fix: stop artifact uploads depending on the cache server reaching Gitea (#1216)

Cache v2 makes the cache server the `ACTIONS_RESULTS_URL` origin, so artifact calls arrived there and were proxied on to Gitea, failing whenever it could not reach the instance.

- Artifact calls are answered with a redirect, so the cache server opens no connection to Gitea. A scheme change or an untrusted instance is still proxied, but there the cache server is the runner itself, which already reaches Gitea.
- Failures answer in twirp, not an empty `502` that clients report as `Unexpected end of JSON input`.
- `cache.v2: false` really points artifacts at Gitea now.
- Cache reservations are bound to the job that made them, so two jobs saving one key cannot commit against each other's upload, and a retry after a lost answer no longer fails a saved entry.
- The toolkit patch, which edits the GitHub-host check out of an action's bundle, was left in the shared checkout where a job running with `runner.patch_actions: false` could inherit it. It is put back after the job's copy.
- `exec` names an origin for the cache v2 it advertises, and masks its runtime token.

Behaviour changes: `no_proxy` no longer exempts `cache.external_server`, and `cache.enabled: false` also stops external registration.

Fixes https://gitea.com/gitea/runner/issues/1208
Fixes https://gitea.com/gitea/runner/issues/1211

Assisted by Claude (Opus 5).

Reviewed-on: https://gitea.com/gitea/runner/pulls/1216
Reviewed-by: bircni <bircni@icloud.com>
Co-authored-by: silverwind <me@silverwind.io>

internal/app/cmd/exec.go

@@ -5,6 +5,7 @@
 package cmd
 
 import (
+	"cmp"
 	"context"
 	"crypto/rand"
 	"encoding/hex"
@@ -187,7 +188,10 @@ func (i *executeArgs) LoadEnvs() map[string]string {
 	envs := parseKVAndFile(i.envs, i.Envfile())
 
 	envs["ACTIONS_CACHE_URL"] = i.cacheHandler.ExternalURL() + "/"
-	// The same server answers the cache service v2 API, so let the actions reach it.
+	// The same server answers cache v2, which docker buildx reads from the results origin alone.
+	if envs["ACTIONS_RESULTS_URL"] == "" {
+		envs["ACTIONS_RESULTS_URL"] = cmp.Or(os.Getenv("ACTIONS_RESULTS_URL"), i.cacheHandler.ExternalURL())
+	}
 	envs[runner.CacheServiceV2Env] = "true"
 
 	return envs
@@ -540,6 +544,7 @@ func runExec(ctx context.Context, execArgs *executeArgs) func(cmd *cobra.Command
 		}
 
 		config.Env["ACT_EXEC"] = "true"
+		config.Secrets[actionsRuntimeTokenEnvName] = actionsRuntimeToken
 
 		if t := config.Secrets["GITEA_TOKEN"]; t != "" {
 			config.Token = t

internal/app/cmd/exec_test.go

@@ -12,9 +12,12 @@ import (
 	"strings"
 	"testing"
 
+	"gitea.com/gitea/runner/act/artifactcache"
+	"gitea.com/gitea/runner/act/runner"
 	"gitea.com/gitea/runner/internal/pkg/config"
 
 	"gitea.dev/actionslib/pkg/model"
+	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"go.yaml.in/yaml/v4"
 )
@@ -307,3 +310,19 @@ func captureStdout(t *testing.T, fn func()) string {
 	require.NoError(t, r.Close())
 	return buf.String()
 }
+
+func TestExecuteArgsLoadEnvsResultsOrigin(t *testing.T) {
+	t.Setenv("ACTIONS_RESULTS_URL", "") // the default is only supplied when nothing else does
+	handler, err := artifactcache.StartHandler(artifactcache.Options{Dir: t.TempDir(), OutboundIP: "127.0.0.1"})
+	require.NoError(t, err)
+	t.Cleanup(func() { _ = handler.Close() })
+	args := &executeArgs{cacheHandler: handler}
+
+	envs := args.LoadEnvs()
+	assert.Equal(t, handler.ExternalURL(), envs["ACTIONS_RESULTS_URL"],
+		"the v2 flag is advertised, so something has to serve that origin")
+	assert.Equal(t, "true", envs[runner.CacheServiceV2Env])
+
+	args.envs = []string{"ACTIONS_RESULTS_URL=https://gitea.example"}
+	assert.Equal(t, "https://gitea.example", args.LoadEnvs()["ACTIONS_RESULTS_URL"], "a supplied origin wins")
+}

internal/app/run/runner.go

@@ -11,6 +11,7 @@ import (
 	"fmt"
 	"maps"
 	"net/http"
+	"net/url"
 	"os"
 	"path/filepath"
 	"runtime"
@@ -89,7 +90,7 @@ func NewRunner(cfg *config.Config, reg *config.Registration, cli client.Client)
 	envs := make(map[string]string, len(cfg.Runner.Envs))
 	maps.Copy(envs, cfg.Runner.Envs)
 	var cacheHandler *artifactcache.Handler
-	if cfg.Cache.Enabled == nil || *cfg.Cache.Enabled {
+	if cacheEnabled(cfg) {
 		if cfg.Cache.ExternalServer != "" {
 			warnIgnoredCachePolicy(cfg)
 			// The v1 client appends its path to this without a separator, so the slash is required.
@@ -109,6 +110,7 @@ func NewRunner(cfg *config.Config, reg *config.Registration, cli client.Client)
 			} else {
 				cacheHandler = handler
 				envs["ACTIONS_CACHE_URL"] = handler.ExternalURL() + "/"
+				warnIfCacheUnreachable(cfg, handler.ExternalURL())
 			}
 		}
 	}
@@ -388,7 +390,7 @@ func (r *Runner) run(ctx context.Context, task *runnerv1.Task, reporter *report.
 
 	// Added per task because this job's service containers must be reached directly, and
 	// act reaches them by their workflow key.
-	proxyEnv := JobProxyEnv(envs, envs["ACTIONS_CACHE_URL"], slices.Sorted(maps.Keys(job.Services)))
+	proxyEnv := JobProxyEnv(envs, r.builtInCacheURL(), slices.Sorted(maps.Keys(job.Services)))
 	maps.Copy(envs, proxyEnv)
 
 	if r.capabilities != "" {
@@ -454,14 +456,7 @@ func (r *Runner) run(ctx context.Context, task *runnerv1.Task, reporter *report.
 	// is that server's responsibility to authenticate requests.
 	revokeCache, resultsURL := r.registerCacheForTask(giteaRuntimeToken, preset.Repository, reporter)
 	defer revokeCache()
-	// A cache server that agreed to forward the artifact half is the whole results service, so the
-	// job is pointed at it.
-	if resultsURL != "" {
-		envs["ACTIONS_RESULTS_URL"] = resultsURL
-		if r.cacheServiceV2() {
-			envs[runner.CacheServiceV2Env] = "true"
-		}
-	}
+	r.setResultsService(envs, resultsURL)
 
 	eventJSON, err := json.Marshal(preset.Event)
 	if err != nil {
@@ -584,12 +579,43 @@ func (r *Runner) run(ctx context.Context, task *runnerv1.Task, reporter *report.
 	return execErr
 }
 
+func (r *Runner) builtInCacheURL() string {
+	if r.cacheHandler == nil {
+		return ""
+	}
+	return r.envs["ACTIONS_CACHE_URL"]
+}
+
+func cacheEnabled(cfg *config.Config) bool {
+	return cfg.Cache.Enabled == nil || *cfg.Cache.Enabled
+}
+
 // cacheServiceV2 reports whether jobs are told the cache service speaks v2. It is all cache.v2
 // turns off: the bundle edit that reaches it is what the artifact actions need too.
 func (r *Runner) cacheServiceV2() bool {
 	return r.cfg.Cache.V2 == nil || *r.cfg.Cache.V2
 }
 
+func (r *Runner) setResultsService(envs map[string]string, resultsURL string) {
+	v2 := resultsURL != "" && r.cacheServiceV2()
+	if v2 {
+		envs[runner.CacheServiceV2Env] = "true"
+	} else {
+		delete(envs, runner.CacheServiceV2Env)
+	}
+	// Cache v2 clients read no other variable for it, and some cannot use the instance address.
+	if v2 || (resultsURL != "" && r.instanceOutOfReach(envs["ACTIONS_RESULTS_URL"])) {
+		envs["ACTIONS_RESULTS_URL"] = resultsURL
+	}
+}
+
+// These clients keep only the origin they are handed, and reach it on the job's own trust.
+func (r *Runner) instanceOutOfReach(instance string) bool {
+	parsed, err := url.Parse(instance)
+	return err != nil || strings.Trim(parsed.Path, "/") != "" ||
+		(r.cfg.Runner.Insecure && parsed.Scheme == "https")
+}
+
 // registerCacheForTask tells the cache server to accept requests authenticated
 // with the given runtime token for the duration of this task. Returns a
 // function the caller must invoke (typically via defer) to revoke the
@@ -616,7 +642,7 @@ func (r *Runner) registerCacheForTask(token, repo string, reporter *report.Repor
 	if r.cacheHandler != nil {
 		return r.cacheHandler.RegisterJob(token, cred), r.cacheHandler.ResultsURL(cred)
 	}
-	if r.cfg.Cache.ExternalServer != "" && r.cfg.Cache.ExternalSecret != "" {
+	if cacheEnabled(r.cfg) && r.cfg.Cache.ExternalServer != "" && r.cfg.Cache.ExternalSecret != "" {
 		return r.registerExternalCacheJob(token, cred, reporter)
 	}
 	// No cache server to register against: caching is disabled, or the built-in server failed to start.
@@ -799,3 +825,13 @@ func warnIgnoredCacheSecret(cfg *config.Config) {
 	}
 	log.Warnf("%s is set but cache.external_server is not; the built-in cache server does not use a shared secret, so the value is ignored", key)
 }
+
+func warnIfCacheUnreachable(cfg *config.Config, cacheURL string) {
+	if cfg.Cache.Host != "" || cfg.Container.Network != "" {
+		return
+	}
+	if _, err := os.Stat("/.dockerenv"); err != nil {
+		return
+	}
+	log.Warnf("jobs are given %s for the cache server; if they cannot reach it, set container.network to a network this runner is on, or cache.host", cacheURL)
+}

internal/app/run/runner_cache_test.go

@@ -194,6 +194,13 @@ func TestRunner_ExternalCacheServer_RegisterRevoke(t *testing.T) {
 	require.Equal(t, http.StatusUnauthorized, probe(),
 		"token must be unknown to the remote server before registration")
 
+	disabled := &Runner{cfg: &config.Config{Cache: config.Cache{
+		Enabled: new(bool), ExternalServer: external, ExternalSecret: secret,
+	}}}
+	disabled.registerCacheForTask(token, repo, nil)
+	require.Equal(t, http.StatusUnauthorized, probe(),
+		"a disabled cache registers nothing, whatever external server is left configured")
+
 	unregister, resultsURL := r.registerCacheForTask(token, repo, nil)
 	require.NotEqual(t, http.StatusUnauthorized, probe(),
 		"token must be accepted after registerCacheForTask")

internal/app/run/runner_test.go

@@ -143,6 +143,7 @@ func TestNewRunnerLeavesProxyToTheTask(t *testing.T) {
 
 	require.NotContains(t, r.envs, "http_proxy")
 	require.NotContains(t, r.envs, "no_proxy")
+	assert.Empty(t, r.builtInCacheURL(), "an external cache server is the operator's to exempt, not ours")
 }
 
 func taskWithDefaultActionsURL(url string) *runnerv1.Task {
@@ -189,11 +190,38 @@ func TestNewRunnerCacheServiceV2(t *testing.T) {
 	defer resp.Body.Close()
 
 	assert.Equal(t, http.StatusOK, resp.StatusCode, "the advertised results service serves no cache service")
+	assert.Equal(t, r.envs["ACTIONS_CACHE_URL"], r.builtInCacheURL(), "the address only the runner knows is bypassed for the operator")
+
+	envs := r.cloneEnvs()
+	r.setResultsService(envs, resultsURL)
+	assert.Equal(t, resultsURL, envs["ACTIONS_RESULTS_URL"])
+	assert.Equal(t, "true", envs[runner.CacheServiceV2Env])
 
-	// Turning v2 off withdraws the advertisement and nothing else.
-	assert.True(t, r.cacheServiceV2())
 	cfg.Cache.V2 = new(bool)
-	assert.False(t, r.cacheServiceV2())
+	envs = r.cloneEnvs()
+	envs[runner.CacheServiceV2Env] = "true"
+	r.setResultsService(envs, resultsURL)
+	assert.Equal(t, "https://gitea.example", envs["ACTIONS_RESULTS_URL"])
+	assert.Empty(t, envs[runner.CacheServiceV2Env], "a runner.envs entry would promise v2 at an origin not serving it")
+
+	envs = r.cloneEnvs()
+	envs[runner.CacheServiceV2Env] = "true"
+	envs["ACTIONS_RESULTS_URL"] = "https://gitea.example/sub"
+	r.setResultsService(envs, "")
+	assert.Equal(t, "https://gitea.example/sub", envs["ACTIONS_RESULTS_URL"], "with no cache server there is nothing to front with")
+	assert.Empty(t, envs[runner.CacheServiceV2Env])
+
+	for instance, insecure := range map[string]bool{
+		"https://gitea.example/sub":   false,
+		"https://self-signed.example": true,
+	} {
+		cfg.Runner.Insecure = insecure
+		envs = r.cloneEnvs()
+		envs["ACTIONS_RESULTS_URL"] = instance
+		r.setResultsService(envs, resultsURL)
+		assert.Equal(t, resultsURL, envs["ACTIONS_RESULTS_URL"], instance)
+		assert.Empty(t, envs[runner.CacheServiceV2Env])
+	}
 }
 
 // The v1 cache client appends its path to ACTIONS_CACHE_URL without a separator, so a configured

internal/pkg/config/config.example.yaml

@@ -139,7 +139,8 @@ runner:
   #  job_completed: ''
 
 cache:
-  # Enable the built-in cache server (used by actions/cache and similar actions).
+  # Enable caching (used by actions/cache and similar actions). Off means no built-in server is
+  # started and no job is registered with an external_server either.
   #enabled: true
   # Directory where cache blobs are stored on disk. Default: $HOME/.cache/actcache
   # Ignored when external_server is set.
@@ -176,9 +177,9 @@ cache:
   # until its cache entry expires or is manually removed.
   #offline_mode: false
   # Serve the actions cache service v2 API. The actions that use it refuse any host they do not
-  # take for GitHub, so reaching it means editing that check out of their own bundle, undone
-  # whenever it is downloaded again. That edit is made either way, this only governs the API
-  # advertised. A bundle that does not match is left alone.
+  # take for GitHub, so reaching it means editing that check out of their own bundle, put back
+  # after the copy into the job. That edit is made either way, this only governs the API
+  # advertised. A bundle that does not match is left alone. With v2, uploads need a reachable cache.
   #v2: true
   # How the cache server discards entries, ignored when external_server is set since that
   # server applies its own. Leave a setting out for its default; 0s or 0 turns the three