feat!: add cache service v2, add toolkit patches (#1110)

Serves `github.actions.results.api.v1.CacheService` next to the v1 cache API, sharing its store, plus the subset of the Azure blob protocol the toolkit uploads with. On by default via `cache.v2`, and works with `external_server`.

Clients reach it through two edits in the action's own bundle: the GHES check is opened, and the cache service URL is taken from `ACTIONS_CACHE_URL`.

The same GHES check is what makes the stock `actions/upload-artifact` and `download-artifact` abort on Gitea. Opening it makes them work without the `gitea-upload-artifact` fork, from `upload-artifact@v4.4.0` on.

Verified against 118 real bundles, every major version of 16 actions: 92 patched, the rest deliberately left alone, and every patched bundle checked with `node --check`. Also end to end against pinned `actions/cache@v6.1.0` with an unreachable results URL, so only the patch can make the cache work.

---------

Co-authored-by: silverwind <me@silverwind.io>
Co-authored-by: bircni <me@silverwind.io>
Reviewed-on: https://gitea.com/gitea/runner/pulls/1110
Reviewed-by: silverwind <2021+silverwind@noreply.gitea.com>
This commit is contained in:
bircni
2026-07-31 12:08:44 +00:00
parent 2398d4a527
commit 47d5b5ad03
18 changed files with 1585 additions and 75 deletions

View File

@@ -242,6 +242,17 @@ A password in a proxy URL is hidden in job logs. Any step can still read it, bec
Each runner starts its own cache server automatically. Cache entries are local to that runner — runners do not share a cache by default.
**Cache service v2**
`actions/cache@v4.2` and later can use the *cache service v2* API. The runner serves it from the same store as v1, on by default, and it works with `external_server`. Turn it off with:
```yaml
cache:
v2: false
```
Those actions refuse any host they do not take for GitHub. Rather than misreport the server URL, the runner edits that check out of the action's own bundle and keeps the untouched copy beside it; a bundle it does not recognise is left alone and keeps to v1. The same edit lets the stock `actions/upload-artifact` and `actions/download-artifact` work from `v4.4.0` on, without the `gitea-upload-artifact` fork.
**Shared cache across multiple runners**
Run one dedicated `gitea-runner cache-server` that all runners point at.

View File

@@ -158,12 +158,13 @@ func StartHandler(dir, outboundIP string, port uint16, internalSecret string, lo
router.POST(apiPath+"/clean", h.bearerAuth(h.clean))
// Artifact GET is signed via query-string HMAC because @actions/cache
// does not attach Authorization when downloading archiveLocation.
router.GET(apiPath+"/artifacts/:id", h.signedURLAuth(h.get))
router.GET(apiPath+"/artifacts/:id", h.signedAuth("", h.get))
// Control-plane: a remote runner registers/revokes per-job tokens so the
// cache API can authenticate them. Always wired so the routes exist; the
// handlers themselves 401 when internalSecret is unset.
router.POST(internalPath+"/register", h.internalAuth(h.internalRegister))
router.POST(internalPath+"/revoke", h.internalAuth(h.internalRevoke))
h.registerV2Routes(router)
h.router = router
@@ -339,7 +340,7 @@ func (h *Handler) find(w http.ResponseWriter, r *http.Request, _ httprouter.Para
}
defer db.Close()
cache, err := findCache(db, cred.Repo, keys, version)
cache, err := h.lookupCache(db, cred.Repo, keys, version)
if err != nil {
h.responseJSON(w, r, 500, err)
return
@@ -348,15 +349,6 @@ func (h *Handler) find(w http.ResponseWriter, r *http.Request, _ httprouter.Para
h.responseJSON(w, r, 204)
return
}
if ok, err := h.storage.Exist(cache.ID); err != nil {
h.responseJSON(w, r, 500, err)
return
} else if !ok {
_ = db.Delete(cache.ID, cache)
h.responseJSON(w, r, 204)
return
}
h.responseJSON(w, r, 200, map[string]any{
"result": "hit",
"archiveLocation": h.signedArtifactURL(cache.ID, time.Now().Add(artifactURLTTL)),
@@ -364,6 +356,25 @@ func (h *Handler) find(w http.ResponseWriter, r *http.Request, _ httprouter.Para
})
}
// lookupCache returns the entry to restore for these keys, or (nil, nil) when there is none:
// either nothing matched, or the match had lost its blob to a prune, in which case the dangling
// entry is dropped on the way out.
func (h *Handler) lookupCache(db *bolthold.Store, repo string, keys []string, version string) (*Cache, error) {
cache, err := findCache(db, repo, keys, version)
if err != nil || cache == nil {
return nil, err
}
ok, err := h.storage.Exist(cache.ID)
if err != nil {
return nil, err
}
if !ok {
_ = db.Delete(cache.ID, cache)
return nil, nil //nolint:nilnil // absence is not an error here
}
return cache, nil
}
// POST /_apis/artifactcache/caches
func (h *Handler) reserve(w http.ResponseWriter, r *http.Request, _ httprouter.Params) {
cred := credFromContext(r.Context())
@@ -438,7 +449,7 @@ func (h *Handler) upload(w http.ResponseWriter, r *http.Request, params httprout
h.responseJSON(w, r, 500, err)
return
}
h.useCache(id)
_ = h.touchCache(uint64(id), false)
h.responseJSON(w, r, 200)
}
@@ -479,23 +490,7 @@ func (h *Handler) commit(w http.ResponseWriter, r *http.Request, params httprout
db.Close()
size, err := h.storage.Commit(cache.ID, cache.Size)
if err != nil {
h.responseJSON(w, r, 500, err)
return
}
// write real size back to cache, it may be different from the current value when the request doesn't specify it.
cache.Size = size
db, err = h.openDB()
if err != nil {
h.responseJSON(w, r, 500, err)
return
}
defer db.Close()
cache.Complete = true
if err := db.Update(cache.ID, cache); err != nil {
if err := h.commitCache(cache); err != nil {
h.responseJSON(w, r, 500, err)
return
}
@@ -503,8 +498,28 @@ func (h *Handler) commit(w http.ResponseWriter, r *http.Request, params httprout
h.responseJSON(w, r, 200)
}
// commitCache assembles the uploaded parts and marks the entry complete. The caller must
// have closed its store first: Commit concatenates the whole archive and would otherwise
// hold bolt's exclusive file lock for the duration.
func (h *Handler) commitCache(cache *Cache) error {
written, err := h.storage.Commit(cache.ID, cache.Size)
if err != nil {
return err
}
// write real size back to cache, it may be different from the current value when the request doesn't specify it.
cache.Size = written
cache.Complete = true
db, err := h.openDB()
if err != nil {
return err
}
defer db.Close()
return db.Update(cache.ID, cache)
}
// GET /_apis/artifactcache/artifacts/:id
// Authenticated via signed URL (see signedURLAuth), not bearer, because the
// Authenticated via signed URL (see signedAuth), not bearer, because the
// @actions/cache toolkit downloads archiveLocation without Authorization.
// Repository scoping is already enforced at find() time; the signature binds
// the URL to the specific cache ID and an expiry.
@@ -514,7 +529,7 @@ func (h *Handler) get(w http.ResponseWriter, r *http.Request, params httprouter.
h.responseJSON(w, r, 400, err)
return
}
h.useCache(id)
_ = h.touchCache(uint64(id), false)
h.storage.Serve(w, r, uint64(id))
}
@@ -548,7 +563,9 @@ func (h *Handler) bearerAuth(handler httprouter.Handle) httprouter.Handle {
}
}
func (h *Handler) signedURLAuth(handler httprouter.Handle) httprouter.Handle {
// signedAuth authenticates a signed URL. purpose separates the flavours of URL the
// handler hands out, so one cannot be replayed as another; see computeSignature.
func (h *Handler) signedAuth(purpose string, handler httprouter.Handle) httprouter.Handle {
return func(w http.ResponseWriter, r *http.Request, params httprouter.Params) {
h.logger.Debugf("%s %s", r.Method, r.URL.Path)
id, err := strconv.ParseInt(params.ByName("id"), 10, 64)
@@ -571,7 +588,7 @@ func (h *Handler) signedURLAuth(handler httprouter.Handle) httprouter.Handle {
h.responseJSON(w, r, http.StatusUnauthorized, errors.New("signature expired"))
return
}
expected := h.computeSignature(id, exp)
expected := h.computeSignature(purpose, id, exp)
if !hmac.Equal([]byte(sig), []byte(expected)) {
h.responseJSON(w, r, http.StatusUnauthorized, errors.New("bad signature"))
return
@@ -655,19 +672,26 @@ func credFromContext(ctx context.Context) JobCredential {
return JobCredential{}
}
func (h *Handler) computeSignature(cacheID, exp int64) string {
// computeSignature signs a URL for one cache entry and expiry. purpose is mixed into the
// message so a URL handed out for writing an entry cannot be replayed to read one, and the
// other way round. Downloads use the empty purpose, the message v1 has always signed.
func (h *Handler) computeSignature(purpose string, cacheID, exp int64) string {
mac := hmac.New(sha256.New, h.secret)
fmt.Fprintf(mac, "%d:%d", cacheID, exp)
fmt.Fprintf(mac, "%s%d:%d", purpose, cacheID, exp)
return hex.EncodeToString(mac.Sum(nil))
}
func (h *Handler) signedArtifactURL(cacheID uint64, exp time.Time) string {
// signedURL builds a URL under path that signedAuth accepts for the same purpose.
func (h *Handler) signedURL(path, purpose string, cacheID uint64, exp time.Time) string {
expUnix := exp.Unix()
sig := h.computeSignature(int64(cacheID), expUnix)
q := url.Values{}
q.Set("exp", strconv.FormatInt(expUnix, 10))
q.Set("sig", sig)
return fmt.Sprintf("%s%s/artifacts/%d?%s", h.ExternalURL(), apiPath, cacheID, q.Encode())
q.Set("sig", h.computeSignature(purpose, int64(cacheID), expUnix))
return fmt.Sprintf("%s%s/%d?%s", h.ExternalURL(), path, cacheID, q.Encode())
}
func (h *Handler) signedArtifactURL(cacheID uint64, exp time.Time) string {
return h.signedURL(apiPath+"/artifacts", "", cacheID, exp)
}
// if not found, return (nil, nil) instead of an error.
@@ -675,16 +699,12 @@ func findCache(db *bolthold.Store, repo string, keys []string, version string) (
cache := &Cache{}
for _, prefix := range keys {
// if a key in the list matches exactly, don't return partial matches
if err := db.FindOne(cache,
bolthold.Where("Repo").Eq(repo).
And("Key").Eq(prefix).
And("Version").Eq(version).
And("Complete").Eq(true).
SortBy("CreatedAt").Reverse()); err == nil || !errors.Is(err, bolthold.ErrNotFound) {
if err != nil {
return nil, fmt.Errorf("find cache: %w", err)
}
return cache, nil
exact, err := findExactCache(db, repo, prefix, version, true)
if err != nil {
return nil, err
}
if exact != nil {
return exact, nil
}
prefixPattern := "^" + regexp.QuoteMeta(prefix)
re, err := regexp.Compile(prefixPattern)
@@ -707,6 +727,34 @@ func findCache(db *bolthold.Store, repo string, keys []string, version string) (
return nil, nil //nolint:nilnil // pre-existing issue from nektos/act
}
// findExactCache returns the entry for exactly this key and version, or (nil, nil) if there is
// none. Unlike findCache it never falls back to a prefix (restore-key) match, which is what both
// its callers need: a new key that is only a prefix of an existing key is not the same entry.
//
// A completed entry is the one to restore, sorted by when it was written. An incomplete one is a
// reservation being uploaded to, sorted by when it was last written to, because the upload route
// touches UsedAt on every part.
func findExactCache(db *bolthold.Store, repo, key, version string, complete bool) (*Cache, error) {
sortBy := "UsedAt"
if complete {
sortBy = "CreatedAt"
}
cache := &Cache{}
err := db.FindOne(cache,
bolthold.Where("Repo").Eq(repo).
And("Key").Eq(key).
And("Version").Eq(version).
And("Complete").Eq(complete).
SortBy(sortBy).Reverse())
if errors.Is(err, bolthold.ErrNotFound) {
return nil, nil //nolint:nilnil // absence is not an error here
}
if err != nil {
return nil, fmt.Errorf("find cache: %w", err)
}
return cache, nil
}
func insertCache(db *bolthold.Store, cache *Cache) error {
if err := db.Insert(bolthold.NextSequence(), cache); err != nil {
return fmt.Errorf("insert cache: %w", err)
@@ -718,18 +766,30 @@ func insertCache(db *bolthold.Store, cache *Cache) error {
return nil
}
func (h *Handler) useCache(id int64) {
// touchCache stamps UsedAt so gcCache does not reap an entry mid-upload. With requireIncomplete
// it also refuses an entry that is already complete, which is what the v2 blob route needs: its
// upload URL outlives the finalize call, and overwriting a finished entry would leave the blob
// other jobs restore no longer matching its recorded size. An entry missing from the store is
// accepted, since the signature proves the id was handed out.
func (h *Handler) touchCache(id uint64, requireIncomplete bool) error {
db, err := h.openDB()
if err != nil {
return
return err
}
defer db.Close()
cache := &Cache{}
if err := db.Get(id, cache); err != nil {
return
if errors.Is(err, bolthold.ErrNotFound) {
return nil
}
return err
}
if requireIncomplete && cache.Complete {
return fmt.Errorf("cache %d: already complete", id)
}
cache.UsedAt = time.Now().Unix()
_ = db.Update(cache.ID, cache)
return db.Update(cache.ID, cache)
}
const (

View File

@@ -1041,14 +1041,14 @@ func TestHandler_SecretPersistsAcrossRestarts(t *testing.T) {
first, err := StartHandler(dir, "127.0.0.1", 0, "", nil)
require.NoError(t, err)
exp := time.Now().Add(artifactURLTTL).Unix()
sig := first.computeSignature(42, exp)
sig := first.computeSignature("", 42, exp)
require.NoError(t, first.Close())
second, err := StartHandler(dir, "127.0.0.1", 0, "", nil)
require.NoError(t, err)
defer second.Close()
assert.Equal(t, sig, second.computeSignature(42, exp))
assert.Equal(t, sig, second.computeSignature("", 42, exp))
}
// TestHandler_ArtifactSignatureDownload is a happy-path round trip that

View File

@@ -0,0 +1,268 @@
// Copyright 2026 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package artifactcache
import (
"cmp"
"encoding/json"
"encoding/xml"
"errors"
"fmt"
"io"
"net/http"
"strconv"
"strings"
"time"
"github.com/julienschmidt/httprouter"
)
// The cache service v2 API. A client on this version talks twirp to
// `github.actions.results.api.v1.CacheService` instead of the /_apis/artifactcache
// endpoints, and uploads the archive to the returned URL with the Azure blob protocol.
// Both API versions are served from the same store, so a repository keeps its cache
// when a workflow moves between action versions.
const (
cacheServiceV2Path = "/twirp/github.actions.results.api.v1.CacheService"
// blobPath authenticates by signature, because the client uploads without an
// Authorization header. Downloads are handed the v1 artifact URL instead.
blobPath = apiPath + "/blobs"
// blobUploadPurpose keeps an upload URL from being replayed to read an entry.
blobUploadPurpose = "upload:"
blobUploadURLTTL = time.Hour
// twirpInternal is the only error code that is not the client's fault.
twirpInternal = "internal"
)
func (h *Handler) registerV2Routes(router *httprouter.Router) {
router.POST(cacheServiceV2Path+"/CreateCacheEntry", h.bearerAuth(h.v2CreateCacheEntry))
router.POST(cacheServiceV2Path+"/FinalizeCacheEntryUpload", h.bearerAuth(h.v2FinalizeCacheEntryUpload))
router.POST(cacheServiceV2Path+"/GetCacheEntryDownloadURL", h.bearerAuth(h.v2GetCacheEntryDownloadURL))
router.PUT(blobPath+"/:id", h.signedAuth(blobUploadPurpose, h.v2UploadBlob))
}
// An entry that already exists is reported as not ok, which is how the client learns to skip
// the upload.
func (h *Handler) v2CreateCacheEntry(w http.ResponseWriter, r *http.Request, _ httprouter.Params) {
cred := credFromContext(r.Context())
req, err := decodeTwirpRequest[v2CreateRequest](r)
if err != nil {
h.twirpError(w, r, "malformed_request", err)
return
}
if req.Key == "" || req.Version == "" {
h.twirpError(w, r, "invalid_argument", errors.New("key and version are required"))
return
}
db, err := h.openDB()
if err != nil {
h.twirpError(w, r, twirpInternal, err)
return
}
defer db.Close()
// An exact (key, version) match means the entry is already cached; the client then skips
// the upload. A prefix match must not count here, or a shorter key would be reported as
// existing and silently never saved.
if existing, err := findExactCache(db, cred.Repo, req.Key, req.Version, true); err != nil {
h.twirpError(w, r, twirpInternal, err)
return
} else if existing != nil {
h.twirpNotOK(w, r)
return
}
now := time.Now().Unix()
cache := &Cache{
Repo: cred.Repo,
Key: req.Key,
Version: req.Version,
Size: -1, // the size is only known at finalize time
CreatedAt: now,
UsedAt: now,
}
if err := insertCache(db, cache); err != nil {
h.twirpError(w, r, twirpInternal, err)
return
}
h.responseJSON(w, r, http.StatusOK, map[string]any{
"ok": true,
"signedUploadUrl": h.signedURL(blobPath, blobUploadPurpose, cache.ID, time.Now().Add(blobUploadURLTTL)),
})
}
func (h *Handler) v2FinalizeCacheEntryUpload(w http.ResponseWriter, r *http.Request, _ httprouter.Params) {
cred := credFromContext(r.Context())
req, err := decodeTwirpRequest[v2FinalizeRequest](r)
if err != nil {
h.twirpError(w, r, "malformed_request", err)
return
}
db, err := h.openDB()
if err != nil {
h.twirpError(w, r, twirpInternal, err)
return
}
defer db.Close()
cache, err := findExactCache(db, cred.Repo, req.Key, req.Version, false)
if err != nil {
h.twirpError(w, r, twirpInternal, err)
return
}
if cache == nil {
h.twirpNotOK(w, r)
return
}
db.Close() // commitCache needs the store closed
cache.Size, _ = cmp.Or(req.SizeBytes, req.SizeBytesCamel).Int64()
if err := h.commitCache(cache); err != nil {
h.logger.Errorf("finalize cache %d (%s): %v", cache.ID, cache.Key, err)
h.twirpNotOK(w, r)
return
}
h.responseJSON(w, r, http.StatusOK, map[string]any{
"ok": true,
// int64 fields travel as strings in the proto JSON mapping.
"entryId": strconv.FormatUint(cache.ID, 10),
})
}
func (h *Handler) v2GetCacheEntryDownloadURL(w http.ResponseWriter, r *http.Request, _ httprouter.Params) {
cred := credFromContext(r.Context())
req, err := decodeTwirpRequest[v2DownloadRequest](r)
if err != nil {
h.twirpError(w, r, "malformed_request", err)
return
}
db, err := h.openDB()
if err != nil {
h.twirpError(w, r, twirpInternal, err)
return
}
defer db.Close()
cache, err := h.lookupCache(db, cred.Repo, req.keys(), req.Version)
if err != nil {
h.twirpError(w, r, twirpInternal, err)
return
}
if cache == nil {
h.twirpNotOK(w, r)
return
}
h.responseJSON(w, r, http.StatusOK, map[string]any{
"ok": true,
"signedDownloadUrl": h.signedArtifactURL(cache.ID, time.Now().Add(artifactURLTTL)),
"matchedKey": cache.Key,
})
}
// The archive arrives over the subset of the Azure blob API the toolkit uses: a small
// cache is a single PUT, a large one is staged as blocks that a final block list puts
// in order.
func (h *Handler) v2UploadBlob(w http.ResponseWriter, r *http.Request, params httprouter.Params) {
id, err := strconv.ParseUint(params.ByName("id"), 10, 64)
if err != nil {
h.responseJSON(w, r, http.StatusBadRequest, err)
return
}
if err := h.touchCache(id, true); err != nil {
h.responseJSON(w, r, http.StatusBadRequest, err)
return
}
query := r.URL.Query()
switch strings.ToLower(query.Get("comp")) {
case "block":
blockID := query.Get("blockid")
if blockID == "" {
h.responseJSON(w, r, http.StatusBadRequest, errors.New("missing blockid"))
return
}
err = h.storage.WriteBlock(id, blockID, r.Body)
case "blocklist":
var list struct{ Latest []string }
if err := xml.NewDecoder(io.LimitReader(r.Body, 8<<20)).Decode(&list); err != nil {
h.responseJSON(w, r, http.StatusBadRequest, fmt.Errorf("malformed block list: %w", err))
return
}
err = h.storage.OrderBlocks(id, list.Latest)
default:
err = h.storage.Write(id, 0, r.Body)
}
if err != nil {
h.responseJSON(w, r, http.StatusInternalServerError, err)
return
}
w.WriteHeader(http.StatusCreated)
}
// twirpNotOK is the negative answer all three endpoints share: no such entry to restore, no
// reservation to finalize, or an entry that already exists and need not be uploaded again.
func (h *Handler) twirpNotOK(w http.ResponseWriter, r *http.Request) {
h.responseJSON(w, r, http.StatusOK, map[string]any{"ok": false})
}
// twirpError reports in the shape a twirp client expects, so the toolkit surfaces the message
// instead of a parse error.
func (h *Handler) twirpError(w http.ResponseWriter, r *http.Request, code string, err error) {
h.logger.Debugf("%s %s: %v", r.Method, r.URL.Path, err)
status := http.StatusBadRequest
if code == twirpInternal {
status = http.StatusInternalServerError
}
h.responseJSON(w, r, status, map[string]any{"code": code, "msg": err.Error()})
}
// The twirp request bodies. The toolkit's client serialises with useProtoFieldName, so the proto
// names are what arrive; the camelCase spellings of the same mapping are accepted too, as are
// int64s sent as a bare number rather than the string the mapping prescribes.
type (
v2CreateRequest struct {
Key string `json:"key"`
Version string `json:"version"`
}
v2FinalizeRequest struct {
Key string `json:"key"`
Version string `json:"version"`
SizeBytes json.Number `json:"size_bytes"`
SizeBytesCamel json.Number `json:"sizeBytes"`
}
v2DownloadRequest struct {
Key string `json:"key"`
Version string `json:"version"`
RestoreKeys []string `json:"restore_keys"`
RestoreKeysCamel []string `json:"restoreKeys"`
}
)
func (d v2DownloadRequest) keys() []string {
restoreKeys := d.RestoreKeys
if len(restoreKeys) == 0 {
restoreKeys = d.RestoreKeysCamel
}
return append([]string{d.Key}, restoreKeys...)
}
func decodeTwirpRequest[T any](r *http.Request) (T, error) {
var req T
err := json.NewDecoder(io.LimitReader(r.Body, 1<<20)).Decode(&req)
return req, err
}

View File

@@ -0,0 +1,236 @@
// Copyright 2026 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package artifactcache
import (
"bytes"
"encoding/base64"
"encoding/json"
"fmt"
"io"
"net/http"
"path/filepath"
"strconv"
"testing"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// v2Call posts a twirp request to the cache service and returns the decoded response.
// Field names are the proto ones, which is what the toolkit's client sends.
func v2Call(t *testing.T, handler *Handler, client *http.Client, method string, request any) map[string]any {
t.Helper()
body, err := json.Marshal(request)
require.NoError(t, err)
resp, err := client.Post(handler.ExternalURL()+cacheServiceV2Path+"/"+method, "application/json", bytes.NewReader(body))
require.NoError(t, err)
defer resp.Body.Close()
require.Equal(t, http.StatusOK, resp.StatusCode)
got := map[string]any{}
require.NoError(t, json.NewDecoder(resp.Body).Decode(&got))
return got
}
// putBlob uploads to a signed URL and returns the status, so a test can assert a refusal.
func putBlob(t *testing.T, url string, content []byte) int {
t.Helper()
req, err := http.NewRequestWithContext(t.Context(), http.MethodPut, url, bytes.NewReader(content))
require.NoError(t, err)
resp, err := http.DefaultClient.Do(req)
require.NoError(t, err)
defer resp.Body.Close()
return resp.StatusCode
}
func getURL(t *testing.T, url string) []byte {
t.Helper()
req, err := http.NewRequestWithContext(t.Context(), http.MethodGet, url, nil)
require.NoError(t, err)
resp, err := http.DefaultClient.Do(req)
require.NoError(t, err)
defer resp.Body.Close()
body, err := io.ReadAll(resp.Body)
require.NoError(t, err)
return body
}
func startTestHandler(t *testing.T) *Handler {
t.Helper()
handler, err := StartHandler(filepath.Join(t.TempDir(), "artifactcache"), "127.0.0.1", 0, "", nil)
require.NoError(t, err)
t.Cleanup(func() { _ = handler.Close() })
handler.RegisterJob(testToken, testRepo)
return handler
}
// saveV2 runs the reserve/upload/finalize sequence and returns the finalize response along
// with the upload URL it used.
func saveV2(t *testing.T, handler *Handler, key, version string, content []byte) (finalized map[string]any, uploadURL string) {
t.Helper()
created := v2Call(t, handler, testClient, "CreateCacheEntry", map[string]any{"key": key, "version": version})
require.Equal(t, true, created["ok"])
uploadURL, _ = created["signedUploadUrl"].(string)
require.NotEmpty(t, uploadURL)
require.Equal(t, http.StatusCreated, putBlob(t, uploadURL, content))
return v2Call(t, handler, testClient, "FinalizeCacheEntryUpload", map[string]any{
"key": key, "version": version,
"size_bytes": strconv.Itoa(len(content)),
}), uploadURL
}
// The whole round trip an actions/cache v2 client makes, plus the guarantees on the signed
// URLs it is handed: unsigned requests are refused, an upload URL cannot be replayed to read
// or to replace a finalized entry.
func TestCacheServiceV2RoundTrip(t *testing.T) {
handler := startTestHandler(t)
content := []byte("the cached archive")
unsigned := fmt.Sprintf("%s%s/1", handler.ExternalURL(), blobPath)
assert.Equal(t, http.StatusUnauthorized, putBlob(t, unsigned, content))
finalized, uploadURL := saveV2(t, handler, "deps-v1", "abc123", content)
require.Equal(t, true, finalized["ok"])
assert.NotEmpty(t, finalized["entryId"])
// The upload URL outlives the finalize call, so replaying it must not poison the entry,
// and it is an upload URL only: nothing reads a blob back through it.
assert.Equal(t, http.StatusBadRequest, putBlob(t, uploadURL, []byte("poisoned")))
resp, err := http.Get(uploadURL) //nolint:noctx // the URL is the server under test
require.NoError(t, err)
defer resp.Body.Close()
assert.Equal(t, http.StatusMethodNotAllowed, resp.StatusCode)
got := v2Call(t, handler, testClient, "GetCacheEntryDownloadURL", map[string]any{"key": "deps-v1", "version": "abc123"})
require.Equal(t, true, got["ok"])
assert.Equal(t, "deps-v1", got["matchedKey"])
downloadURL, _ := got["signedDownloadUrl"].(string)
require.NotEmpty(t, downloadURL)
assert.Equal(t, content, getURL(t, downloadURL))
}
// A large archive is staged as blocks and only put in order by the final block list, so
// blocks that arrive out of order must still be assembled the way the client asked.
func TestCacheServiceV2BlockUpload(t *testing.T) {
handler := startTestHandler(t)
created := v2Call(t, handler, testClient, "CreateCacheEntry", map[string]any{"key": "blocks", "version": "v1"})
uploadURL, _ := created["signedUploadUrl"].(string)
require.NotEmpty(t, uploadURL)
blocks := map[string][]byte{}
var order []string
for i, part := range []string{"hello ", "world", "!"} {
blockID := base64.StdEncoding.EncodeToString(fmt.Appendf(nil, "block-%d", i))
blocks[blockID] = []byte(part)
order = append(order, blockID)
}
// Upload in an order that is not the block list order.
for _, blockID := range []string{order[2], order[0], order[1]} {
require.Equal(t, http.StatusCreated, putBlob(t, uploadURL+"&comp=block&blockid="+blockID, blocks[blockID]))
}
var list bytes.Buffer
list.WriteString(`<?xml version="1.0" encoding="utf-8"?><BlockList>`)
for _, blockID := range order {
fmt.Fprintf(&list, "<Latest>%s</Latest>", blockID)
}
list.WriteString(`</BlockList>`)
require.Equal(t, http.StatusCreated, putBlob(t, uploadURL+"&comp=blocklist", list.Bytes()))
finalized := v2Call(t, handler, testClient, "FinalizeCacheEntryUpload", map[string]any{
"key": "blocks", "version": "v1", "size_bytes": len("hello world!"),
})
require.Equal(t, true, finalized["ok"])
got := v2Call(t, handler, testClient, "GetCacheEntryDownloadURL", map[string]any{"key": "blocks", "version": "v1"})
require.Equal(t, true, got["ok"])
assert.Equal(t, "hello world!", string(getURL(t, got["signedDownloadUrl"].(string))))
}
func TestCacheServiceV2Lookups(t *testing.T) {
handler := startTestHandler(t)
saved, _ := saveV2(t, handler, "deps-abc", "v1", []byte("x"))
require.Equal(t, true, saved["ok"])
t.Run("reports a miss for an unknown key", func(t *testing.T) {
got := v2Call(t, handler, testClient, "GetCacheEntryDownloadURL", map[string]any{"key": "nothing", "version": "v1"})
assert.Equal(t, false, got["ok"])
})
// The toolkit serialises with the proto field names; the camelCase spellings of the same
// proto JSON mapping are accepted alongside them.
for _, field := range []string{"restore_keys", "restoreKeys"} {
t.Run("restore keys match by prefix, spelled "+field, func(t *testing.T) {
got := v2Call(t, handler, testClient, "GetCacheEntryDownloadURL", map[string]any{
"key": "deps-zzz", field: []string{"deps-"}, "version": "v1",
})
require.Equal(t, true, got["ok"])
assert.Equal(t, "deps-abc", got["matchedKey"])
})
}
t.Run("an existing entry is not reserved twice", func(t *testing.T) {
again := v2Call(t, handler, testClient, "CreateCacheEntry", map[string]any{"key": "deps-abc", "version": "v1"})
assert.Equal(t, false, again["ok"])
})
// A key that is only a prefix of an existing one is a different entry, so the
// reservation check must be exact and not a restore-key prefix match, or the shorter
// key would be reported as existing and silently never saved.
t.Run("a prefix of an existing key is still reserved", func(t *testing.T) {
reserved := v2Call(t, handler, testClient, "CreateCacheEntry", map[string]any{"key": "deps", "version": "v1"})
require.Equal(t, true, reserved["ok"])
assert.NotEmpty(t, reserved["signedUploadUrl"])
})
t.Run("finalizing without a reservation is not ok", func(t *testing.T) {
got := v2Call(t, handler, testClient, "FinalizeCacheEntryUpload", map[string]any{
"key": "never-reserved", "version": "v1", "size_bytes": 1,
})
assert.Equal(t, false, got["ok"])
})
// The size the client declares is what Commit validates the assembled archive against.
t.Run("finalizing with the wrong size is not ok", func(t *testing.T) {
created := v2Call(t, handler, testClient, "CreateCacheEntry", map[string]any{"key": "wrong-size", "version": "v1"})
require.Equal(t, http.StatusCreated, putBlob(t, created["signedUploadUrl"].(string), []byte("four")))
got := v2Call(t, handler, testClient, "FinalizeCacheEntryUpload", map[string]any{
"key": "wrong-size", "version": "v1", "size_bytes": 99,
})
assert.Equal(t, false, got["ok"])
})
// Both API versions are served from one store, so an entry written through v2 is a hit for
// a v1 client asking for the same key and version.
t.Run("a v1 client sees an entry written through v2", func(t *testing.T) {
resp, err := testClient.Get(fmt.Sprintf("%s%s/cache?keys=deps-abc&version=v1", handler.ExternalURL(), apiPath))
require.NoError(t, err)
defer resp.Body.Close()
require.Equal(t, http.StatusOK, resp.StatusCode)
got := map[string]any{}
require.NoError(t, json.NewDecoder(resp.Body).Decode(&got))
assert.Equal(t, "deps-abc", got["cacheKey"])
assert.NotEmpty(t, got["archiveLocation"])
})
// The cache of one repository must stay invisible to another, as it does for the v1 API.
t.Run("another repository sees nothing", func(t *testing.T) {
handler.RegisterJob("other-runtime-token", "other/repo")
otherClient := &http.Client{Transport: &bearerTransport{token: "other-runtime-token"}}
got := v2Call(t, handler, otherClient, "GetCacheEntryDownloadURL", map[string]any{"key": "deps-abc", "version": "v1"})
assert.Equal(t, false, got["ok"])
})
}

View File

@@ -5,12 +5,15 @@
package artifactcache
import (
"crypto/sha256"
"encoding/hex"
"fmt"
"io"
"net/http"
"os"
"path/filepath"
"strconv"
"strings"
)
type Storage struct {
@@ -37,7 +40,10 @@ func (s *Storage) Exist(id uint64) (bool, error) {
}
func (s *Storage) Write(id uint64, offset int64, reader io.Reader) error {
name := s.tempName(id, offset)
return s.writeFile(s.tempName(id, offset), reader)
}
func (s *Storage) writeFile(name string, reader io.Reader) error {
if err := os.MkdirAll(filepath.Dir(name), 0o755); err != nil {
return err
}
@@ -51,6 +57,26 @@ func (s *Storage) Write(id uint64, offset int64, reader io.Reader) error {
return err
}
func (s *Storage) WriteBlock(id uint64, blockID string, reader io.Reader) error {
return s.writeFile(s.blockName(id, blockID), reader)
}
// OrderBlocks renames the staged blocks into the order the block list gives. A block the list
// does not name keeps its staged name, which is how Commit leaves it out, as Azure drops it. One
// rename pass is safe because a staged name always carries blockFilePrefix and a target name
// never does, so no rename can collide with a block not yet moved.
func (s *Storage) OrderBlocks(id uint64, blockIDs []string) error {
for i, blockID := range blockIDs {
if err := os.Rename(s.blockName(id, blockID), s.tempName(id, int64(i))); err != nil {
if os.IsNotExist(err) {
return fmt.Errorf("block %q of cache %d was never uploaded: %w", blockID, id, err)
}
return err
}
}
return nil
}
func (s *Storage) Commit(id uint64, size int64) (int64, error) {
defer func() {
_ = os.RemoveAll(s.tempDir(id))
@@ -65,6 +91,31 @@ func (s *Storage) Commit(id uint64, size int64) (int64, error) {
if err := os.MkdirAll(filepath.Dir(name), 0o755); err != nil {
return 0, err
}
written, err := assemble(name, tempNames)
if err != nil {
return 0, err
}
// If size is less than 0, it means the size is unknown.
// We can't check the size of the file, just skip the check.
// It happens when the request comes from old versions of actions, like `actions/cache@v2`.
if size >= 0 && written != size {
_ = os.Remove(name)
return 0, fmt.Errorf("broken file: %v != %v", written, size)
}
return written, nil
}
// assemble concatenates the uploaded parts into name. A single part, which is what the v2 API
// produces below the client's block threshold, is already the whole archive and is moved.
func assemble(name string, tempNames []string) (int64, error) {
if len(tempNames) == 1 {
info, err := os.Stat(tempNames[0])
if err != nil {
return 0, err
}
return info.Size(), os.Rename(tempNames[0], name)
}
file, err := os.Create(name)
if err != nil {
return 0, err
@@ -84,16 +135,6 @@ func (s *Storage) Commit(id uint64, size int64) (int64, error) {
}
written += n
}
// If size is less than 0, it means the size is unknown.
// We can't check the size of the file, just skip the check.
// It happens when the request comes from old versions of actions, like `actions/cache@v2`.
if size >= 0 && written != size {
_ = file.Close()
_ = os.Remove(name)
return 0, fmt.Errorf("broken file: %v != %v", written, size)
}
return written, nil
}
@@ -119,6 +160,17 @@ func (s *Storage) tempName(id uint64, offset int64) string {
return filepath.Join(s.tempDir(id), fmt.Sprintf("%016x", offset))
}
// blockFilePrefix marks a staged, not yet ordered block, so that tempNames can keep it out of
// Commit's name-ordered concatenation.
const blockFilePrefix = "block-"
func (s *Storage) blockName(id uint64, blockID string) string {
// The block id is client-chosen (base64), so it is hashed rather than trusted as a
// path element.
sum := sha256.Sum256([]byte(blockID))
return filepath.Join(s.tempDir(id), blockFilePrefix+hex.EncodeToString(sum[:]))
}
func (s *Storage) tempNames(id uint64) ([]string, error) {
dir := s.tempDir(id)
files, err := os.ReadDir(dir)
@@ -127,7 +179,7 @@ func (s *Storage) tempNames(id uint64) ([]string, error) {
}
var names []string
for _, v := range files {
if !v.IsDir() {
if !v.IsDir() && !strings.HasPrefix(v.Name(), blockFilePrefix) {
names = append(names, filepath.Join(dir, v.Name()))
}
}

View File

@@ -595,8 +595,8 @@ func actionStagePaths(step actionStep) (actionDir, actionPath, actionName, conta
rc := step.getRunContext()
stepModel := step.getStepModel()
if _, ok := step.(*stepActionRemote); ok {
actionDir = fmt.Sprintf("%s/%s", rc.ActionCacheDir(), stepModel.UsesHash())
if sar, ok := step.(*stepActionRemote); ok {
actionDir = sar.actionDir()
actionPath = newRemoteAction(stepModel.Uses).Path
} else {
actionDir = filepath.Join(rc.Config.Workdir, stepModel.Uses)

View File

@@ -69,6 +69,8 @@ func (sar *stepActionRemote) prepareActionExecutor() common.Executor {
github.Token = sar.RunContext.Config.ReplaceGheActionTokenWithGithubCom
}
}
// Actions served from the action cache are read out of a git object store rather than a
// directory, so they never reach the bundle patch below and keep to the v1 cache API.
if sar.RunContext.Config.ActionCache != nil {
cache := sar.RunContext.Config.ActionCache
@@ -112,7 +114,7 @@ func (sar *stepActionRemote) prepareActionExecutor() common.Executor {
return err
}
actionDir := fmt.Sprintf("%s/%s", sar.RunContext.ActionCacheDir(), sar.Step.UsesHash())
actionDir := sar.actionDir()
defaultActionURL := sar.RunContext.Config.DefaultActionURL()
// For Gitea
// A composite RunContext nils Config.Secrets, so getGitCloneToken would yield an
@@ -171,6 +173,9 @@ func (sar *stepActionRemote) prepareActionExecutor() common.Executor {
sar.action = actionModel
return err
},
// A stage of its own: it takes the same clone lock, and it has to land before
// runAction copies the action into the job container.
sar.patchActionToolkit,
)(ctx)
}
}
@@ -189,7 +194,7 @@ func (sar *stepActionRemote) pre() common.Executor {
return common.NewPipelineExecutor(
sar.prepareActionExecutor(),
runStepExecutor(sar, stepStagePre, runPreStep(sar)).If(hasPreStep(sar)).If(shouldRunPreStep(sar)))
runStepExecutor(sar, stepStagePre, sar.revertToolkitOnFailure(runPreStep(sar))).If(hasPreStep(sar)).If(shouldRunPreStep(sar)))
}
func (sar *stepActionRemote) main() common.Executor {
@@ -211,15 +216,51 @@ func (sar *stepActionRemote) main() common.Executor {
return sar.RunContext.JobContainer.CopyDir(copyToPath, sar.RunContext.Config.Workdir+string(filepath.Separator)+".", sar.RunContext.Config.UseGitIgnore)(ctx)
}
actionDir := fmt.Sprintf("%s/%s", sar.RunContext.ActionCacheDir(), sar.Step.UsesHash())
actionDir := sar.actionDir()
return sar.runAction(sar, actionDir, sar.remoteAction)(ctx)
return sar.revertToolkitOnFailure(sar.runAction(sar, actionDir, sar.remoteAction))(ctx)
}),
)
}
func (sar *stepActionRemote) post() common.Executor {
return runStepExecutor(sar, stepStagePost, runPostStep(sar)).If(hasPostStep(sar)).If(shouldRunPostStep(sar))
return runStepExecutor(sar, stepStagePost, sar.revertToolkitOnFailure(runPostStep(sar))).If(hasPostStep(sar)).If(shouldRunPostStep(sar))
}
// toolkitBundles is the action directory and the entrypoints the toolkit may live in.
func (sar *stepActionRemote) toolkitBundles() (string, []string) {
if sar.remoteAction == nil {
return "", nil
}
dir := sar.actionDir()
return dir, actionScriptPaths(filepath.Join(dir, sar.remoteAction.Path), sar.action)
}
// patchActionToolkit edits the bundled toolkit so it works against Gitea, which lets the cache
// client use the v2 API this runner serves. A no-op unless the runner serves it.
func (sar *stepActionRemote) patchActionToolkit(ctx context.Context) error {
if sar.RunContext.GetEnv()[CacheServiceV2Env] != "" {
dir, scripts := sar.toolkitBundles()
patchToolkit(ctx, dir, scripts)
}
return nil
}
// revertToolkitOnFailure restores the untouched bundles when the action fails, so a later job
// runs it as shipped rather than repeating a failure the patch may have caused.
func (sar *stepActionRemote) revertToolkitOnFailure(exec common.Executor) common.Executor {
return func(ctx context.Context) error {
err := exec(ctx)
if err != nil {
dir, scripts := sar.toolkitBundles()
revertToolkit(ctx, dir, scripts)
}
return err
}
}
func (sar *stepActionRemote) actionDir() string {
return fmt.Sprintf("%s/%s", sar.RunContext.ActionCacheDir(), sar.Step.UsesHash())
}
func (sar *stepActionRemote) getRunContext() *RunContext {
@@ -270,7 +311,7 @@ func (sar *stepActionRemote) getActionModel() *model.Action {
func (sar *stepActionRemote) getCompositeRunContext(ctx context.Context) *RunContext {
if sar.compositeRunContext == nil {
actionDir := fmt.Sprintf("%s/%s", sar.RunContext.ActionCacheDir(), sar.Step.UsesHash())
actionDir := sar.actionDir()
actionLocation := path.Join(actionDir, sar.remoteAction.Path)
_, containerActionDir := getContainerActionPaths(sar.getStepModel(), actionLocation, sar.RunContext)

262
act/runner/toolkit_patch.go Normal file
View File

@@ -0,0 +1,262 @@
// Copyright 2026 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package runner
import (
"bytes"
"context"
"os"
"path/filepath"
"regexp"
"strings"
"gitea.com/gitea/runner/act/common"
"gitea.com/gitea/runner/act/common/git"
"gitea.com/gitea/runner/act/model"
)
// Actions bundle the @actions toolkit into their own JavaScript, and two of its lines keep it
// from working against Gitea. Both are edited out of the bundle the runner downloaded.
//
// isGhes() takes any host that is not github.com, *.ghe.com or *.localhost for GitHub
// Enterprise. @actions/cache then forces the v1 API, and @actions/artifact refuses outright,
// which is why the stock upload-artifact aborts here. The edit empties the last of the three
// hostname tests, so `endsWith('.LOCALHOST')` becomes `endsWith(”)`, which every hostname
// satisfies: one string literal, no call sites to resolve, and the same answer the toolkit's own
// proposed ACTIONS_VENDOR switch would give. Gitea already makes this edit by hand in its fork
// of upload-artifact.
//
// getCacheServiceURL() then resolves the cache service from ACTIONS_RESULTS_URL alone, where v1
// reads ACTIONS_CACHE_URL first. Both reads there are given the same preference, which is what
// keeps the runner out of the artifact path: the results URL still points at Gitea.
//
// Either of these landing upstream makes this file deletable:
//
// https://github.com/actions/toolkit/pull/2123 — an ACTIONS_VENDOR switch, naming Gitea
// https://github.com/actions/toolkit/issues/2439 — treat ACTIONS_RESULTS_URL as the signal
const (
CacheServiceV2Env = "ACTIONS_CACHE_SERVICE_V2"
cacheURLEnv = "ACTIONS_CACHE_URL"
resultsURLEnv = "ACTIONS_RESULTS_URL"
// localhostHost is the suffix isGhes accepts; emptying the test is what opens the gate,
// because every hostname ends with the empty string.
localhostHost = ".LOCALHOST"
// artifactRefusal is the only thing the gate guards in @actions/artifact, which is what makes
// such a bundle safe to open. A bundle carrying neither toolkit uses isGhes for something this
// runner has not looked at, and is left alone.
artifactRefusal = "GHESNotSupportedError"
// sidecarSuffix names the directory of untouched copies, a sibling of the action directory
// because that directory is copied wholesale into job containers.
sidecarSuffix = ".toolkit-patch"
// skipMarker in the sidecar means a patched bundle already failed once here.
skipMarker = "skip"
maxBundleSize = 64 << 20
)
var (
// localhostTest matches the third hostname test of isGhes, in any quoting. The match is case
// sensitive on purpose, and that is load-bearing: isGhes uppercases the hostname before
// testing it, while undici, bundled into all of these actions, tests a lowercase ".localhost"
// in isURLPotentiallyTrustworthy. Opening that one would tell its HTTP client that every URL
// is trustworthy. Uppercase, the literal occurs nowhere but this test, across 118 bundles
// covering every major version of sixteen actions.
localhostTest = regexp.MustCompile(`endsWith\s*\(\s*` + quoted(regexp.QuoteMeta(localhostHost)) + `\s*\)`)
// serviceURLBranches matches both branches of getCacheServiceURL at once: the v1 branch reads
// the cache URL and falls back to the results URL, the v2 branch just below reads the results
// URL alone. That `||` pairing is the only place the two variables are read together, so
// matching them as one expression is what keeps the edit inside this function rather than
// anywhere they happen to sit near each other. The branches are 21 bytes apart minified and
// 63 not, across every bundle measured.
serviceURLBranches = regexp.MustCompile(`(` + envRead(cacheURLEnv) + `\s*\|\|\s*)(` +
envRead(resultsURLEnv) + `)((?s).{0,256}?)(` + envRead(resultsURLEnv) + `)`)
// cacheURLFirst gives both reads the preference the v1 branch already had.
cacheURLFirst = []byte(`${1}(process.env.` + cacheURLEnv + `||${2})${3}(process.env.` + cacheURLEnv + `||${4})`)
)
func envRead(name string) string {
return `process\s*\.\s*env\s*(?:\.\s*` + name + `\b|\[\s*` + quoted(name) + `\s*\])`
}
// quoted matches a string literal in any of the three quote characters. RE2 has no
// backreferences, so the pairs are spelled out.
func quoted(pattern string) string {
return "(?:'" + pattern + "'|\"" + pattern + "\"|`" + pattern + "`)"
}
// actionScriptPaths returns the entrypoints of a node action, the only kind with a bundle. Only
// remote actions get here: a local one lives in the user's checkout, which the runner does not
// rewrite.
func actionScriptPaths(dir string, action *model.Action) []string {
if action == nil || !action.Runs.Using.IsNode() {
return nil
}
var paths []string
for _, script := range []string{action.Runs.Pre, action.Runs.Main, action.Runs.Post} {
if script != "" {
paths = append(paths, filepath.Join(dir, script))
}
}
return paths
}
// patchToolkit edits the toolkit in an action's bundles, keeping each original beside them. Every
// failure is silent and leaves the bundle as it was, which costs the cache client the v2 API and
// an artifact action nothing at all.
func patchToolkit(ctx context.Context, actionDir string, scripts []string) {
if len(scripts) == 0 {
return
}
if _, err := os.Stat(filepath.Join(sidecarDir(actionDir), skipMarker)); err == nil {
return
}
defer git.AcquireCloneLock(actionDir)()
for _, script := range scripts {
if err := patchBundle(script, originalFor(actionDir, script)); err != nil {
common.Logger(ctx).Debugf("actions toolkit: %s left unpatched: %v", filepath.Base(script), err)
}
}
}
// revertToolkit puts the originals back and stops this action being patched again, so the next job
// runs it exactly as shipped. Called when a step failed with a patched bundle; it does not re-run
// the step, because a step's outputs and env-file writes are already recorded by then.
func revertToolkit(ctx context.Context, actionDir string, scripts []string) {
if len(scripts) == 0 {
return
}
if _, err := os.Stat(sidecarDir(actionDir)); err != nil {
return
}
defer git.AcquireCloneLock(actionDir)()
reverted := false
for _, script := range scripts {
original := originalFor(actionDir, script)
if !isPatchOf(original, script) {
continue
}
if err := os.Rename(original, script); err == nil {
reverted = true
}
}
if reverted {
_ = os.WriteFile(filepath.Join(sidecarDir(actionDir), skipMarker), nil, 0o600)
common.Logger(ctx).Warnf("actions toolkit: restored the original %s, it will not be patched again", filepath.Base(actionDir))
}
}
// sidecarDir holds an action's untouched bundles, and the marker that stops it being patched.
func sidecarDir(actionDir string) string {
return actionDir + sidecarSuffix
}
// originalFor is where a script's untouched copy lives, or "" for a script the action's own
// `runs` keys placed outside its directory, which is not this runner's to rewrite.
func originalFor(actionDir, script string) string {
rel, err := filepath.Rel(actionDir, script)
if err != nil || strings.HasPrefix(rel, "..") {
return ""
}
return filepath.Join(sidecarDir(actionDir), rel)
}
// patchBundle rewrites one entrypoint in place. The untouched copy kept beside it is what marks
// the bundle as already patched.
func patchBundle(script, original string) error {
if original == "" {
return nil
}
if _, err := os.Stat(original); err == nil {
if isPatchOf(original, script) {
return nil
}
// The action's ref moved and git checked the new bundle out over the patched one, so
// the pair no longer belongs together. Patch afresh rather than keep an original that
// would restore an older version of the action.
if err := os.Remove(original); err != nil {
return err
}
}
info, err := os.Stat(script)
if err != nil {
return err
}
if info.Size() > maxBundleSize {
return nil
}
data, err := os.ReadFile(script)
if err != nil {
return err
}
patched, ok := patchedBundle(data)
if !ok {
return nil
}
if err := os.MkdirAll(filepath.Dir(original), 0o755); err != nil {
return err
}
// The copy is taken before the bundle is replaced, so a write that fails part way can put the
// action back as it was. A crash needs no handling: the clone executor checks the action out
// and hard resets it on every prepare, so a half-written bundle never outlives the job.
if err := os.WriteFile(original, data, info.Mode().Perm()); err != nil {
return err
}
if err := os.WriteFile(script, patched, info.Mode().Perm()); err != nil {
_ = os.Rename(original, script)
return err
}
return nil
}
// isPatchOf reports whether script is exactly what patching original produced. It is what proves
// the two still belong together: an action whose ref moved is checked out over the patched bundle,
// leaving an original that would restore the version before the move.
func isPatchOf(original, script string) bool {
data, err := os.ReadFile(original)
if err != nil {
return false
}
current, err := os.ReadFile(script)
if err != nil {
return false
}
patched, ok := patchedBundle(data)
return ok && bytes.Equal(patched, current)
}
// patchedBundle opens the GHES gate, and where the cache toolkit is present, points the cache
// service at the cache server. A bundle this runner cannot account for comes back untouched.
func patchedBundle(data []byte) ([]byte, bool) {
if !localhostTest.Match(data) {
return data, false
}
switch {
case bytes.Contains(data, []byte(CacheServiceV2Env)):
// The cache toolkit: both edits or neither, because choosing v2 without redirecting the
// URL would send the client to a results URL that serves no cache service.
if !serviceURLBranches.Match(data) {
return data, false
}
case bytes.Contains(data, []byte(artifactRefusal)):
// The artifact toolkit, where the gate is a plain refusal and there is no URL to move:
// artifacts already go to Gitea, which implements that service.
default:
return data, false
}
opened := localhostTest.ReplaceAllFunc(data, func(test []byte) []byte {
// Drop the hostname from the test rather than rewriting the call, so the bundle's own
// quoting survives and the result stays valid even inside a string literal.
return bytes.Replace(test, []byte(localhostHost), nil, 1)
})
return serviceURLBranches.ReplaceAll(opened, cacheURLFirst), true
}

View File

@@ -0,0 +1,196 @@
// Copyright 2026 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package runner
import (
"context"
"io"
"net/http"
"os"
"os/exec"
"path/filepath"
"strings"
"testing"
"time"
"gitea.com/gitea/runner/act/artifactcache"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// actionsCacheRef pins the actions/cache release this is verified against. Bump it
// deliberately: a new release is exactly what can stop the patch matching.
const actionsCacheRef = "v6.1.0"
// bundleFromGitHub downloads one entrypoint, keeping it in the user cache dir so repeated runs
// cost nothing. The bundles are megabytes, too large to vendor.
func bundleFromGitHub(t *testing.T, repo, ref, path string) string {
t.Helper()
cacheDir, err := os.UserCacheDir()
require.NoError(t, err)
dir := filepath.Join(cacheDir, "gitea-runner-test", strings.ReplaceAll(repo, "/", "-")+"-"+ref)
bundle := filepath.Join(dir, strings.ReplaceAll(path, "/", "-"))
if _, err := os.Stat(bundle); err == nil {
return bundle
}
require.NoError(t, os.MkdirAll(dir, 0o755))
url := "https://raw.githubusercontent.com/" + repo + "/" + ref + "/" + path
ctx, cancel := context.WithTimeout(t.Context(), time.Minute)
defer cancel()
req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
require.NoError(t, err)
resp, err := http.DefaultClient.Do(req)
if err != nil {
t.Skipf("cannot reach %s: %v", url, err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
t.Skipf("GET %s: %s", url, resp.Status)
}
file, err := os.Create(bundle)
require.NoError(t, err)
_, err = io.Copy(file, resp.Body)
require.NoError(t, file.Close())
require.NoError(t, err)
return bundle
}
// runCacheAction runs one entrypoint the way a job would: a real Gitea server URL, and a results
// URL that points at Gitea rather than at the runner. Nothing about the environment is rewritten,
// so only the patch can make the client choose v2 and find the cache server.
func runCacheAction(t *testing.T, script, workspace, runnerTemp, cacheURL, token, key string) string {
t.Helper()
state := filepath.Join(runnerTemp, "state")
output := filepath.Join(runnerTemp, "output")
for _, name := range []string{state, output} {
require.NoError(t, os.WriteFile(name, nil, 0o600))
}
cmd := exec.CommandContext(t.Context(), "node", script)
cmd.Dir = workspace
cmd.Env = append(os.Environ(),
"INPUT_PATH=to-cache",
"INPUT_KEY="+key,
"ACTIONS_RUNTIME_TOKEN="+token,
"ACTIONS_CACHE_URL="+cacheURL+"/",
// Unreachable on purpose: the artifact service lives here, the cache service must not.
"ACTIONS_RESULTS_URL=https://gitea.example",
"ACTIONS_CACHE_SERVICE_V2=true",
"GITHUB_SERVER_URL=https://gitea.example.com",
"GITHUB_REF=refs/heads/main",
"GITHUB_EVENT_NAME=push",
"GITHUB_WORKSPACE="+workspace,
"RUNNER_TEMP="+runnerTemp,
"GITHUB_STATE="+state,
"GITHUB_OUTPUT="+output,
)
out, err := cmd.CombinedOutput()
t.Logf("%s:\n%s", filepath.Base(filepath.Dir(script)), out)
require.NoError(t, err, "%s failed", script)
return string(out)
}
// tempDirPath is TempDir with symlinks resolved, because macOS hands out /var paths that resolve
// to /private/var and the client derives archive paths relative to the workspace.
func tempDirPath(t *testing.T) string {
t.Helper()
dir, err := filepath.EvalSymlinks(t.TempDir())
require.NoError(t, err)
return dir
}
// The whole chain against the pinned release, whose bundles ship unminified: patch them, run the
// real client with an ordinary Gitea server URL and a results URL that goes nowhere, and have it
// save and restore through this runner's cache server. The unreachable results URL is the point,
// it is what proves the cache reaches the runner without the runner fronting Gitea. If a release
// stops matching the patch the client falls back to v1 and this fails on the version line, which
// is the signal to look at the new bundle.
func TestCacheServiceV2EndToEnd(t *testing.T) {
requireHostTools(t, "node")
// A stand-in action directory, patched exactly as a downloaded one would be.
actionDir := tempDirPath(t)
scripts := map[string]string{}
for _, stage := range []string{"restore", "save"} {
body, err := os.ReadFile(bundleFromGitHub(t, "actions/cache", actionsCacheRef, "dist/"+stage+"/index.js"))
require.NoError(t, err)
scripts[stage] = filepath.Join(actionDir, stage+".js")
require.NoError(t, os.WriteFile(scripts[stage], body, 0o600))
}
patchToolkit(t.Context(), actionDir, []string{scripts["restore"], scripts["save"]})
handler, err := artifactcache.StartHandler(filepath.Join(t.TempDir(), "cache"), "127.0.0.1", 0, "", nil)
require.NoError(t, err)
t.Cleanup(func() { _ = handler.Close() })
const token, repo = "e2e-runtime-token", "testuser/testrepo"
handler.RegisterJob(token, repo)
workspace, runnerTemp := tempDirPath(t), tempDirPath(t)
require.NoError(t, os.MkdirAll(filepath.Join(workspace, "to-cache"), 0o755))
content := []byte("cached through the patched gate")
require.NoError(t, os.WriteFile(filepath.Join(workspace, "to-cache", "data.txt"), content, 0o600))
const key = "patched-gate-key"
missed := runCacheAction(t, scripts["restore"], workspace, runnerTemp, handler.ExternalURL(), token, key)
require.Contains(t, missed, "Cache service version: v2", "the patch did not take, the client stayed on v1")
require.Contains(t, missed, "Cache not found for input keys: "+key)
saved := runCacheAction(t, scripts["save"], workspace, runnerTemp, handler.ExternalURL(), token, key)
require.Contains(t, saved, "Cache saved with key: "+key)
restored := tempDirPath(t)
hit := runCacheAction(t, scripts["restore"], restored, runnerTemp, handler.ExternalURL(), token, key)
require.Contains(t, hit, "Cache restored from key: "+key)
got, err := os.ReadFile(filepath.Join(restored, "to-cache", "data.txt"))
require.NoError(t, err)
assert.Equal(t, content, got)
}
// The gate and the URL getter are separate functions, and a bundler may put either first: the gap
// between them runs from 159 to 1179 bytes across these actions, which is why neither edit is
// anchored on that distance. One entrypoint from each of the families that bundle the cache
// toolkit, patched but not run, is what keeps a future release from quietly matching only one of
// the two shapes and leaving every cache on v1.
func TestToolkitPatchAcrossActions(t *testing.T) {
for _, tc := range []struct{ repo, ref, path string }{
{"actions/setup-go", "v7.0.0", "dist/setup/index.js"},
{"actions/setup-node", "v6.0.0", "dist/cache-save/index.js"},
{"actions/setup-python", "v6.0.0", "dist/setup/index.js"},
{"ruby/setup-ruby", "v1.271.0", "dist/index.js"},
{"pnpm/action-setup", "v6.0.9", "dist/index.js"},
// The artifact toolkit, where the gate is a refusal and there is nothing to redirect.
// v4.4.0 is the first release whose gate carries the localhost test this matches; the
// releases before it refuse in a shape the runner leaves alone.
{"actions/upload-artifact", "v4.4.0", "dist/upload/index.js"},
{"actions/upload-artifact", "v7.0.1", "dist/upload/index.js"},
{"actions/download-artifact", "v6.0.0", "dist/index.js"},
{"oven-sh/setup-bun", "v2.2.0", "dist/setup/index.js"},
} {
t.Run(tc.repo+"@"+tc.ref, func(t *testing.T) {
t.Parallel()
data, err := os.ReadFile(bundleFromGitHub(t, tc.repo, tc.ref, tc.path))
require.NoError(t, err)
out, patched := patchedBundle(data)
assert.True(t, patched, "the version gate was not patched")
assert.NotContains(t, string(out), ".LOCALHOST", "a copy of the gate was missed")
if !strings.Contains(string(data), CacheServiceV2Env) {
return // the artifact toolkit: a refusal to open, and no URL to move
}
// Only the reads inside getCacheServiceURL are rewritten. The others, such as the
// feature-availability check, must be left as they are.
assert.NotZero(t, strings.Count(string(out), "(process.env."+cacheURLEnv+"||process.env"),
"the cache service URL was not redirected")
assert.Equal(t, strings.Count(string(data), resultsURLEnv), strings.Count(string(out), resultsURLEnv),
"a read of the results URL was lost, it must stay as the fallback")
})
}
}

View File

@@ -0,0 +1,328 @@
// Copyright 2026 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package runner
import (
"context"
"errors"
"os"
"os/exec"
"path/filepath"
"strings"
"testing"
"gitea.com/gitea/runner/act/model"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// The three shapes real bundlers emit, reduced to the bytes that matter: the version gate, and
// the URL getter that follows it. tsc keeps the names, webpack prefixes them, esbuild mangles
// them, writes ternaries in place of the switch, and records the real name in the export
// assignment. Each carries both reads of the results URL, as the real getter does.
const (
urlTSC = `function getCacheServiceURL() {` + "\n" + ` switch (getCacheServiceVersion()) {` + "\n" + ` case 'v1':` + "\n" + ` return (process.env['ACTIONS_CACHE_URL'] || process.env['ACTIONS_RESULTS_URL'] || '');` + "\n" + ` case 'v2':` + "\n" + ` return process.env['ACTIONS_RESULTS_URL'] || '';` + "\n" + ` }` + "\n" + `}`
urlEsbuild = `function YK(){let e=XK();return e==="v1"?process.env.ACTIONS_CACHE_URL||process.env.ACTIONS_RESULTS_URL||"":e==="v2"?process.env.ACTIONS_RESULTS_URL||"":""}`
isGhesTSC = `function isGhes(){const h=new URL(process.env['GITHUB_SERVER_URL']||'https://github.com').hostname.toUpperCase();return h!=='GITHUB.COM'&&!h.endsWith('.GHE.COM')&&!h.endsWith('.LOCALHOST')}`
gateTSC = isGhesTSC + "\n" + `function getCacheServiceVersion() {` + "\n" + ` if (isGhes())` + "\n" + ` return 'v1';` + "\n" + ` return process.env['ACTIONS_CACHE_SERVICE_V2'] ? 'v2' : 'v1';` + "\n" + `}` + "\n" + urlTSC
gateWebpack = `function config_isGhes(){const h=new URL(process.env['GITHUB_SERVER_URL']||'https://github.com').hostname.toUpperCase();return h!=='GITHUB.COM'&&!h.endsWith('.GHE.COM')&&!h.endsWith('.LOCALHOST')}` + "\n" + `function config_getCacheServiceVersion() {` + "\n" + ` if (config_isGhes())` + "\n" + ` return 'v1';` + "\n" + ` return process.env['ACTIONS_CACHE_SERVICE_V2'] ? 'v2' : 'v1';` + "\n" + `}` + "\n" + urlTSC
gateEsbuild = `vu.isGhes=$K;vu.getCacheServiceVersion=XK;function $K(){let e=new URL(process.env.GITHUB_SERVER_URL||"https://github.com").hostname.toUpperCase(),r=e==="GITHUB.COM",n=e.endsWith(".GHE.COM"),i=e.endsWith(".LOCALHOST");return!r&&!n&&!i}function XK(){return $K()?"v1":process.env.ACTIONS_CACHE_SERVICE_V2?"v2":"v1"}` + urlEsbuild
)
func TestPatchedBundle(t *testing.T) {
for _, tc := range []struct {
name, body string
wantPatched bool
}{
{"tsc keeps the names", gateTSC, true},
{"webpack prefixes them", gateWebpack, true},
{"esbuild mangles and minifies them", gateEsbuild, true},
// A bundler picks its own quoting; gateTSC is single-quoted already.
{"double-quoted", requoted(`"`), true},
{"backtick-quoted", requoted("`"), true},
// sccache-action sets the variable itself; there is no gate to open.
{"mentions the variable without the gate", `core.exportVariable("ACTIONS_CACHE_SERVICE_V2","on")`, false},
// Both edits or neither: a gate patched without the URL would send the client to a
// results URL that serves no cache service.
{"gate without a recognisable url getter", strings.TrimSuffix(gateTSC, "\n"+urlTSC), false},
// And the other way round: an action that reads both variables but has no gate to open.
{"url getter without a gate", urlTSC, false},
} {
t.Run(tc.name, func(t *testing.T) {
out, patched := patchedBundle([]byte(tc.body))
assert.Equal(t, tc.wantPatched, patched)
if !tc.wantPatched {
assert.Equal(t, tc.body, string(out), "an unpatched bundle must come back byte for byte")
return
}
assert.True(t, gateOpened(string(out)))
// The other two hostname tests are left alone, so a host that really is GitHub or
// GHES is still recognised as such.
assert.NotContains(t, string(out), ".LOCALHOST", "the localhost test is the one that opens")
assert.Contains(t, string(out), ".GHE.COM")
// Every read of the results URL now prefers the cache URL, and none was lost: the
// results URL stays the fallback, so a runner not serving the cache still works.
assert.Equal(t, strings.Count(tc.body, "ACTIONS_RESULTS_URL"), strings.Count(string(out), "ACTIONS_RESULTS_URL"))
assert.Equal(t, strings.Count(tc.body, "ACTIONS_RESULTS_URL"),
strings.Count(string(out), "(process.env.ACTIONS_CACHE_URL||process.env"))
})
}
}
// undici, bundled into every one of these actions, decides whether to trust a URL with a
// lowercase test that reads almost the same. Opening it would tell the HTTP client that every URL
// is trustworthy, so the uppercase the toolkit produces is what separates them.
func TestPatchedBundleLeavesTrustworthyURLCheckAlone(t *testing.T) {
const undici = `if(n.hostname==="localhost"||n.hostname.includes("localhost.")||n.hostname.endsWith(".localhost")){return true}`
out, patched := patchedBundle([]byte(undici + gateTSC))
require.True(t, patched)
assert.Contains(t, string(out), undici, "the trustworthy-URL check must survive byte for byte")
assert.True(t, gateOpened(string(out)))
}
// The artifact toolkit puts the same gate in front of a plain refusal, with no URL to move, so
// opening it is what lets the stock upload-artifact work against Gitea instead of aborting.
func TestPatchedBundleOpensTheArtifactRefusal(t *testing.T) {
const artifact = isGhesTSC + "\n" + `uploadArtifact(){if(isGhes()){throw new GHESNotSupportedError()}}`
out, patched := patchedBundle([]byte(artifact))
assert.True(t, patched)
assert.True(t, gateOpened(string(out)))
assert.Contains(t, string(out), "GHESNotSupportedError", "the refusal itself is left in place, it just stops firing")
// A bundle using the gate for something this runner has not accounted for is not touched.
unknown := strings.Replace(artifact, "GHESNotSupportedError", "SomeOtherError", 1)
out, patched = patchedBundle([]byte(unknown))
assert.False(t, patched)
assert.Equal(t, unknown, string(out))
}
// requoted respells gateTSC's string literals with another quote character.
func requoted(quote string) string {
gate := strings.ReplaceAll(gateTSC, `'.LOCALHOST'`, quote+".LOCALHOST"+quote)
gate = strings.ReplaceAll(gate, `['ACTIONS_RESULTS_URL']`, "["+quote+"ACTIONS_RESULTS_URL"+quote+"]")
return strings.ReplaceAll(gate, `['ACTIONS_CACHE_URL']`, "["+quote+"ACTIONS_CACHE_URL"+quote+"]")
}
// gateOpened reports whether the hostname test was emptied, in whatever quoting the bundle used.
func gateOpened(body string) bool {
return strings.Contains(body, "endsWith(") && !strings.Contains(body, ".LOCALHOST")
}
// The patched bundle must still be JavaScript, and must resolve the way the runner needs: v2 for
// an ordinary Gitea host, the cache server for the service URL, and the results URL when there is
// no cache server. Unpatched, the same bundle must still choose v1, or the patch proves nothing.
func TestPatchedBundleBehavesInNode(t *testing.T) {
requireHostTools(t, "node")
eval := func(t *testing.T, bundle, prelude, cacheURL string) string {
t.Helper()
script := prelude + bundle + "\nprocess.stdout.write(getCacheServiceVersion()+' '+getCacheServiceURL())"
cmd := exec.CommandContext(t.Context(), "node", "-e", script)
cmd.Env = append(os.Environ(),
"ACTIONS_CACHE_SERVICE_V2=true",
"ACTIONS_CACHE_URL="+cacheURL,
"ACTIONS_RESULTS_URL=https://gitea.example",
"GITHUB_SERVER_URL=https://gitea.example",
)
out, err := cmd.CombinedOutput()
require.NoError(t, err, "%s", out)
return string(out)
}
for _, tc := range []struct{ name, bundle, prelude string }{
{"tsc", gateTSC, ""},
{"webpack", gateWebpack, "const getCacheServiceVersion=()=>config_getCacheServiceVersion();"},
{"esbuild", gateEsbuild, "var vu={};const getCacheServiceVersion=()=>XK(),getCacheServiceURL=()=>YK();"},
} {
t.Run(tc.name, func(t *testing.T) {
// Unpatched, a Gitea host is taken for GHES: v1, whose branch already reads the
// cache URL. The patch has to move the version without moving that.
assert.Equal(t, "v1 http://cache:8088/", eval(t, tc.bundle, tc.prelude, "http://cache:8088/"))
patched, ok := patchedBundle([]byte(tc.bundle))
require.True(t, ok)
assert.Equal(t, "v2 http://cache:8088/", eval(t, string(patched), tc.prelude, "http://cache:8088/"))
assert.Equal(t, "v2 https://gitea.example", eval(t, string(patched), tc.prelude, ""),
"with no cache server the results URL is still the fallback")
})
}
}
// A bundler that embeds module sources as strings, such as webpack with devtool: eval, carries
// the gate inside a double-quoted literal. Rewriting the call rather than emptying its argument
// would end that string early and leave the bundle unparseable.
func TestPatchedBundleSurvivesInsideAStringLiteral(t *testing.T) {
requireHostTools(t, "node")
escaped := strings.ReplaceAll(gateTSC, `"`, `\"`)
embedded := `eval("` + strings.ReplaceAll(escaped, "\n", `\n`) + `");`
out, patched := patchedBundle([]byte(embedded))
require.True(t, patched)
file := filepath.Join(t.TempDir(), "bundle.js")
require.NoError(t, os.WriteFile(file, out, 0o600))
checked, err := exec.CommandContext(t.Context(), "node", "--check", file).CombinedOutput()
require.NoError(t, err, "%s", checked)
}
func TestPatchBundleKeepsTheOriginal(t *testing.T) {
dir, script := bundleFile(t, gateTSC)
original := originalFor(dir, script)
require.NoError(t, patchBundle(script, original))
patched, err := os.ReadFile(script)
require.NoError(t, err)
assert.True(t, gateOpened(string(patched)))
kept, err := os.ReadFile(original)
require.NoError(t, err)
assert.Equal(t, gateTSC, string(kept), "the untouched bundle is kept outside the action tree")
assert.NotContains(t, original, dir+string(filepath.Separator), "originals must not ship into job containers")
// Patching again must not stack, and must not overwrite the kept original.
require.NoError(t, patchBundle(script, original))
again, err := os.ReadFile(script)
require.NoError(t, err)
assert.Equal(t, string(patched), string(again))
kept, err = os.ReadFile(original)
require.NoError(t, err)
assert.Equal(t, gateTSC, string(kept))
}
// A bundle with nothing to patch is left exactly as it was, with no original kept beside it.
func TestPatchBundleLeavesOtherActionsAlone(t *testing.T) {
dir, script := bundleFile(t, `console.log("checkout")`)
original := originalFor(dir, script)
require.NoError(t, patchBundle(script, original))
body, err := os.ReadFile(script)
require.NoError(t, err)
assert.Equal(t, `console.log("checkout")`, string(body))
_, err = os.Stat(original)
assert.True(t, os.IsNotExist(err), "no original is kept for a bundle that was not patched")
}
// bundleFile writes one entrypoint into a fresh action directory.
func bundleFile(t *testing.T, body string) (dir, script string) {
t.Helper()
dir = t.TempDir()
script = filepath.Join(dir, "index.js")
require.NoError(t, os.WriteFile(script, []byte(body), 0o600))
return dir, script
}
func TestActionScriptPaths(t *testing.T) {
node := &model.Action{Runs: model.ActionRuns{Using: "node20", Main: "dist/restore/index.js", Post: "dist/save/index.js"}}
assert.Equal(t, []string{"/a/dist/restore/index.js", "/a/dist/save/index.js"}, actionScriptPaths("/a", node))
// Only a node action has a bundle to patch.
assert.Nil(t, actionScriptPaths("/a", &model.Action{Runs: model.ActionRuns{Using: "docker", Image: "alpine"}}))
assert.Nil(t, actionScriptPaths("/a", nil))
}
// A step that fails with a patched bundle gets the untouched bundle back, and the action is not
// patched again, so later jobs run it exactly as its author shipped it.
func TestRevertToolkit(t *testing.T) {
dir, script := bundleFile(t, gateTSC)
scripts := []string{script}
patchToolkit(t.Context(), dir, scripts)
body, err := os.ReadFile(script)
require.NoError(t, err)
require.True(t, gateOpened(string(body)), "precondition: the bundle is patched")
revertToolkit(t.Context(), dir, scripts)
body, err = os.ReadFile(script)
require.NoError(t, err)
assert.Equal(t, gateTSC, string(body), "the original bundle is back")
// The skip marker survives, so the action stays unpatched from now on.
patchToolkit(t.Context(), dir, scripts)
body, err = os.ReadFile(script)
require.NoError(t, err)
assert.Equal(t, gateTSC, string(body), "a reverted action stays unpatched")
}
// An action whose ref moves is checked out over the patched bundle. The kept original then
// belongs to the version before the move, and must not be restored over the new one.
func TestPatchBundleAfterTheActionMoved(t *testing.T) {
dir, script := bundleFile(t, gateTSC)
original := originalFor(dir, script)
scripts := []string{script}
patchToolkit(t.Context(), dir, scripts)
require.NoError(t, os.WriteFile(script, []byte(gateWebpack), 0o600)) // the new version lands
// Reverting must not roll the action back to the version the original came from.
revertToolkit(t.Context(), dir, scripts)
body, err := os.ReadFile(script)
require.NoError(t, err)
assert.Equal(t, gateWebpack, string(body))
// Nothing was reverted, so the action is not marked off either: the new version is patched
// in its own right, and keeps its own original.
require.NoFileExists(t, filepath.Join(sidecarDir(dir), skipMarker))
require.NoError(t, patchBundle(script, original))
body, err = os.ReadFile(script)
require.NoError(t, err)
assert.True(t, gateOpened(string(body)))
kept, err := os.ReadFile(original)
require.NoError(t, err)
assert.Equal(t, gateWebpack, string(kept))
}
// The wiring: a step patches its own bundles only when the runner serves the v2 API, and a step
// that fails gets them back. The action's path inside its repository is part of where they live.
func TestStepActionRemoteToolkitPatch(t *testing.T) {
newStep := func(t *testing.T, env map[string]string) (*stepActionRemote, string) {
t.Helper()
sar := &stepActionRemote{
Step: &model.Step{Uses: "owner/repo/sub@v1"},
remoteAction: &remoteAction{Org: "owner", Repo: "repo", Path: "sub", Ref: "v1"},
action: &model.Action{Runs: model.ActionRuns{Using: "node20", Main: "index.js"}},
RunContext: &RunContext{
Env: env,
Config: &Config{ActionCacheDir: t.TempDir()},
},
}
script := filepath.Join(sar.actionDir(), "sub", "index.js")
require.NoError(t, os.MkdirAll(filepath.Dir(script), 0o755))
require.NoError(t, os.WriteFile(script, []byte(gateTSC), 0o600))
return sar, script
}
t.Run("left alone when the runner does not serve the v2 API", func(t *testing.T) {
sar, script := newStep(t, map[string]string{})
require.NoError(t, sar.patchActionToolkit(t.Context()))
body, err := os.ReadFile(script)
require.NoError(t, err)
assert.Equal(t, gateTSC, string(body))
})
t.Run("patched, and put back when the step fails", func(t *testing.T) {
sar, script := newStep(t, map[string]string{CacheServiceV2Env: "true"})
require.NoError(t, sar.patchActionToolkit(t.Context()))
body, err := os.ReadFile(script)
require.NoError(t, err)
require.True(t, gateOpened(string(body)))
failed := errors.New("the step failed")
require.ErrorIs(t, sar.revertToolkitOnFailure(func(context.Context) error { return failed })(t.Context()), failed)
body, err = os.ReadFile(script)
require.NoError(t, err)
assert.Equal(t, gateTSC, string(body))
})
}

View File

@@ -141,6 +141,11 @@ func runDaemon(ctx context.Context, daemArgs *daemonArgs, configFile *string) fu
)
runner := run.NewRunner(cfg, reg, cli)
defer func() {
if err := runner.Close(); err != nil {
log.Warnf("runner %s: cache server shutdown: %v", reg.Name, err)
}
}()
// declare the labels of the runner before fetching tasks
resp, err := runner.Declare(ctx, ls.Names())

View File

@@ -132,6 +132,8 @@ func (i *executeArgs) LoadEnvs() map[string]string {
_ = readEnvs(i.Envfile(), envs)
envs["ACTIONS_CACHE_URL"] = i.cacheHandler.ExternalURL() + "/"
// The same server answers the cache service v2 API, so let the actions reach it.
envs[runner.CacheServiceV2Env] = "true"
return envs
}

View File

@@ -109,6 +109,12 @@ func NewRunner(cfg *config.Config, reg *config.Registration, cli client.Client)
}
}
if envs["ACTIONS_CACHE_URL"] != "" && (cfg.Cache.V2 == nil || *cfg.Cache.V2) {
// act patches the GHES check out of an action's bundle when it sees this, so the client
// uses the cache service v2 API this server also answers; see act/runner/toolkit_patch.go.
envs[runner.CacheServiceV2Env] = "true"
}
// set artifact gitea api
artifactGiteaAPI := strings.TrimSuffix(cli.Address(), "/") + "/api/actions_pipeline/"
envs["ACTIONS_RUNTIME_URL"] = artifactGiteaAPI
@@ -132,6 +138,11 @@ func NewRunner(cfg *config.Config, reg *config.Registration, cli client.Client)
return runner
}
// Close shuts down the cache server this runner exposes to job containers.
func (r *Runner) Close() error {
return r.cacheHandler.Close()
}
// removeOrphanNetworks is a variable so tests can substitute one that needs no Docker daemon.
var removeOrphanNetworks = container.RemoveOrphanNetworks

View File

@@ -7,12 +7,14 @@ import (
"context"
"testing"
"gitea.com/gitea/runner/act/runner"
clientmocks "gitea.com/gitea/runner/internal/pkg/client/mocks"
"gitea.com/gitea/runner/internal/pkg/config"
"gitea.com/gitea/runner/internal/pkg/ver"
"connectrpc.com/connect"
runnerv1 "gitea.dev/actions-proto-go/runner/v1"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/mock"
"github.com/stretchr/testify/require"
"google.golang.org/protobuf/types/known/structpb"
@@ -91,6 +93,7 @@ func TestNewRunnerInitializesLabelsAndEnvironment(t *testing.T) {
require.Equal(t, "true", r.envs["GITEA_ACTIONS"])
require.NotEmpty(t, r.envs["GITEA_ACTIONS_RUNNER_VERSION"])
require.Nil(t, r.cacheHandler)
require.Empty(t, r.envs[runner.CacheServiceV2Env], "no cache server, nothing to serve v2 from")
}
// Proxy variables are assembled per task, because a job's service containers have to be
@@ -120,3 +123,22 @@ func taskWithDefaultActionsURL(url string) *runnerv1.Task {
},
}
}
// The cache service v2 API is announced to jobs unless it is turned off. Announcing it is what
// makes act patch the GHES check out of an action's bundle, so the client can reach it.
func TestNewRunnerCacheServiceV2(t *testing.T) {
announced := func(v2 *bool) string {
cfg := &config.Config{}
cfg.Cache.V2, cfg.Cache.Dir, cfg.Cache.Host = v2, t.TempDir(), "127.0.0.1"
cli := clientmocks.NewClient(t)
cli.On("Address").Return("https://gitea.example/").Maybe()
r := NewRunner(cfg, &config.Registration{Name: "runner"}, cli)
t.Cleanup(func() { _ = r.Close() })
return r.envs[runner.CacheServiceV2Env]
}
off := false
assert.Equal(t, "true", announced(nil))
assert.Empty(t, announced(&off))
}

View File

@@ -155,6 +155,11 @@ cache:
# A moved tag (e.g. a re-tagged "v6") or an updated branch stays at the cached commit
# until its cache entry expires or is manually removed.
offline_mode: false
# Serve the actions cache service v2 API, used by actions/cache@v4.2 and later. Those actions
# refuse any host they do not take for GitHub, so reaching it means editing that check out of
# the action's own bundle, keeping the untouched copy beside it. The same edit lets the stock
# upload-artifact and download-artifact work here. A bundle that does not match is left alone.
v2: true
container:
# Specifies the network to which the container will connect.

View File

@@ -74,6 +74,7 @@ type Cache struct {
ExternalSecret string `yaml:"external_secret"` // ExternalSecret is a shared secret between this runner and an external gitea-runner cache-server, enabling per-job ACTIONS_RUNTIME_TOKEN authentication and repo scoping over the network. Required whenever ExternalServer is set; ExternalSecretFile is the alternative way to provide it.
ExternalSecretFile string `yaml:"external_secret_file"` // ExternalSecretFile is the path to a file holding the ExternalSecret value, so the secret can be mounted instead of stored in the config file. LoadDefault reads it into ExternalSecret; setting both is an error.
OfflineMode bool `yaml:"offline_mode"` // OfflineMode reuses a cached action without fetching from the remote; a moved tag or branch stays at the cached commit until the cache entry is removed.
V2 *bool `yaml:"v2"` // V2 serves the actions cache service v2 API to jobs, used by actions/cache@v4.2 and later, and edits the action bundles that would otherwise refuse it. Unset means enabled.
}
// Container represents the configuration for the container.

View File

@@ -347,3 +347,13 @@ cache:
require.Error(t, err)
assert.Contains(t, err.Error(), "contains no secret")
}
// The shipped example must parse, and every key in it must be one the config knows.
func TestLoadDefault_ExampleConfigParses(t *testing.T) {
hook := test.NewGlobal()
defer hook.Reset()
_, err := LoadDefault("config.example.yaml")
require.NoError(t, err)
assert.Empty(t, hook.AllEntries())
}