beads/internal/httpapi/release_test.go
Julian Knutsen b858a29e4a
fix(contracts): JSON contract fixes before 1.3.0 freezes them — string revision tokens, --set-metadata scalar typing, create --json nano timestamps (#6053)
* fix(metadata): restore v1.2.2 scalar typing for --set-metadata values

8ebd53e88 moved the CLI's metadata-edit helpers into
internal/storage/metadata.go and, in the move, replaced v1.2.2's
toJSONValue inference with a MetadataEditValue that always json.Marshals
the value as a string. The intent was GH#4146 (a numeric-looking id or a
version like "1e3" came back as a JSON number, breaking map[string]string
Go consumers), but it silently changed the shipped CLI contract:
`bd update --set-metadata k=5` started storing "5" instead of 5. That
regression shipped only in the deprecated/superseded v1.2.1, never in
v1.2.2, and 1.3.0 would freeze it.

Restore v1.2.2's inference body-for-body inside MetadataEditValue, so the
parity claim stays checkable against `git show v1.2.2:cmd/bd/update.go`:
"null" is JSON null, "true"/"false" are booleans, a spelling that scans
as a float AND is valid JSON on its own is emitted as that literal
(5 -> 5, 1e3 -> 1e3, 0.50 -> 0.50, while 05, +5 and NaN stay strings),
and everything else is a string. Every --set-metadata caller funnels
through this one function: embedded (cmd/bd/update.go ->
storage.ApplyMetadataEdits), proxied (parseSetMetadataFlags) and the
update executor's []string path.

GH#4146's concern is served by the explicitly typed routes, which are
untouched: --metadata / --metadata-json on the CLI (MergeMetadataJSON)
and the typed set map on the HTTP/Go path (applyTypedMetadataEdits).

No data migration: a "5" written by an in-window build is
indistinguishable from a string a user set intentionally through the
typed paths, so normalizing stored metadata would corrupt intended data.
Re-running the same --set-metadata k=5 retypes any given key.

Tests: new table in internal/storage/metadata_edit_value_test.go pinning
the v1.2.2 matrix plus an ApplyMetadataEdits end-to-end; the tests that
pinned the string-always behavior are flipped (domain/db mergeops suite,
update_lifecycle patch mapping, cmd/bd toJSONValue table) and the
embedded + proxied CLI suites now assert a stored JSON number. The
contract corpus is regenerated: its update/list/ready goldens run
`--set-metadata phase=2`, so `"phase": "2"` becomes `"phase": 2` and
nothing else moves.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(create): echo full-precision timestamps from bd create, not the DATETIME(0) re-read

b92442d1a routed `bd create` through issueops.Lifecycle, whose executors
build their result by re-reading the row they just wrote
(HydrateIssueOperationResult -> GetIssueInTx). created_at/updated_at are
DATETIME(0) columns, so that snapshot comes back at whole-second
precision and cmd/bd/create.go echoes it. v1.2.2 returned the caller's
own in-memory struct, which PrepareIssueForInsert stamps with
time.Now().UTC() including nanoseconds, so `bd create --json` used to
print e.g. 2026-08-28T12:34:56.789012345Z. Agents that create several
issues in a burst order them by exactly that fraction; rounded to the
second they all tie.

Restore the echo with OverlayCreateTimestamps: after hydration, put back
the created_at/updated_at (and closed_at, when the write set one) the
INSERT actually sent. Everything else on the snapshot stays DB-truth,
which is the facade's contract — only create, and only these fields.

Applied at every create body, not just the two the CLI reaches:
issueops.ExecuteCreate and ExecuteCreateBatch behind the dolt and
embedded-dolt stores, and the unit-of-work provider's own Create,
CreateBatch and batch-apply create item, which is what the proxied server
and POST /v0/beads/issues run on. Fixing only the shared executors would
have left the two bodies disagreeing about the same promise.

What is NOT changed, because it is not a regression: `bd show --json` and
`bd list --json` answered whole seconds in v1.2.2 too — they read the
column — and update/close/reopen re-read after writing there as well.
Widening the columns to DATETIME(6) would give reads sub-second times
too, but that is a schema migration on user Dolt DBs and touches paths
that deliberately compare at second precision; it is a 1.4 change.

The accepted skew, which predates this and stays: the engine ROUNDS to
the column rather than truncating, so a later read of the same row can
sit up to half a second either side of the create echo.

Tests: two new conformance cases, run on all three backends —
RunLifecycleCreateEchoesSubSecondTimestamps (a caller-supplied
nanosecond stamp must survive to the result exactly; the auto-stamped arm
stays a bounded smoke check, since requiring a non-zero fraction off the
clock is flaky by a billionth) and RunBatchCreatorEchoesSubSecondTimestamps
(per item, since both bodies build the result one snapshot at a time).
Plus a cmd/bd embedded test asserting `bd create --json` prints a
fractional part, and one httpapi assertion that the wire does not round
the role's snapshot on the way out. All four fail on the pre-fix tree.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(httpapi,types)!: revision / expected_version tokens are opaque JSON strings

The optimistic-concurrency token is minted by freshRowLock over the FULL
int64 range (crypto-random, non-zero) and published on the wire as a JSON
number: `revision` on responses, `expected_version` on guarded requests.
A JSON number does not survive a consumer that decodes into an IEEE-754
double — JavaScript's JSON.parse, Go's `any`, Python's `float`, jq. Live
tokens run past 5e17, where a double's ulp is already 64, so such a
client reads back a value NEAR the token, echoes that as its guard, and
earns a 409 against a row nobody else touched. The spec already carried
"DECODE IT AS A 64-BIT INTEGER" paragraphs telling clients to work around
this; that is a workaround for a wire type that should not have been a
number.

Emit and accept the token as a decimal STRING instead. It is documented
as opaque and EQUALITY-ONLY — compare it, never order or interpret it —
and a string round-trips exactly in every JSON consumer. It also leaves a
future backend free to mint a token that is not an int64 at all.

Internals are untouched: freshRowLock, the row_lock BIGINT column, its
migration, types.Issue.RowVersion and the ExpectedVersion *int64 guards
on the issueops requests all stay int64. Only the wire projection moves,
through the one spelling of the encoding — types.RevisionToken and
types.ParseRevisionToken.

STRICT STRING ONLY, with no transitional number-or-string acceptance.
This API already refuses unknown members by name and keeps one spelling
of the guard, and a number is precisely the shape a double-based client
has already rounded before it left the client — accepting it would
preserve the defect. A numeric guard is a 400 naming the member.

Chosen over clamping freshRowLock to 53 bits: a clamp does not fix rows
already minted with full-range tokens by in-window builds, so a JS
consumer reading one still rounds past 2^53 — repairing that would need a
row_lock normalization sweep over issues and wisps, a migration days
before a tag.

Surface: types.IssueDetails.Revision (what `bd show --json` prints and
GET /v0/beads/issues/{id} returns), the five hand-assigned httpapi
response fields, applyVersionGuardMember (now the reader for the batch
update item too, ahead of its typed decode, so a mistyped token still
earns a 400 naming the member), Problem.expected_version, and 14 OpenAPI
members regenerated through `make api-gen`. actual_version goes with them
— it is documented as "the `revision` the row was found holding", so
leaving it an integer would make the expected/actual pair disagree about
its own type; it is declared but never populated, so nothing changes
behaviorally.

Nothing shipped in a tagged release ever carried this member: v1.2.2 has
no internal/httpapi at all and no `revision` anywhere on the wire, so the
whole surface is in-window and can still be changed compatibly. Legacy
migration-0054 rows read as the string "0" with the same CAS semantics.

Corpus goldens are UNCHANGED: the canonicalizer already replaced any
`revision` value with the string "<REVISION>" whatever its JSON type.

Tests: types-level pins that the marshaled member matches `"revision":
"-?\d+"` and that RevisionToken/ParseRevisionToken round-trip MinInt64,
MaxInt64 and 2^53+1, plus a JS-realistic decode of 2^53+1 with a float64
control proving the token really is one a number would corrupt; httpapi
refusal tests for a numeric guard on update, close, reopen, delete and
both batch item kinds; a 409 echo asserted as the string; and the proxied
serve integration tests now splice the response's own token back in
verbatim rather than re-rendering a parsed one.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-08-29 05:20:27 +00:00

480 lines
21 KiB
Go

package httpapi
import (
"errors"
"fmt"
"net/http"
"strings"
"testing"
"github.com/steveyegge/beads/internal/types"
"github.com/steveyegge/beads/issueops"
)
// These are pure: the release path runs end to end over a real listener against
// a fake ROLE, so the wire edge — the path split, the media type, the body
// vocabulary, the guard pair, the response shape and every problem document —
// is covered on every pull request by the unconditional Go test job.
//
// What a fake cannot prove is that a REFUSED release left the claim standing,
// and that is the property the whole non-idempotence argument rests on. That
// lives in cmd/bd's proxied-server integration test against real Dolt
// (TestProxiedServerServeRelease). The role-level transition — what the row
// looks like afterwards, the lease, the event — is owned against a real store
// by internal/storage/uow/releaser_contract_test.go, which this slice cites
// rather than duplicates.
const releasePath = "/v0/beads/issues/bd-1:release"
func (ts *testServer) releaseIssue(t *testing.T, path, body string) *http.Response {
t.Helper()
return ts.postBody(t, path, "application/json", body)
}
// newReleaseServer wires a server over the store-shaped source with a releaser
// the case controls. Every other role is a placeholder: Listen refuses a
// partial source, and a release reaches none of them.
func newReleaseServer(t *testing.T, releaser *roleReleaser) *testServer {
t.Helper()
return newTestServer(t, rolesConfig(Config{Releaser: releaser}))
}
// releasedIssue is the ANONYMOUS post-state: assignee cleared, status open. It
// is the same row no matter who emptied it, which is the whole reason this
// operation refuses rather than answering an idempotent 200.
func releasedIssue(id string, revision int64) *types.Issue {
issue := seededIssue(id, "", types.StatusOpen)
issue.RowVersion = revision
return issue
}
// TestReleaseWritesOnceAndAnswersWithTheRowItWrote is the happy path and the
// three things a client depends on: the role receives exactly what the body
// asked for, the response carries the row as it stands after the release, and
// `revision` carries the REMINTED token off that same row.
func TestReleaseWritesOnceAndAnswersWithTheRowItWrote(t *testing.T) {
releaser := &roleReleaser{result: issueops.ReleaseResult{Issue: releasedIssue("bd-1", 77), Changed: true}}
ts := newReleaseServer(t, releaser)
resp := ts.releaseIssue(t, releasePath, `{"actor":"alice"}`)
if resp.StatusCode != http.StatusOK {
t.Fatalf("status = %d, want 200: %s", resp.StatusCode, readAll(t, resp))
}
got := releaser.releaseRequests()
if len(got) != 1 {
t.Fatalf("the role received %+v, want exactly one request", got)
}
// The unconditional path: nil selects it, and nothing else does.
if got[0].Actor != "alice" || got[0].IssueID != "bd-1" || got[0].ExpectedAssignee != nil || got[0].Force {
t.Fatalf("the role received %+v, want an unconditional release of bd-1 by alice", got[0])
}
body := decodeBody(t, resp)
if body["changed"] != true {
t.Errorf("changed = %v, want true — the role reports it true on every answer it returns without an error", body["changed"])
}
if body["revision"] != "77" {
t.Errorf("revision = %v, want the post-release token off the row the role answered with, as a decimal string", body["revision"])
}
issue, ok := body["issue"].(map[string]any)
if !ok {
t.Fatalf("issue = %#v, want an object", body["issue"])
}
// The ANONYMOUS post-state, as the wire spells it: status open, and no
// `assignee` at all — types.Issue omits the member when it is empty, so a
// client reads the release's outcome from the member's ABSENCE.
if issue["status"] != string(types.StatusOpen) {
t.Errorf("status = %v, want %s: %v", issue["status"], types.StatusOpen, issue)
}
if assignee, present := issue["assignee"]; present && assignee != "" {
t.Errorf("assignee = %v; a released row holds no claim", assignee)
}
}
// TestReleaseHasNoIdempotentAnswer is the operation's defining property, tested
// from the wire rather than assumed from the role: an unheld row is a REFUSAL,
// and it must not arrive as a 200 carrying `changed: false`.
//
// The distinction is not stylistic. A 200 here would report one answer for "I
// released this twice", "a reaper beat me to it" and "nothing ever claimed it",
// because the post-state is identical in all three and there is nothing left on
// the row to tell them apart. It would also be a value no role produces:
// ReleaseResult.Changed is true on every non-error answer.
func TestReleaseHasNoIdempotentAnswer(t *testing.T) {
releaser := &roleReleaser{err: fmt.Errorf("%w: bd-1 has no assignee to release", issueops.ErrNotClaimed)}
ts := newReleaseServer(t, releaser)
resp := ts.releaseIssue(t, releasePath, `{"actor":"alice"}`)
if resp.StatusCode != http.StatusConflict {
t.Fatalf("status = %d, want 409: %s", resp.StatusCode, readAll(t, resp))
}
body := decodeBody(t, resp)
if body["code"] != string(CodeNotReleasable) {
t.Fatalf("code = %v, want %s", body["code"], CodeNotReleasable)
}
// No member distinguishes the two conditions this code covers, and a client
// must not learn to dispatch on one appearing.
for _, member := range []string{"assignee", "issue_status", "expected_assignee", "actual_assignee"} {
if _, present := body[member]; present {
t.Errorf("not_releasable carries %q; the code documents no discriminating member", member)
}
}
}
// TestReleaseAnswersTheRefusalsTheRoleRaises walks the whole 409 vocabulary and
// the 404, each from the sentinel the role actually returns.
//
// EVERY ARM IS MATCHED BEFORE ErrValidation AND ErrNotFound in failRelease, and
// that is what this table exists to pin: none of these sentinels is wrapped in
// ErrValidation by any leg, so an arm placed below a generic one would be a 500
// on every leg for a condition the document names by code. Deleting any single
// case from failRelease must fail a row here.
func TestReleaseAnswersTheRefusalsTheRoleRaises(t *testing.T) {
for _, tc := range []struct {
name string
body string
err error
status int
code Code
param string
members map[string]any
}{
{
name: "a status that does not accept a release",
body: `{"actor":"alice"}`,
err: fmt.Errorf("%w: bd-1 has status %q, which is neither %q nor %q", issueops.ErrNotReleasable, "closed", "open", "in_progress"),
status: http.StatusConflict,
code: CodeNotReleasable,
},
{
name: "a row that holds no claim",
body: `{"actor":"alice"}`,
err: fmt.Errorf("%w: bd-1 has no assignee to release", issueops.ErrNotClaimed),
status: http.StatusConflict,
code: CodeNotReleasable,
},
{
// The fence, answered with updateIssue's code for the identical
// situation. No `assignee` member: the fence refuses without naming
// the holder, so absence means "re-read the row".
name: "an unforced release by an actor that is not the holder",
body: `{"actor":"mallory"}`,
err: fmt.Errorf("%w: bd-1 is held by alice", issueops.ErrNotOwner),
status: http.StatusConflict,
code: CodeAlreadyClaimed,
},
{
// The guard, carrying the REQUEST's expectation and no observation.
name: "a guard that missed",
body: `{"actor":"supervisor","expected_assignee":"alice"}`,
err: fmt.Errorf("%w: bd-1 is held by %q, expected %q", issueops.ErrAssigneeMismatch, "bob", "alice"),
status: http.StatusConflict,
code: CodePreconditionFailed,
param: releaseExpectedAssigneeMember,
members: map[string]any{"expected_assignee": "alice"},
},
{
name: "an id naming no row on either plane",
body: `{"actor":"alice"}`,
err: fmt.Errorf("get bd-1: %w", issueops.ErrNotFound),
status: http.StatusNotFound,
code: CodeNotFound,
},
{
// Defensive: every validation refusal the role has is refused at the
// edge. It must still be a 400 rather than a 500 if that stops
// being true.
name: "the role's own validation, which the edge should have caught",
body: `{"actor":"alice"}`,
err: fmt.Errorf("%w: release requires an actor", issueops.ErrValidation),
status: http.StatusBadRequest,
code: CodeInvalidArgument,
},
} {
t.Run(tc.name, func(t *testing.T) {
ts := newReleaseServer(t, &roleReleaser{err: tc.err})
resp := ts.releaseIssue(t, releasePath, tc.body)
if resp.StatusCode != tc.status {
t.Fatalf("status = %d, want %d: %s", resp.StatusCode, tc.status, readAll(t, resp))
}
body := decodeBody(t, resp)
if body["code"] != string(tc.code) {
t.Fatalf("code = %v, want %s", body["code"], tc.code)
}
if tc.param != "" && body["param"] != tc.param {
t.Errorf("param = %v, want %q", body["param"], tc.param)
}
for member, want := range tc.members {
if body[member] != want {
t.Errorf("%s = %v, want %v", member, body[member], want)
}
}
// The refusals replace the role's prose with the server's own
// words, so the role's message must not travel on the wire: it
// names holders and statuses this surface publishes typed or not
// at all.
if detail, _ := body["detail"].(string); detail != "" && tc.err != nil {
if detail == tc.err.Error() {
t.Errorf("detail quotes the role's message verbatim: %q", detail)
}
}
// `actual_assignee` is declared on the envelope for an operation
// whose role can report what it found. This one cannot.
if _, present := body["actual_assignee"]; present {
t.Errorf("the refusal carries actual_assignee, which no release refusal can honestly fill")
}
})
}
}
// TestReleaseRefusesTheGuardPair pins the one refusal this handler owns rather
// than forwards. `force` and `expected_assignee` are answers to the same
// question and they disagree, so honoring either would be the server picking
// which half of the request the caller meant.
//
// It is refused at the EDGE, before any database work, which is what lets the
// 400 name a member.
func TestReleaseRefusesTheGuardPair(t *testing.T) {
releaser := &roleReleaser{result: issueops.ReleaseResult{Issue: releasedIssue("bd-1", 1), Changed: true}}
ts := newReleaseServer(t, releaser)
resp := ts.releaseIssue(t, releasePath, `{"actor":"alice","expected_assignee":"bob","force":true}`)
if resp.StatusCode != http.StatusBadRequest {
t.Fatalf("status = %d, want 400: %s", resp.StatusCode, readAll(t, resp))
}
body := decodeBody(t, resp)
if body["code"] != string(CodeInvalidArgument) || body["param"] != releaseForceMember {
t.Errorf("code/param = %v/%v, want %s on `%s`", body["code"], body["param"], CodeInvalidArgument, releaseForceMember)
}
if got := releaser.releaseRequests(); len(got) != 0 {
t.Errorf("the pair reached the role: %+v", got)
}
}
// TestReleaseExpectedAssigneeDistinguishesAbsentFromEmpty is the member's whole
// contract, and the one place it disagrees with updateIssue's same-named guard.
//
// Absent selects the unconditional path. The empty string — and anything blank
// after trimming — is a 400 rather than a guard meaning "expected unassigned",
// because "release a row nobody holds" describes no release at all. A caller
// asking whether a row is unheld is asking a READER a question.
func TestReleaseExpectedAssigneeDistinguishesAbsentFromEmpty(t *testing.T) {
for _, tc := range []struct {
name string
body string
status int
expected *string
}{
{name: "absent selects the unconditional path", body: `{"actor":"alice"}`, status: http.StatusOK},
{name: "the empty string is refused", body: `{"actor":"alice","expected_assignee":""}`, status: http.StatusBadRequest},
{name: "blank after trimming is refused", body: `{"actor":"alice","expected_assignee":" "}`, status: http.StatusBadRequest},
{name: "explicit null is refused", body: `{"actor":"alice","expected_assignee":null}`, status: http.StatusBadRequest},
{name: "a non-string is refused", body: `{"actor":"alice","expected_assignee":7}`, status: http.StatusBadRequest},
{
name: "a real holder guards the release",
body: `{"actor":"supervisor","expected_assignee":"alice"}`,
status: http.StatusOK,
expected: strPtr("alice"),
},
{
// The value travels UNTRIMMED, matching the role: a padded
// expectation must lose every time rather than intermittently.
name: "a padded holder reaches the role padded",
body: `{"actor":"supervisor","expected_assignee":" alice"}`,
status: http.StatusOK,
expected: strPtr(" alice"),
},
} {
t.Run(tc.name, func(t *testing.T) {
releaser := &roleReleaser{result: issueops.ReleaseResult{Issue: releasedIssue("bd-1", 1), Changed: true}}
ts := newReleaseServer(t, releaser)
resp := ts.releaseIssue(t, releasePath, tc.body)
if resp.StatusCode != tc.status {
t.Fatalf("status = %d, want %d: %s", resp.StatusCode, tc.status, readAll(t, resp))
}
if tc.status != http.StatusOK {
if body := decodeBody(t, resp); body["param"] != releaseExpectedAssigneeMember {
t.Errorf("param = %v, want %q", body["param"], releaseExpectedAssigneeMember)
}
if got := releaser.releaseRequests(); len(got) != 0 {
t.Errorf("a refused guard reached the role: %+v", got)
}
return
}
got := releaser.releaseRequests()
if len(got) != 1 {
t.Fatalf("the role received %+v, want exactly one request", got)
}
switch {
case tc.expected == nil && got[0].ExpectedAssignee != nil:
t.Errorf("ExpectedAssignee = %q, want nil — only absence selects the unconditional path", *got[0].ExpectedAssignee)
case tc.expected != nil && got[0].ExpectedAssignee == nil:
t.Errorf("ExpectedAssignee = nil, want %q", *tc.expected)
case tc.expected != nil && *got[0].ExpectedAssignee != *tc.expected:
t.Errorf("ExpectedAssignee = %q, want %q", *got[0].ExpectedAssignee, *tc.expected)
}
})
}
}
// TestReleaseForwardsForce: the fence is the ROLE's, so the handler's whole job
// is to pass the bypass through unexamined.
func TestReleaseForwardsForce(t *testing.T) {
releaser := &roleReleaser{result: issueops.ReleaseResult{Issue: releasedIssue("bd-1", 2), Changed: true}}
ts := newReleaseServer(t, releaser)
resp := ts.releaseIssue(t, releasePath, `{"actor":"supervisor","force":true}`)
if resp.StatusCode != http.StatusOK {
t.Fatalf("status = %d, want 200: %s", resp.StatusCode, readAll(t, resp))
}
got := releaser.releaseRequests()
if len(got) != 1 || !got[0].Force {
t.Fatalf("the role received %+v, want one forced release", got)
}
}
// TestReleaseRefusesUnknownAndMalformedBodies is the body vocabulary, refused BY
// NAME. The schema is additionalProperties: false, and a client that has stopped
// parsing prose can only enforce that if the server names the offender.
func TestReleaseRefusesUnknownAndMalformedBodies(t *testing.T) {
for _, tc := range []struct {
name string
body string
param string
}{
{name: "an unknown member", body: `{"actor":"alice","expected_version":"3"}`, param: "expected_version"},
{name: "a missing actor", body: `{"force":true}`, param: claimActorMember},
{name: "an actor that is blank after trimming", body: `{"actor":" "}`, param: claimActorMember},
{name: "an actor carrying a newline", body: `{"actor":"alice\nbd: released by mallory"}`, param: claimActorMember},
{name: "a non-boolean force", body: `{"actor":"alice","force":"yes"}`, param: releaseForceMember},
{name: "an explicitly null force", body: `{"actor":"alice","force":null}`, param: releaseForceMember},
} {
t.Run(tc.name, func(t *testing.T) {
releaser := &roleReleaser{}
ts := newReleaseServer(t, releaser)
resp := ts.releaseIssue(t, releasePath, tc.body)
if resp.StatusCode != http.StatusBadRequest {
t.Fatalf("status = %d, want 400: %s", resp.StatusCode, readAll(t, resp))
}
body := decodeBody(t, resp)
if body["code"] != string(CodeInvalidArgument) {
t.Errorf("code = %v, want %s", body["code"], CodeInvalidArgument)
}
if body["param"] != tc.param {
t.Errorf("param = %v, want %q", body["param"], tc.param)
}
if got := releaser.releaseRequests(); len(got) != 0 {
t.Errorf("a refused body reached the role: %+v", got)
}
})
}
}
// TestReleaseRefusesAReleaserThatAnswersWithNothing pins checkedReleaser. The
// handler dereferences the result's issue and reads its RowVersion, so a
// caller-supplied role reporting success without a row is a nil dereference on
// a live server — and there is no wire code that honestly describes it.
func TestReleaseRefusesAReleaserThatAnswersWithNothing(t *testing.T) {
ts := newReleaseServer(t, &roleReleaser{result: issueops.ReleaseResult{Changed: true}})
resp := ts.releaseIssue(t, releasePath, `{"actor":"alice"}`)
if resp.StatusCode != http.StatusInternalServerError {
t.Fatalf("status = %d, want 500: %s", resp.StatusCode, readAll(t, resp))
}
if body := decodeBody(t, resp); body["code"] != string(CodeInternal) {
t.Errorf("code = %v, want %s", body["code"], CodeInternal)
}
// THE STATUS ALONE PROVES NOTHING, which is why this asserts the log. The
// server recovers panics into the same 500 with the same code, so a test
// that stopped at the status would pass with the wrapper deleted and the
// handler dereferencing nil on a live server. What the wrapper buys is the
// fault ARRIVING AS AN ERROR NAMING ITSELF, which is exactly what the panic
// it replaces did not produce.
line := findLogLine(t, ts.stderr.String(), "the releaser reported success without an issue")
if !strings.Contains(line, "bd-1") {
t.Errorf("the logged fault does not name the issue it was asked about:\n%s", line)
}
}
// TestReleasePathReachesItsHandler drives the path the DOCUMENT spells, which
// is the one thing route parity cannot check for a custom-method row: it
// declares its spec path instead of deriving it from the pattern. The parity
// test bounds the shape of that exception; only a request proves the pattern
// serves the documented path — and, on a shared dispatcher, that this verb is
// not answered by whichever row happens to be first.
func TestReleasePathReachesItsHandler(t *testing.T) {
releaser := &roleReleaser{result: issueops.ReleaseResult{Issue: releasedIssue("bd-1", 1), Changed: true}}
ts := newReleaseServer(t, releaser)
resp := ts.releaseIssue(t, releasePath, `{"actor":"alice"}`)
if resp.StatusCode != http.StatusOK {
t.Fatalf("POST the documented release path: status = %d, want 200: %s", resp.StatusCode, readAll(t, resp))
}
line := findLogLine(t, ts.stderr.String(), "path="+releasePath)
if !strings.Contains(line, "op="+OpReleaseIssue) {
t.Errorf("the documented release path is served by another operation:\n%s", line)
}
}
// TestReleaseIsNotReachableThroughTheClaimsErrorVocabulary guards the reuse
// decision from the direction that would make it wrong. `already_claimed` is
// borrowed for the ownership fence; the claim's OTHER conflict, not_claimable,
// is deliberately absent from this operation, because an open unassigned row is
// the most claimable row a workspace has and answering that about a release
// refusal would send a client somewhere there is nothing to find.
func TestReleaseIsNotReachableThroughTheClaimsErrorVocabulary(t *testing.T) {
for _, code := range operationCodes[OpReleaseIssue] {
if code == CodeNotClaimable {
t.Fatalf("releaseIssue documents %s; see CodeNotReleasable for why that reuse was refused", code)
}
}
// And the borrowed one is really borrowed rather than re-minted.
if !containsCode(operationCodes[OpReleaseIssue], CodeAlreadyClaimed) {
t.Errorf("releaseIssue does not document %s; the ownership fence has no other code", CodeAlreadyClaimed)
}
}
func containsCode(codes []Code, want Code) bool {
for _, code := range codes {
if code == want {
return true
}
}
return false
}
func strPtr(s string) *string { return &s }
// releaseSentinels is the set failRelease must classify, asserted as a set so a
// sentinel added to the role without a wire arm fails here rather than reaching
// a client as a 500.
func TestEveryReleaseSentinelIsClassified(t *testing.T) {
for _, sentinel := range []error{
issueops.ErrNotClaimed,
issueops.ErrNotReleasable,
issueops.ErrNotOwner,
issueops.ErrAssigneeMismatch,
issueops.ErrNotFound,
} {
t.Run(sentinel.Error(), func(t *testing.T) {
ts := newReleaseServer(t, &roleReleaser{err: fmt.Errorf("release bd-1: %w", sentinel)})
resp := ts.releaseIssue(t, releasePath, `{"actor":"alice","expected_assignee":"bob"}`)
if resp.StatusCode >= http.StatusInternalServerError {
t.Fatalf("%v reached the client as a %d; it must be classified in failRelease: %s",
sentinel, resp.StatusCode, readAll(t, resp))
}
code := Code(decodeBody(t, resp)["code"].(string))
if !containsCode(operationCodes[OpReleaseIssue], code) {
t.Errorf("%v produced %s, which this operation does not document", sentinel, code)
}
if errors.Is(sentinel, issueops.ErrValidation) {
t.Fatalf("%v is wrapped in ErrValidation; failRelease's ordering assumes it is not", sentinel)
}
})
}
}