Support upload outputs and use needs context on Actions (#24230)

See [Defining outputs for
jobs](https://docs.github.com/en/actions/using-jobs/defining-outputs-for-jobs)
and [Example usage of the needs
context](https://docs.github.com/en/actions/learn-github-actions/contexts#example-usage-of-the-needs-context).

Related to:
- [actions-proto-def
#5](https://gitea.com/gitea/actions-proto-def/pulls/5)
- [act_runner #133](https://gitea.com/gitea/act_runner/pulls/133)

<details>
<summary>Tests & screenshots</summary>

Test workflow file:
```yaml
name: outputs
on: push

jobs:
  job1:
    runs-on: ubuntu-latest
    outputs:
      output1: ${{ steps.step1.outputs.output1 }}
      output2: ${{ steps.step2.outputs.output2 }}
    steps:
      - name: step1
        id: step1
        run: |
          date -Is > output1
          cat output1
          echo "output1=$(cat output1)" >> $GITHUB_OUTPUT
      - name: step2
        id: step2
        run: |
          cat /proc/sys/kernel/random/uuid > output2
          cat output2
          echo "output2=$(cat output2)" >> $GITHUB_OUTPUT
  job2:
    needs: job1
    runs-on: ubuntu-latest
    steps:
      - run: echo ${{ needs.job1.outputs.output1 }}
      - run: echo ${{ needs.job1.outputs.output2 }}
      - run: echo ${{ needs.job1.result }}
```

<img width="397" alt="image"
src="https://user-images.githubusercontent.com/9418365/233313322-903e7ebf-49a7-48e2-8c17-95a4581b3284.png">
<img width="385" alt="image"
src="https://user-images.githubusercontent.com/9418365/233313442-30909135-1711-4b78-a5c6-133fcc79f47c.png">



</details>

---------

Co-authored-by: Giteabot <teabot@gitea.io>

go.mod

@@ -3,7 +3,7 @@ module code.gitea.io/gitea
 go 1.19
 
 require (
-	code.gitea.io/actions-proto-go v0.2.0
+	code.gitea.io/actions-proto-go v0.2.1
 	code.gitea.io/gitea-vet v0.2.2
 	code.gitea.io/sdk/gitea v0.15.1
 	codeberg.org/gusted/mcaptcha v0.0.0-20220723083913-4f3072e1d570

go.sum

@@ -40,8 +40,8 @@ cloud.google.com/go/storage v1.6.0/go.mod h1:N7U0C8pVQ/+NIKOBQyamJIeKQKkZ+mxpohl
 cloud.google.com/go/storage v1.8.0/go.mod h1:Wv1Oy7z6Yz3DshWRJFhqM/UCfaWIRTdp0RXyy7KQOVs=
 cloud.google.com/go/storage v1.10.0/go.mod h1:FLPqc6j+Ki4BU591ie1oL6qBQGu2Bl/tZ9ullr3+Kg0=
 cloud.google.com/go/storage v1.14.0/go.mod h1:GrKmX003DSIwi9o29oFT7YDnHYwZoctc3fOKtUw0Xmo=
-code.gitea.io/actions-proto-go v0.2.0 h1:nYh9nhhfk67YA4wVNLsCzd//RCvXnljwXClJ33+HPVk=
-code.gitea.io/actions-proto-go v0.2.0/go.mod h1:00ys5QDo1iHN1tHNvvddAcy2W/g+425hQya1cCSvq9A=
+code.gitea.io/actions-proto-go v0.2.1 h1:ToMN/8thz2q10TuCq8dL2d8mI+/pWpJcHCvG+TELwa0=
+code.gitea.io/actions-proto-go v0.2.1/go.mod h1:00ys5QDo1iHN1tHNvvddAcy2W/g+425hQya1cCSvq9A=
 code.gitea.io/gitea-vet v0.2.1/go.mod h1:zcNbT/aJEmivCAhfmkHOlT645KNOf9W2KnkLgFjGGfE=
 code.gitea.io/gitea-vet v0.2.2 h1:TEOV/Glf38iGmKzKP0EB++Z5OSL4zGg3RrAvlwaMuvk=
 code.gitea.io/gitea-vet v0.2.2/go.mod h1:zcNbT/aJEmivCAhfmkHOlT645KNOf9W2KnkLgFjGGfE=

models/actions/task_output.go

@@ -0,0 +1,51 @@
+// Copyright 2023 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+package actions
+
+import (
+	"context"
+
+	"code.gitea.io/gitea/models/db"
+)
+
+// ActionTaskOutput represents an output of ActionTask.
+// So the outputs are bound to a task, that means when a completed job has been rerun,
+// the outputs of the job will be reset because the task is new.
+// It's by design, to avoid the outputs of the old task to be mixed with the new task.
+type ActionTaskOutput struct {
+	ID          int64
+	TaskID      int64  `xorm:"INDEX UNIQUE(task_id_output_key)"`
+	OutputKey   string `xorm:"VARCHAR(255) UNIQUE(task_id_output_key)"`
+	OutputValue string `xorm:"MEDIUMTEXT"`
+}
+
+// FindTaskOutputByTaskID returns the outputs of the task.
+func FindTaskOutputByTaskID(ctx context.Context, taskID int64) ([]*ActionTaskOutput, error) {
+	var outputs []*ActionTaskOutput
+	return outputs, db.GetEngine(ctx).Where("task_id=?", taskID).Find(&outputs)
+}
+
+// FindTaskOutputKeyByTaskID returns the keys of the outputs of the task.
+func FindTaskOutputKeyByTaskID(ctx context.Context, taskID int64) ([]string, error) {
+	var keys []string
+	return keys, db.GetEngine(ctx).Table(ActionTaskOutput{}).Where("task_id=?", taskID).Cols("output_key").Find(&keys)
+}
+
+// InsertTaskOutputIfNotExist inserts a new task output if it does not exist.
+func InsertTaskOutputIfNotExist(ctx context.Context, taskID int64, key, value string) error {
+	return db.WithTx(ctx, func(ctx context.Context) error {
+		sess := db.GetEngine(ctx)
+		if exist, err := sess.Exist(&ActionTaskOutput{TaskID: taskID, OutputKey: key}); err != nil {
+			return err
+		} else if exist {
+			return nil
+		}
+		_, err := sess.Insert(&ActionTaskOutput{
+			TaskID:      taskID,
+			OutputKey:   key,
+			OutputValue: value,
+		})
+		return err
+	})
+}

models/migrations/migrations.go

@@ -485,6 +485,8 @@ var migrations = []Migration{
 	NewMigration("Fix incorrect admin team unit access mode", v1_20.FixIncorrectAdminTeamUnitAccessMode),
 	// v253 -> v254
 	NewMigration("Fix ExternalTracker and ExternalWiki accessMode in owner and admin team", v1_20.FixExternalTrackerAndExternalWikiAccessModeInOwnerAndAdminTeam),
+	// v254 -> v255
+	NewMigration("Add ActionTaskOutput table", v1_20.AddActionTaskOutputTable),
 }
 
 // GetCurrentDBVersion returns the current db version

models/migrations/v1_20/v254.go

@@ -0,0 +1,18 @@
+// Copyright 2023 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+package v1_20 //nolint
+
+import (
+	"xorm.io/xorm"
+)
+
+func AddActionTaskOutputTable(x *xorm.Engine) error {
+	type ActionTaskOutput struct {
+		ID          int64
+		TaskID      int64  `xorm:"INDEX UNIQUE(task_id_output_key)"`
+		OutputKey   string `xorm:"VARCHAR(255) UNIQUE(task_id_output_key)"`
+		OutputValue string `xorm:"MEDIUMTEXT"`
+	}
+	return x.Sync(new(ActionTaskOutput))
+}

routers/api/actions/runner/runner.go

@@ -97,7 +97,7 @@ func (s *Service) Register(
 // FetchTask assigns a task to the runner
 func (s *Service) FetchTask(
 	ctx context.Context,
-	req *connect.Request[runnerv1.FetchTaskRequest],
+	_ *connect.Request[runnerv1.FetchTaskRequest],
 ) (*connect.Response[runnerv1.FetchTaskResponse], error) {
 	runner := GetRunner(ctx)
 
@@ -145,6 +145,31 @@ func (s *Service) UpdateTask(
 		return nil, status.Errorf(codes.Internal, "update task: %v", err)
 	}
 
+	for k, v := range req.Msg.Outputs {
+		if len(k) > 255 {
+			log.Warn("Ignore the output of task %d because the key is too long: %q", task.ID, k)
+			continue
+		}
+		// The value can be a maximum of 1 MB
+		if l := len(v); l > 1024*1024 {
+			log.Warn("Ignore the output %q of task %d because the value is too long: %v", k, task.ID, l)
+			continue
+		}
+		// There's another limitation on GitHub that the total of all outputs in a workflow run can be a maximum of 50 MB.
+		// We don't check the total size here because it's not easy to do, and it doesn't really worth it.
+		// See https://docs.github.com/en/actions/using-jobs/defining-outputs-for-jobs
+
+		if err := actions_model.InsertTaskOutputIfNotExist(ctx, task.ID, k, v); err != nil {
+			log.Warn("Failed to insert the output %q of task %d: %v", k, task.ID, err)
+			// It's ok not to return errors, the runner will resend the outputs.
+		}
+	}
+	sentOutputs, err := actions_model.FindTaskOutputKeyByTaskID(ctx, task.ID)
+	if err != nil {
+		log.Warn("Failed to find the sent outputs of task %d: %v", task.ID, err)
+		// It's not to return errors, it can be handled when the runner resends sent outputs.
+	}
+
 	if err := task.LoadJob(ctx); err != nil {
 		return nil, status.Errorf(codes.Internal, "load job: %v", err)
 	}
@@ -162,6 +187,7 @@ func (s *Service) UpdateTask(
 			Id:     req.Msg.State.Id,
 			Result: task.Status.AsResult(),
 		},
+		SentOutputs: sentOutputs,
 	}), nil
 }
 

routers/api/actions/runner/utils.go

@@ -37,6 +37,17 @@ func pickTask(ctx context.Context, runner *actions_model.ActionRunner) (*runnerv
 		Context:         generateTaskContext(t),
 		Secrets:         getSecretsOfTask(ctx, t),
 	}
+
+	if needs, err := findTaskNeeds(ctx, t); err != nil {
+		log.Error("Cannot find needs for task %v: %v", t.ID, err)
+		// Go on with empty needs.
+		// If return error, the task will be wild, which means the runner will never get it when it has been assigned to the runner.
+		// In contrast, missing needs is less serious.
+		// And the task will fail and the runner will report the error in the logs.
+	} else {
+		task.Needs = needs
+	}
+
 	return task, true, nil
 }
 
@@ -124,3 +135,46 @@ func generateTaskContext(t *actions_model.ActionTask) *structpb.Struct {
 
 	return taskContext
 }
+
+func findTaskNeeds(ctx context.Context, task *actions_model.ActionTask) (map[string]*runnerv1.TaskNeed, error) {
+	if err := task.LoadAttributes(ctx); err != nil {
+		return nil, fmt.Errorf("LoadAttributes: %w", err)
+	}
+	if len(task.Job.Needs) == 0 {
+		return nil, nil
+	}
+	needs := map[string]struct{}{}
+	for _, v := range task.Job.Needs {
+		needs[v] = struct{}{}
+	}
+
+	jobs, _, err := actions_model.FindRunJobs(ctx, actions_model.FindRunJobOptions{RunID: task.Job.RunID})
+	if err != nil {
+		return nil, fmt.Errorf("FindRunJobs: %w", err)
+	}
+
+	ret := make(map[string]*runnerv1.TaskNeed, len(needs))
+	for _, job := range jobs {
+		if _, ok := needs[job.JobID]; !ok {
+			continue
+		}
+		if job.TaskID == 0 || !job.Status.IsDone() {
+			// it shouldn't happen, or the job has been rerun
+			continue
+		}
+		outputs := make(map[string]string)
+		got, err := actions_model.FindTaskOutputByTaskID(ctx, job.TaskID)
+		if err != nil {
+			return nil, fmt.Errorf("FindTaskOutputByTaskID: %w", err)
+		}
+		for _, v := range got {
+			outputs[v.OutputKey] = v.OutputValue
+		}
+		ret[job.JobID] = &runnerv1.TaskNeed{
+			Outputs: outputs,
+			Result:  runnerv1.Result(job.Status),
+		}
+	}
+
+	return ret, nil
+}