where that belongs.

The pins stay correct (golangci-lint 2.13.2, whose vendored gofumpt v0.11.0
equals GOFUMPT_VERSION) and the comment above them still states the invariant,
so the reasoning survives without the build-time machinery. The pin change
itself reaches the agent image through the normal release path: CI rebuilds
and pushes archie-agent on an archie/vX.Y.Z release, and a gateway-only
release deliberately does not.

* feat(workflow): add the remediate workflow for PR review remediation

Implements step 5 of docs/prds/pr-review-remediation.md: a remediate
workflow that resumes the task's existing worktree onto its PR branch
(worktree.Manager.Resume, not Prepare's refresh), runs a builder agent
against the review's actionable comments, commits and pushes to the
same branch, and replies via the forge's ReplyToReview. Task.RetryCount
doubles as the remediation round counter, capped by the repo's
max_retries; the round-cap stage parks the task and posts one comment
when it's hit.

Task gains a ReviewPayload field (JSON-encoded review unit) with full
store/proto/staterpc plumbing. workflow.Forger gains Comment and
ReplyToReview, proxied to archie-agent via a new forgerpc RPC. Resume
and its Dir lookup are added as an unexported optional Trees capability
rather than widening the exported interface, which is already at
golangci-lint's interfacebloat cap and is projected into wfextract's
Yaegi symbol table for repository-authored stage plugins.

The daemon-side reaction consumer that dispatches into this workflow
(subscribing ARCHIE_REACTIONS, grouping comments by review, enforcing
per-task serialization) is tracked separately in archie-core-8li9.4.

* feat(messaging): extract standalone messaging service binary and process composition

* feat(messaging): adapt channel frontends to consume Gateway ChatContract

* feat(messaging): remove in-process channels and Router from daemon bootstrap

Telegram, email and webhook now run only in the Messaging Service, which
reaches this process through the Gateway's ChatContract. The daemon builds
no channel and no gateway.Router, so the wiring that existed only to serve
them goes with it: setupGateways, setupEmailGateway, setupWebhookGateway,
setupTelegramGateway, buildTelegramRouter, makeTelegramReload and the
release-announcement path helper.

The seams those paths owned move to where they can still be honoured:

- Channel ValidateConfig runs in archiemessaging.compose, which now returns
  an error, so a configured-but-invalid front-end fails the service rather
  than being dropped silently.
- The inbound rate limiter moves to the Gateway process, which owns the only
  Router and therefore the only place a per-sender budget applies to every
  channel's turns.
- The daemon's gateway readiness probe becomes a ContractProbe against the
  Gateway. It cannot observe a channel lifecycle it no longer runs, and a
  channel-count probe would have reported every daemon degraded.
- telegramSetup is now chatSetup (in chat_setup.go), trimmed to the fields a
  chat turn runner actually reads.

Telegram operator features that were wired only in the daemon (/update,
release announcements, version reporting, SIGHUP reload) are not yet wired
in the Messaging Service: archie-core-8cda.6.6. Channel lifecycle reporting:
archie-core-8cda.6.8. The Gateway process builds no worktree manager, so
review_pr is unadvertised on every channel: archie-core-8cda.6.7.

* feat(messaging): re-wire the Telegram operator surface in the Messaging Service

setupTelegramGateway satisfied eight telegram.Gateway seams from the daemon's
own process. Each is resolved by where the fact it needs is owned, not by
where it used to be built; the decision table is in
docs/architecture/migration-decisions.md.

Wired here: /update (built from this service's own [chat.telegram] check and
install commands), the update-report relay (same sha256(bot_user)[:8] path the
daemon wrote, so a pending report is found rather than written afresh),
show_tool_calls, and /restart's config reload, which re-resolves the token and
allowlist from this service's own file and refuses a reload that resolves to
an empty token.

/version is sourced from the Gateway: ChatSnapshot gains a Version field
carrying the Gateway's rendered build block, read per call and bounded by the
dependency timeout. cmd/archie-messaging deliberately takes no version
ldflags, so it cannot report a partial upgrade as a matched one.

RunningVersions and ReleaseAnnouncements stay nil. Both need a component's
real running version, and the only versions this process can obtain are a
different binary's stamps -- reporting them would manufacture exactly the
false success the update check exists to catch, and the announcer without
parseable versions is a no-op that reads as configured. archie-core-8cda.6.9
carries the contract that unblocks them. Dangerous stays nil: no composition
has ever supplied a command authority.

Enrich is dropped from the update service: componentInstallTypeEnricher read
[nats], which this service must not decode, so the check command's own
reported install type stands.

* fix(workflow): guard inline review comments against head drift

Two findings from the review of q9au.

A comment's line numbers describe exactly one revision of the branch, but the
forge resolved the pull request's head at posting time -- after the review had
already read the worktree. A push in that window (a collaborator, a retry) meant
the findings were anchored to whatever now occupied those lines, mislabelling
instead of failing. The revision the review measured now travels with the
comment set: OpenPR records it from the same worktree read it already performs
(TaskContext.ReviewedHeadSHA), StagePostReviewComments threads it through
Forger to forgerpc, and the forge refuses the whole set once the head it reads
has moved off it -- which the best-effort stage logs, the PR body remaining the
record. An unmeasured revision posts unverified rather than dropping every
finding, so a worktree that cannot report one degrades instead of regressing.

Suggestion was documented single-line only, because GitHub applies a suggestion
fence up to its first line and a multi-line replacement would silently corrupt
the file. Only the verdict was validated, so a multi-line suggestion reached the
renderer and posted as that truncated fence. Validate now rejects an embedded
newline, which makes record_finding refuse it with feedback the model can act on
instead of the reviewer discovering the corruption later.
2026-09-19 23:37:23 +10:00

360 lines
14 KiB
Go

// Package forgerpc lets archie-agent call the forge methods workflow stages
// invoke mid-run -- CloseIssue, CreatePR, LinkBranch, CreateReviewComments,
// Comment and ReplyToReview, the workflow.Forger set -- over core NATS
// request/reply, instead of the agent
// container holding a live forge API token. archied remains the sole holder of
// forge credentials and the sole caller of forge.Forge.
//
// The rest of forge.Forge (issue polling, invitations, reactions, PR-state
// reconciliation) is used exclusively by the daemon's own poll/reconcile
// loops, never from inside a workflow stage, so there is nothing for
// archie-agent to call.
//
// The server also still answers Comment and SetStateLabel. No current agent
// calls them; the handlers exist so an older archie-agent image keeps working
// against a newer daemon. The compatibility is one-directional: a NEW agent
// against an OLD daemon has no LinkBranch handler to reach, and the request
// times out, so an image skew that way parks tasks. Note it in release notes
// rather than assuming either side can lag.
//
// CreateReviewComments is the one method whose absence does NOT park a task: the
// stage that calls it is best-effort by design, so the same new-agent/old-daemon
// skew costs the inline comments and leaves the PR-body findings list intact.
package forgerpc
import (
"context"
"encoding/json"
"errors"
"fmt"
"log/slog"
"strings"
"time"
"github.com/nats-io/nats.go"
"github.com/samcharles93/archie-core/internal/domain/workflow"
"github.com/samcharles93/archie-core/internal/forge"
"github.com/samcharles93/archie-core/internal/natsrpc"
)
const (
SubjectComment = "archie.forge.comment"
SubjectCloseIssue = "archie.forge.close_issue"
SubjectCreatePR = "archie.forge.create_pr"
SubjectLinkBranch = "archie.forge.link_branch"
SubjectSetStateLabel = "archie.forge.set_state_label"
SubjectCreateReviewComment = "archie.forge.create_review_comments"
SubjectReplyToReview = "archie.forge.reply_to_review"
)
// SubjectFor returns the subject for base, scoped to identity when set.
// An empty identity uses the root subject, which is how single-identity
// deployments and agent images that predate identity routing behave. The
// daemon registers one server per identity on these scoped subjects so a
// container-mode task owned by a non-root identity has its RPC calls
// served by that identity's own forge client, not the root's.
func SubjectFor(identity, base string) string {
if identity == "" {
return base
}
return "archie.forge." + identity + "." + strings.TrimPrefix(base, "archie.forge.")
}
type CommentRequest struct {
Owner, Repo string
Number int
Body string
}
type CommentResponse struct {
ID int64 `json:"id"`
natsrpc.Envelope
}
type CloseIssueRequest struct {
Owner, Repo string
Number int
Comment string
}
type CreatePRRequest struct {
Owner, Repo, Title, Head, Base, Body string
}
type CreatePRResponse struct {
Number int `json:"number"`
natsrpc.Envelope
}
type LinkBranchRequest struct {
Owner, Repo string
IssueNumber int
Branch string
}
type SetStateLabelRequest struct {
Owner, Repo string
Number int
Label string
KnownLabels []string
}
// CreateReviewCommentsRequest carries one review's worth of line-anchored
// comments. The payload mirrors workflow.ReviewComment rather than reusing it:
// the wire shape is a contract between two processes that can be at different
// versions, so it is stated here instead of inherited from a type that may be
// renamed or reshaped for reasons that never crossed the wire.
type CreateReviewCommentsRequest struct {
Owner, Repo string
Number int
// ReviewedHeadSHA is the revision the comments' line numbers were measured
// on, which the worker records when it opens the pull request. It is
// additive on the wire: a worker that predates it omits the field and the
// server posts unverified (the pre-existing behaviour) rather than refusing
// every set, and a server that predates it ignores the field. Neither
// direction breaks, and neither silently mints a revision that was never
// measured.
ReviewedHeadSHA string
Comments []InlineReviewCommentPayload
}
// InlineReviewCommentPayload is one comment on the wire.
type InlineReviewCommentPayload struct {
Path string
Line int
Body string
}
// ReplyToReviewRequest carries one threaded reply to a review comment.
type ReplyToReviewRequest struct {
Owner, Repo string
Number int
CommentID int64
Body string
}
// Response is a bare success/error envelope for calls with no return value.
type Response struct {
natsrpc.Envelope
}
// Server proxies forgerpc requests to a real forge.Forge implementation.
type Server struct {
Forge forge.Forge
Log *slog.Logger
}
// Register subscribes all handlers on nc under the root (identity-less)
// subjects. RegisterFor(nc, identity) registers the same handlers under
// that identity's scoped subjects; the daemon calls it once per identity
// so container-mode tasks route to their own forge client.
func (s *Server) Register(nc *nats.Conn) (unsubscribe func(), err error) {
return s.RegisterFor(nc, "")
}
// RegisterFor registers all handlers under the subjects scoped to identity.
func (s *Server) RegisterFor(nc *nats.Conn, identity string) (unsubscribe func(), err error) {
return natsrpc.RegisterAll(nc, []natsrpc.Registration{
{Subject: SubjectFor(identity, SubjectComment), Handler: s.handleComment},
{Subject: SubjectFor(identity, SubjectCloseIssue), Handler: s.handleCloseIssue},
{Subject: SubjectFor(identity, SubjectCreatePR), Handler: s.handleCreatePR},
{Subject: SubjectFor(identity, SubjectLinkBranch), Handler: s.handleLinkBranch},
{Subject: SubjectFor(identity, SubjectSetStateLabel), Handler: s.handleSetStateLabel},
{Subject: SubjectFor(identity, SubjectCreateReviewComment), Handler: s.handleCreateReviewComments},
{Subject: SubjectFor(identity, SubjectReplyToReview), Handler: s.handleReplyToReview},
})
}
func (s *Server) handleComment(msg *nats.Msg) {
var req CommentRequest
if err := json.Unmarshal(msg.Data, &req); err != nil {
s.respond(msg, CommentResponse{Envelope: natsrpc.NewEnvelope(fmt.Errorf("decode comment request: %w", err))})
return
}
id, err := s.Forge.Comment(context.Background(), req.Owner, req.Repo, req.Number, req.Body)
s.respond(msg, CommentResponse{ID: id, Envelope: natsrpc.NewEnvelope(err)})
}
func (s *Server) handleCloseIssue(msg *nats.Msg) {
var req CloseIssueRequest
if err := json.Unmarshal(msg.Data, &req); err != nil {
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(fmt.Errorf("decode close_issue request: %w", err))})
return
}
err := s.Forge.CloseIssue(context.Background(), req.Owner, req.Repo, req.Number, req.Comment)
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(err)})
}
func (s *Server) handleCreatePR(msg *nats.Msg) {
var req CreatePRRequest
if err := json.Unmarshal(msg.Data, &req); err != nil {
s.respond(msg, CreatePRResponse{Envelope: natsrpc.NewEnvelope(fmt.Errorf("decode create_pr request: %w", err))})
return
}
num, err := s.Forge.CreatePR(context.Background(), req.Owner, req.Repo, req.Title, req.Head, req.Base, req.Body)
s.respond(msg, CreatePRResponse{Number: num, Envelope: natsrpc.NewEnvelope(err)})
}
func (s *Server) handleLinkBranch(msg *nats.Msg) {
var req LinkBranchRequest
if err := json.Unmarshal(msg.Data, &req); err != nil {
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(fmt.Errorf("decode link_branch request: %w", err))})
return
}
err := s.Forge.LinkBranch(context.Background(), req.Owner, req.Repo, req.IssueNumber, req.Branch)
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(err)})
}
func (s *Server) handleCreateReviewComments(msg *nats.Msg) {
var req CreateReviewCommentsRequest
if err := json.Unmarshal(msg.Data, &req); err != nil {
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(fmt.Errorf("decode create_review_comments request: %w", err))})
return
}
writer, ok := s.Forge.(forge.ReviewCommentWriter)
if !ok {
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(errors.New("this forge cannot post review comments"))})
return
}
comments := make([]forge.InlineReviewComment, 0, len(req.Comments))
for _, cm := range req.Comments {
comments = append(comments, forge.InlineReviewComment{Path: cm.Path, Line: cm.Line, Body: cm.Body})
}
err := writer.CreateReviewComments(context.Background(), req.Owner, req.Repo, req.Number, req.ReviewedHeadSHA, comments)
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(err)})
}
func (s *Server) handleReplyToReview(msg *nats.Msg) {
var req ReplyToReviewRequest
if err := json.Unmarshal(msg.Data, &req); err != nil {
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(fmt.Errorf("decode reply_to_review request: %w", err))})
return
}
reader, ok := s.Forge.(forge.PullRequestReviewReader)
if !ok {
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(errors.New("this forge cannot reply to reviews"))})
return
}
err := reader.ReplyToReview(context.Background(), req.Owner, req.Repo, req.Number, req.CommentID, req.Body)
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(err)})
}
func (s *Server) handleSetStateLabel(msg *nats.Msg) {
var req SetStateLabelRequest
if err := json.Unmarshal(msg.Data, &req); err != nil {
s.respond(msg, Response{Envelope: natsrpc.NewEnvelope(fmt.Errorf("decode set_state_label request: %w", err))})
return
}
s.Forge.SetStateLabel(context.Background(), req.Owner, req.Repo, req.Number, req.Label, req.KnownLabels)
s.respond(msg, Response{})
}
func (s *Server) respond(msg *nats.Msg, v any) {
natsrpc.Respond(msg, s.Log, "forgerpc", v)
}
// Client calls the forgerpc Server from archie-agent's process. It
// implements the four proxied forge.Forge methods with matching
// signatures so it can be used as a drop-in for workflow stages.
type Client struct {
Conn *nats.Conn
// Timeout bounds each call when ctx has no deadline of its own.
Timeout time.Duration
// Identity scopes this client's calls to one identity's RPC server.
// Empty uses the root subjects (single-identity deployments).
Identity string
}
func (c *Client) rpc() *natsrpc.Client { return &natsrpc.Client{Conn: c.Conn, Timeout: c.Timeout} }
func (c *Client) subject(base string) string { return SubjectFor(c.Identity, base) }
// Comment and SetStateLabel are not in workflow.Forger, so no current agent
// calls them. They stay as the client half of handlers kept for image skew
// (see the package doc) -- and as the only way to exercise those handlers
// end to end, which is why removing them would leave the compatibility path
// untested rather than merely unused.
func (c *Client) Comment(ctx context.Context, owner, repo string, number int, body string) (int64, error) {
req := CommentRequest{Owner: owner, Repo: repo, Number: number, Body: body}
resp, err := natsrpc.Call[CommentResponse](ctx, c.rpc(), c.subject(SubjectComment), req)
if err != nil {
return 0, err
}
if err := resp.Err(); err != nil {
return 0, err
}
return resp.ID, nil
}
func (c *Client) CloseIssue(ctx context.Context, owner, repo string, number int, comment string) error {
req := CloseIssueRequest{Owner: owner, Repo: repo, Number: number, Comment: comment}
resp, err := natsrpc.Call[Response](ctx, c.rpc(), c.subject(SubjectCloseIssue), req)
if err != nil {
return err
}
return resp.Err()
}
func (c *Client) CreatePR(ctx context.Context, owner, repo, title, head, base, body string) (int, error) {
req := CreatePRRequest{Owner: owner, Repo: repo, Title: title, Head: head, Base: base, Body: body}
resp, err := natsrpc.Call[CreatePRResponse](ctx, c.rpc(), c.subject(SubjectCreatePR), req)
if err != nil {
return 0, err
}
if err := resp.Err(); err != nil {
return 0, err
}
return resp.Number, nil
}
func (c *Client) LinkBranch(ctx context.Context, owner, repo string, issueNumber int, branch string) error {
req := LinkBranchRequest{Owner: owner, Repo: repo, IssueNumber: issueNumber, Branch: branch}
resp, err := natsrpc.Call[Response](ctx, c.rpc(), c.subject(SubjectLinkBranch), req)
if err != nil {
return err
}
return resp.Err()
}
// SetStateLabel remains available for RPC compatibility with older workers;
// the current workflow deliberately never calls it.
func (c *Client) SetStateLabel(ctx context.Context, owner, repo string, number int, label string, knownLabels []string) {
req := SetStateLabelRequest{Owner: owner, Repo: repo, Number: number, Label: label, KnownLabels: knownLabels}
_, _ = natsrpc.Call[Response](ctx, c.rpc(), c.subject(SubjectSetStateLabel), req)
}
// CreateReviewComments posts the review's line-anchored findings on the pull
// request. It is the one proxied method a workflow stage treats as best-effort,
// which is why it is also the one whose failure cannot park a task.
//
// reviewedHeadSHA travels with the call so the daemon -- the only side holding
// forge credentials, and so the only side that can read a pull request's head --
// can refuse a set whose line numbers describe a revision the PR has left.
func (c *Client) CreateReviewComments(ctx context.Context, owner, repo string, number int, reviewedHeadSHA string, comments []workflow.ReviewComment) error {
payload := make([]InlineReviewCommentPayload, 0, len(comments))
for _, cm := range comments {
payload = append(payload, InlineReviewCommentPayload{Path: cm.Path, Line: cm.Line, Body: cm.Body})
}
req := CreateReviewCommentsRequest{
Owner: owner, Repo: repo, Number: number,
ReviewedHeadSHA: reviewedHeadSHA, Comments: payload,
}
resp, err := natsrpc.Call[Response](ctx, c.rpc(), c.subject(SubjectCreateReviewComment), req)
if err != nil {
return err
}
return resp.Err()
}
// ReplyToReview posts a threaded reply to one review comment, on behalf of
// the remediate workflow.
func (c *Client) ReplyToReview(ctx context.Context, owner, repo string, number int, commentID int64, body string) error {
req := ReplyToReviewRequest{Owner: owner, Repo: repo, Number: number, CommentID: commentID, Body: body}
resp, err := natsrpc.Call[Response](ctx, c.rpc(), c.subject(SubjectReplyToReview), req)
if err != nil {
return err
}
return resp.Err()
}