Merge pull request '[gitea] week 2024-33 cherry pick (gitea/main -> forgejo)' (#4924) from earl-warren/wcp/2024-33 into forgejo

Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/4924
Reviewed-by: Michael Kriese <michael.kriese@gmx.de>

.deadcode-out

@@ -260,6 +260,11 @@ code.gitea.io/gitea/modules/web
 code.gitea.io/gitea/modules/web/middleware
 	DeleteLocaleCookie
 
+code.gitea.io/gitea/modules/zstd
+	NewWriter
+	Writer.Write
+	Writer.Close
+
 code.gitea.io/gitea/routers/web
 	NotFound
 

custom/conf/app.example.ini

@@ -2712,6 +2712,12 @@ LEVEL = Info
 ;DEFAULT_ACTIONS_URL = https://code.forgejo.org
 ;; Logs retention time in days. Old logs will be deleted after this period.
 ;LOG_RETENTION_DAYS = 365
+;; Log compression type, `none` for no compression, `zstd` for zstd compression.
+;; Other compression types like `gzip` are NOT supported, since seekable stream is required for log view.
+;; It's always recommended to use compression when using local disk as log storage if CPU or memory is not a bottleneck.
+;; And for object storage services like S3, which is billed for requests, it would cause extra 2 times of get requests for each log view.
+;; But it will save storage space and network bandwidth, so it's still recommended to use compression.
+;LOG_COMPRESSION = zstd
 ;; Default artifact retention time in days. Artifacts could have their own retention periods by setting the `retention-days` option in `actions/upload-artifact` step.
 ;ARTIFACT_RETENTION_DAYS = 90
 ;; Timeout to stop the task which have running status, but haven't been updated for a long time

go.mod

@@ -19,6 +19,7 @@ require (
 	github.com/Azure/go-ntlmssp v0.0.0-20221128193559-754e69321358
 	github.com/ProtonMail/go-crypto v1.0.0
 	github.com/PuerkitoBio/goquery v1.9.2
+	github.com/SaveTheRbtz/zstd-seekable-format-go/pkg v0.7.2
 	github.com/alecthomas/chroma/v2 v2.14.0
 	github.com/blakesmith/ar v0.0.0-20190502131153-809d4375e1fb
 	github.com/blevesearch/bleve/v2 v2.4.2
@@ -200,6 +201,7 @@ require (
 	github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da // indirect
 	github.com/golang/protobuf v1.5.4 // indirect
 	github.com/golang/snappy v0.0.4 // indirect
+	github.com/google/btree v1.1.2 // indirect
 	github.com/google/go-cmp v0.6.0 // indirect
 	github.com/google/go-querystring v1.1.0 // indirect
 	github.com/google/go-tpm v0.9.0 // indirect

go.sum

@@ -65,6 +65,8 @@ github.com/PuerkitoBio/goquery v1.9.2 h1:4/wZksC3KgkQw7SQgkKotmKljk0M6V8TUvA8Wb4
 github.com/PuerkitoBio/goquery v1.9.2/go.mod h1:GHPCaP0ODyyxqcNoFGYlAprUFH81NuRPd0GX3Zu2Mvk=
 github.com/RoaringBitmap/roaring v1.9.3 h1:t4EbC5qQwnisr5PrP9nt0IRhRTb9gMUgQF4t4S2OByM=
 github.com/RoaringBitmap/roaring v1.9.3/go.mod h1:6AXUsoIEzDTFFQCe1RbGA6uFONMhvejWj5rqITANK90=
+github.com/SaveTheRbtz/zstd-seekable-format-go/pkg v0.7.2 h1:cSXom2MoKJ9KPPw29RoZtHvUETY4F4n/kXl8m9btnQ0=
+github.com/SaveTheRbtz/zstd-seekable-format-go/pkg v0.7.2/go.mod h1:JitQWJ8JuV4Y87l8VsHiiwhb3cgdyn68mX40s7NT6PA=
 github.com/alecthomas/assert/v2 v2.7.0 h1:QtqSACNS3tF7oasA8CU6A6sXZSBDqnm7RfpLl9bZqbE=
 github.com/alecthomas/assert/v2 v2.7.0/go.mod h1:Bze95FyfUr7x34QZrjL+XP+0qgp/zg8yS+TtBj1WA3k=
 github.com/alecthomas/chroma/v2 v2.2.0/go.mod h1:vf4zrexSH54oEjJ7EdB65tGNHmH3pGZmVkgTP5RHvAs=
@@ -350,6 +352,8 @@ github.com/golang/snappy v0.0.1/go.mod h1:/XxbfmMg8lxefKM7IXC3fBNl/7bRcc72aCRzEW
 github.com/golang/snappy v0.0.2/go.mod h1:/XxbfmMg8lxefKM7IXC3fBNl/7bRcc72aCRzEWrmP2Q=
 github.com/golang/snappy v0.0.4 h1:yAGX7huGHXlcLOEtBnF4w7FQwA26wojNCwOYAEhLjQM=
 github.com/golang/snappy v0.0.4/go.mod h1:/XxbfmMg8lxefKM7IXC3fBNl/7bRcc72aCRzEWrmP2Q=
+github.com/google/btree v1.1.2 h1:xf4v41cLI2Z6FxbKm+8Bu+m8ifhj15JuZ9sa0jZCMUU=
+github.com/google/btree v1.1.2/go.mod h1:qOPhT0dTNdNzV6Z/lhRX0YXUafgPLFUh+gZMl761Gm4=
 github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
 github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
 github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=

models/actions/task.go

@@ -502,7 +502,13 @@ func convertTimestamp(timestamp *timestamppb.Timestamp) timeutil.TimeStamp {
 }
 
 func logFileName(repoFullName string, taskID int64) string {
-	return fmt.Sprintf("%s/%02x/%d.log", repoFullName, taskID%256, taskID)
+	ret := fmt.Sprintf("%s/%02x/%d.log", repoFullName, taskID%256, taskID)
+
+	if setting.Actions.LogCompression.IsZstd() {
+		ret += ".zst"
+	}
+
+	return ret
 }
 
 func getTaskIDFromCache(token string) int64 {

models/issues/pull.go

@@ -163,6 +163,7 @@ type PullRequest struct {
 	Issue                      *Issue `xorm:"-"`
 	Index                      int64
 	RequestedReviewers         []*user_model.User `xorm:"-"`
+	RequestedReviewersTeams    []*org_model.Team  `xorm:"-"`
 	isRequestedReviewersLoaded bool               `xorm:"-"`
 
 	HeadRepoID          int64                  `xorm:"INDEX"`
@@ -303,8 +304,29 @@ func (pr *PullRequest) LoadRequestedReviewers(ctx context.Context) error {
 	}
 	pr.isRequestedReviewersLoaded = true
 	for _, review := range reviews {
+		if review.ReviewerID != 0 {
 			pr.RequestedReviewers = append(pr.RequestedReviewers, review.Reviewer)
 		}
+	}
+
+	return nil
+}
+
+// LoadRequestedReviewersTeams loads the requested reviewers teams.
+func (pr *PullRequest) LoadRequestedReviewersTeams(ctx context.Context) error {
+	reviews, err := GetReviewsByIssueID(ctx, pr.Issue.ID)
+	if err != nil {
+		return err
+	}
+	if err = reviews.LoadReviewersTeams(ctx); err != nil {
+		return err
+	}
+
+	for _, review := range reviews {
+		if review.ReviewerTeamID != 0 {
+			pr.RequestedReviewersTeams = append(pr.RequestedReviewersTeams, review.ReviewerTeam)
+		}
+	}
 
 	return nil
 }

models/issues/review_list.go

@@ -7,6 +7,7 @@ import (
 	"context"
 
 	"code.gitea.io/gitea/models/db"
+	organization_model "code.gitea.io/gitea/models/organization"
 	user_model "code.gitea.io/gitea/models/user"
 	"code.gitea.io/gitea/modules/container"
 	"code.gitea.io/gitea/modules/optional"
@@ -37,6 +38,34 @@ func (reviews ReviewList) LoadReviewers(ctx context.Context) error {
 	return nil
 }
 
+// LoadReviewersTeams loads reviewers teams
+func (reviews ReviewList) LoadReviewersTeams(ctx context.Context) error {
+	reviewersTeamsIDs := make([]int64, 0)
+	for _, review := range reviews {
+		if review.ReviewerTeamID != 0 {
+			reviewersTeamsIDs = append(reviewersTeamsIDs, review.ReviewerTeamID)
+		}
+	}
+
+	teamsMap := make(map[int64]*organization_model.Team, 0)
+	for _, teamID := range reviewersTeamsIDs {
+		team, err := organization_model.GetTeamByID(ctx, teamID)
+		if err != nil {
+			return err
+		}
+
+		teamsMap[teamID] = team
+	}
+
+	for _, review := range reviews {
+		if review.ReviewerTeamID != 0 {
+			review.ReviewerTeam = teamsMap[review.ReviewerTeamID]
+		}
+	}
+
+	return nil
+}
+
 func (reviews ReviewList) LoadIssues(ctx context.Context) error {
 	issueIDs := container.FilterSlice(reviews, func(review *Review) (int64, bool) {
 		return review.IssueID, true

modules/actions/log.go

@@ -15,6 +15,7 @@ import (
 	"code.gitea.io/gitea/models/dbfs"
 	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/storage"
+	"code.gitea.io/gitea/modules/zstd"
 
 	runnerv1 "code.gitea.io/actions-proto-go/runner/v1"
 	"google.golang.org/protobuf/types/known/timestamppb"
@@ -28,6 +29,9 @@ const (
 	defaultBufSize = MaxLineSize
 )
 
+// WriteLogs appends logs to DBFS file for temporary storage.
+// It doesn't respect the file format in the filename like ".zst", since it's difficult to reopen a closed compressed file and append new content.
+// Why doesn't it store logs in object storage directly? Because it's not efficient to append content to object storage.
 func WriteLogs(ctx context.Context, filename string, offset int64, rows []*runnerv1.LogRow) ([]int, error) {
 	flag := os.O_WRONLY
 	if offset == 0 {
@@ -106,6 +110,17 @@ func ReadLogs(ctx context.Context, inStorage bool, filename string, offset, limi
 	return rows, nil
 }
 
+const (
+	// logZstdBlockSize is the block size for zstd compression.
+	// 128KB leads the compression ratio to be close to the regular zstd compression.
+	// And it means each read from the underlying object storage will be at least 128KB*(compression ratio).
+	// The compression ratio is about 30% for text files, so the actual read size is about 38KB, which should be acceptable.
+	logZstdBlockSize = 128 * 1024 // 128KB
+)
+
+// TransferLogs transfers logs from DBFS to object storage.
+// It happens when the file is complete and no more logs will be appended.
+// It respects the file format in the filename like ".zst", and compresses the content if needed.
 func TransferLogs(ctx context.Context, filename string) (func(), error) {
 	name := DBFSPrefix + filename
 	remove := func() {
@@ -119,7 +134,26 @@ func TransferLogs(ctx context.Context, filename string) (func(), error) {
 	}
 	defer f.Close()
 
-	if _, err := storage.Actions.Save(filename, f, -1); err != nil {
+	var reader io.Reader = f
+	if strings.HasSuffix(filename, ".zst") {
+		r, w := io.Pipe()
+		reader = r
+		zstdWriter, err := zstd.NewSeekableWriter(w, logZstdBlockSize)
+		if err != nil {
+			return nil, fmt.Errorf("zstd NewSeekableWriter: %w", err)
+		}
+		go func() {
+			defer func() {
+				_ = w.CloseWithError(zstdWriter.Close())
+			}()
+			if _, err := io.Copy(zstdWriter, f); err != nil {
+				_ = w.CloseWithError(err)
+				return
+			}
+		}()
+	}
+
+	if _, err := storage.Actions.Save(filename, reader, -1); err != nil {
 		return nil, fmt.Errorf("storage save %q: %w", filename, err)
 	}
 	return remove, nil
@@ -150,11 +184,22 @@ func OpenLogs(ctx context.Context, inStorage bool, filename string) (io.ReadSeek
 		}
 		return f, nil
 	}
+
 	f, err := storage.Actions.Open(filename)
 	if err != nil {
 		return nil, fmt.Errorf("storage open %q: %w", filename, err)
 	}
-	return f, nil
+
+	var reader io.ReadSeekCloser = f
+	if strings.HasSuffix(filename, ".zst") {
+		r, err := zstd.NewSeekableReader(f)
+		if err != nil {
+			return nil, fmt.Errorf("zstd NewSeekableReader: %w", err)
+		}
+		reader = r
+	}
+
+	return reader, nil
 }
 
 func FormatLog(timestamp time.Time, content string) string {

modules/git/repo_branch_test.go

@@ -95,3 +95,103 @@ func BenchmarkGetRefsBySha(b *testing.B) {
 	_, _ = bareRepo5.GetRefsBySha("c83380d7056593c51a699d12b9c00627bd5743e9", "")
 	_, _ = bareRepo5.GetRefsBySha("58a4bcc53ac13e7ff76127e0fb518b5262bf09af", "")
 }
+
+func TestRepository_IsObjectExist(t *testing.T) {
+	repo, err := openRepositoryWithDefaultContext(filepath.Join(testReposDir, "repo1_bare"))
+	require.NoError(t, err)
+	defer repo.Close()
+
+	supportShortHash := true
+
+	tests := []struct {
+		name string
+		arg  string
+		want bool
+	}{
+		{
+			name: "empty",
+			arg:  "",
+			want: false,
+		},
+		{
+			name: "branch",
+			arg:  "master",
+			want: false,
+		},
+		{
+			name: "commit hash",
+			arg:  "ce064814f4a0d337b333e646ece456cd39fab612",
+			want: true,
+		},
+		{
+			name: "short commit hash",
+			arg:  "ce06481",
+			want: supportShortHash,
+		},
+		{
+			name: "blob hash",
+			arg:  "153f451b9ee7fa1da317ab17a127e9fd9d384310",
+			want: true,
+		},
+		{
+			name: "short blob hash",
+			arg:  "153f451",
+			want: supportShortHash,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			assert.Equal(t, tt.want, repo.IsObjectExist(tt.arg))
+		})
+	}
+}
+
+func TestRepository_IsReferenceExist(t *testing.T) {
+	repo, err := openRepositoryWithDefaultContext(filepath.Join(testReposDir, "repo1_bare"))
+	require.NoError(t, err)
+	defer repo.Close()
+
+	supportBlobHash := true
+
+	tests := []struct {
+		name string
+		arg  string
+		want bool
+	}{
+		{
+			name: "empty",
+			arg:  "",
+			want: false,
+		},
+		{
+			name: "branch",
+			arg:  "master",
+			want: true,
+		},
+		{
+			name: "commit hash",
+			arg:  "ce064814f4a0d337b333e646ece456cd39fab612",
+			want: true,
+		},
+		{
+			name: "short commit hash",
+			arg:  "ce06481",
+			want: true,
+		},
+		{
+			name: "blob hash",
+			arg:  "153f451b9ee7fa1da317ab17a127e9fd9d384310",
+			want: supportBlobHash,
+		},
+		{
+			name: "short blob hash",
+			arg:  "153f451",
+			want: supportBlobHash,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			assert.Equal(t, tt.want, repo.IsReferenceExist(tt.arg))
+		})
+	}
+}

modules/markup/html.go

@@ -1197,7 +1197,7 @@ func hashCurrentPatternProcessor(ctx *RenderContext, node *html.Node) {
 				})
 			}
 
-			exist = ctx.GitRepo.IsObjectExist(hash)
+			exist = ctx.GitRepo.IsReferenceExist(hash)
 			ctx.ShaExistCache[hash] = exist
 		}
 

modules/packages/conda/metadata.go

@@ -13,8 +13,7 @@ import (
 	"code.gitea.io/gitea/modules/json"
 	"code.gitea.io/gitea/modules/util"
 	"code.gitea.io/gitea/modules/validation"
-
+	"code.gitea.io/gitea/modules/zstd"
-	"github.com/klauspost/compress/zstd"
 )
 
 var (

modules/packages/conda/metadata_test.go

@@ -10,8 +10,9 @@ import (
 	"io"
 	"testing"
 
+	"code.gitea.io/gitea/modules/zstd"
+
 	"github.com/dsnet/compress/bzip2"
-	"github.com/klauspost/compress/zstd"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )

modules/packages/debian/metadata.go

@@ -14,9 +14,9 @@ import (
 
 	"code.gitea.io/gitea/modules/util"
 	"code.gitea.io/gitea/modules/validation"
+	"code.gitea.io/gitea/modules/zstd"
 
 	"github.com/blakesmith/ar"
-	"github.com/klauspost/compress/zstd"
 	"github.com/ulikunitz/xz"
 )
 

modules/packages/debian/metadata_test.go

@@ -10,8 +10,9 @@ import (
 	"io"
 	"testing"
 
+	"code.gitea.io/gitea/modules/zstd"
+
 	"github.com/blakesmith/ar"
-	"github.com/klauspost/compress/zstd"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"github.com/ulikunitz/xz"

modules/setting/actions.go

@@ -15,6 +15,7 @@ var (
 		Enabled               bool
 		LogStorage            *Storage          // how the created logs should be stored
 		LogRetentionDays      int64             `ini:"LOG_RETENTION_DAYS"`
+		LogCompression        logCompression    `ini:"LOG_COMPRESSION"`
 		ArtifactStorage       *Storage          // how the created artifacts should be stored
 		ArtifactRetentionDays int64             `ini:"ARTIFACT_RETENTION_DAYS"`
 		DefaultActionsURL     defaultActionsURL `ini:"DEFAULT_ACTIONS_URL"`
@@ -50,6 +51,20 @@ const (
 	defaultActionsURLSelf    = "self"   // the root URL of the self-hosted instance
 )
 
+type logCompression string
+
+func (c logCompression) IsValid() bool {
+	return c.IsNone() || c.IsZstd()
+}
+
+func (c logCompression) IsNone() bool {
+	return strings.ToLower(string(c)) == "none"
+}
+
+func (c logCompression) IsZstd() bool {
+	return c == "" || strings.ToLower(string(c)) == "zstd"
+}
+
 func loadActionsFrom(rootCfg ConfigProvider) error {
 	sec := rootCfg.Section("actions")
 	err := sec.MapTo(&Actions)
@@ -83,5 +98,9 @@ func loadActionsFrom(rootCfg ConfigProvider) error {
 	Actions.EndlessTaskTimeout = sec.Key("ENDLESS_TASK_TIMEOUT").MustDuration(3 * time.Hour)
 	Actions.AbandonedJobTimeout = sec.Key("ABANDONED_JOB_TIMEOUT").MustDuration(24 * time.Hour)
 
+	if !Actions.LogCompression.IsValid() {
+		return fmt.Errorf("invalid [actions] LOG_COMPRESSION: %q", Actions.LogCompression)
+	}
+
 	return nil
 }

modules/structs/pull.go

@@ -20,6 +20,7 @@ type PullRequest struct {
 	Assignee                *User      `json:"assignee"`
 	Assignees               []*User    `json:"assignees"`
 	RequestedReviewers      []*User    `json:"requested_reviewers"`
+	RequestedReviewersTeams []*Team    `json:"requested_reviewers_teams"`
 	State                   StateType  `json:"state"`
 	Draft                   bool       `json:"draft"`
 	IsLocked                bool       `json:"is_locked"`

modules/zstd/option.go

@@ -0,0 +1,46 @@
+// Copyright 2024 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+package zstd
+
+import "github.com/klauspost/compress/zstd"
+
+type WriterOption = zstd.EOption
+
+var (
+	WithEncoderCRC               = zstd.WithEncoderCRC
+	WithEncoderConcurrency       = zstd.WithEncoderConcurrency
+	WithWindowSize               = zstd.WithWindowSize
+	WithEncoderPadding           = zstd.WithEncoderPadding
+	WithEncoderLevel             = zstd.WithEncoderLevel
+	WithZeroFrames               = zstd.WithZeroFrames
+	WithAllLitEntropyCompression = zstd.WithAllLitEntropyCompression
+	WithNoEntropyCompression     = zstd.WithNoEntropyCompression
+	WithSingleSegment            = zstd.WithSingleSegment
+	WithLowerEncoderMem          = zstd.WithLowerEncoderMem
+	WithEncoderDict              = zstd.WithEncoderDict
+	WithEncoderDictRaw           = zstd.WithEncoderDictRaw
+)
+
+type EncoderLevel = zstd.EncoderLevel
+
+const (
+	SpeedFastest           EncoderLevel = zstd.SpeedFastest
+	SpeedDefault           EncoderLevel = zstd.SpeedDefault
+	SpeedBetterCompression EncoderLevel = zstd.SpeedBetterCompression
+	SpeedBestCompression   EncoderLevel = zstd.SpeedBestCompression
+)
+
+type ReaderOption = zstd.DOption
+
+var (
+	WithDecoderLowmem      = zstd.WithDecoderLowmem
+	WithDecoderConcurrency = zstd.WithDecoderConcurrency
+	WithDecoderMaxMemory   = zstd.WithDecoderMaxMemory
+	WithDecoderDicts       = zstd.WithDecoderDicts
+	WithDecoderDictRaw     = zstd.WithDecoderDictRaw
+	WithDecoderMaxWindow   = zstd.WithDecoderMaxWindow
+	WithDecodeAllCapLimit  = zstd.WithDecodeAllCapLimit
+	WithDecodeBuffersBelow = zstd.WithDecodeBuffersBelow
+	IgnoreChecksum         = zstd.IgnoreChecksum
+)

modules/zstd/zstd.go

@@ -0,0 +1,163 @@
+// Copyright 2024 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+// Package zstd provides a high-level API for reading and writing zstd-compressed data.
+// It supports both regular and seekable zstd streams.
+// It's not a new wheel, but a wrapper around the zstd and zstd-seekable-format-go packages.
+package zstd
+
+import (
+	"errors"
+	"io"
+
+	seekable "github.com/SaveTheRbtz/zstd-seekable-format-go/pkg"
+	"github.com/klauspost/compress/zstd"
+)
+
+type Writer zstd.Encoder
+
+var _ io.WriteCloser = (*Writer)(nil)
+
+// NewWriter returns a new zstd writer.
+func NewWriter(w io.Writer, opts ...WriterOption) (*Writer, error) {
+	zstdW, err := zstd.NewWriter(w, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return (*Writer)(zstdW), nil
+}
+
+func (w *Writer) Write(p []byte) (int, error) {
+	return (*zstd.Encoder)(w).Write(p)
+}
+
+func (w *Writer) Close() error {
+	return (*zstd.Encoder)(w).Close()
+}
+
+type Reader zstd.Decoder
+
+var _ io.ReadCloser = (*Reader)(nil)
+
+// NewReader returns a new zstd reader.
+func NewReader(r io.Reader, opts ...ReaderOption) (*Reader, error) {
+	zstdR, err := zstd.NewReader(r, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return (*Reader)(zstdR), nil
+}
+
+func (r *Reader) Read(p []byte) (int, error) {
+	return (*zstd.Decoder)(r).Read(p)
+}
+
+func (r *Reader) Close() error {
+	(*zstd.Decoder)(r).Close() // no error returned
+	return nil
+}
+
+type SeekableWriter struct {
+	buf []byte
+	n   int
+	w   seekable.Writer
+}
+
+var _ io.WriteCloser = (*SeekableWriter)(nil)
+
+// NewSeekableWriter returns a zstd writer to compress data to seekable format.
+// blockSize is an important parameter, it should be decided according to the actual business requirements.
+// If it's too small, the compression ratio could be very bad, even no compression at all.
+// If it's too large, it could cost more traffic when reading the data partially from underlying storage.
+func NewSeekableWriter(w io.Writer, blockSize int, opts ...WriterOption) (*SeekableWriter, error) {
+	zstdW, err := zstd.NewWriter(nil, opts...)
+	if err != nil {
+		return nil, err
+	}
+
+	seekableW, err := seekable.NewWriter(w, zstdW)
+	if err != nil {
+		return nil, err
+	}
+
+	return &SeekableWriter{
+		buf: make([]byte, blockSize),
+		w:   seekableW,
+	}, nil
+}
+
+func (w *SeekableWriter) Write(p []byte) (int, error) {
+	written := 0
+	for len(p) > 0 {
+		n := copy(w.buf[w.n:], p)
+		w.n += n
+		written += n
+		p = p[n:]
+
+		if w.n == len(w.buf) {
+			if _, err := w.w.Write(w.buf); err != nil {
+				return written, err
+			}
+			w.n = 0
+		}
+	}
+	return written, nil
+}
+
+func (w *SeekableWriter) Close() error {
+	if w.n > 0 {
+		if _, err := w.w.Write(w.buf[:w.n]); err != nil {
+			return err
+		}
+	}
+	return w.w.Close()
+}
+
+type SeekableReader struct {
+	r seekable.Reader
+	c func() error
+}
+
+var _ io.ReadSeekCloser = (*SeekableReader)(nil)
+
+// NewSeekableReader returns a zstd reader to decompress data from seekable format.
+func NewSeekableReader(r io.ReadSeeker, opts ...ReaderOption) (*SeekableReader, error) {
+	zstdR, err := zstd.NewReader(nil, opts...)
+	if err != nil {
+		return nil, err
+	}
+
+	seekableR, err := seekable.NewReader(r, zstdR)
+	if err != nil {
+		return nil, err
+	}
+
+	ret := &SeekableReader{
+		r: seekableR,
+	}
+	if closer, ok := r.(io.Closer); ok {
+		ret.c = closer.Close
+	}
+
+	return ret, nil
+}
+
+func (r *SeekableReader) Read(p []byte) (int, error) {
+	return r.r.Read(p)
+}
+
+func (r *SeekableReader) Seek(offset int64, whence int) (int64, error) {
+	return r.r.Seek(offset, whence)
+}
+
+func (r *SeekableReader) Close() error {
+	return errors.Join(
+		func() error {
+			if r.c != nil {
+				return r.c()
+			}
+			return nil
+		}(),
+		r.r.Close(),
+	)
+}

modules/zstd/zstd_test.go

@@ -0,0 +1,304 @@
+// Copyright 2024 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+package zstd
+
+import (
+	"bytes"
+	"io"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestWriterReader(t *testing.T) {
+	testData := prepareTestData(t, 20_000_000)
+
+	result := bytes.NewBuffer(nil)
+
+	t.Run("regular", func(t *testing.T) {
+		result.Reset()
+		writer, err := NewWriter(result)
+		require.NoError(t, err)
+
+		_, err = io.Copy(writer, bytes.NewReader(testData))
+		require.NoError(t, err)
+		require.NoError(t, writer.Close())
+
+		t.Logf("original size: %d, compressed size: %d, rate: %.2f%%", len(testData), result.Len(), float64(result.Len())/float64(len(testData))*100)
+
+		reader, err := NewReader(result)
+		require.NoError(t, err)
+
+		data, err := io.ReadAll(reader)
+		require.NoError(t, err)
+		require.NoError(t, reader.Close())
+
+		assert.Equal(t, testData, data)
+	})
+
+	t.Run("with options", func(t *testing.T) {
+		result.Reset()
+		writer, err := NewWriter(result, WithEncoderLevel(SpeedBestCompression))
+		require.NoError(t, err)
+
+		_, err = io.Copy(writer, bytes.NewReader(testData))
+		require.NoError(t, err)
+		require.NoError(t, writer.Close())
+
+		t.Logf("original size: %d, compressed size: %d, rate: %.2f%%", len(testData), result.Len(), float64(result.Len())/float64(len(testData))*100)
+
+		reader, err := NewReader(result, WithDecoderLowmem(true))
+		require.NoError(t, err)
+
+		data, err := io.ReadAll(reader)
+		require.NoError(t, err)
+		require.NoError(t, reader.Close())
+
+		assert.Equal(t, testData, data)
+	})
+}
+
+func TestSeekableWriterReader(t *testing.T) {
+	testData := prepareTestData(t, 20_000_000)
+
+	result := bytes.NewBuffer(nil)
+
+	t.Run("regular", func(t *testing.T) {
+		result.Reset()
+		blockSize := 100_000
+
+		writer, err := NewSeekableWriter(result, blockSize)
+		require.NoError(t, err)
+
+		_, err = io.Copy(writer, bytes.NewReader(testData))
+		require.NoError(t, err)
+		require.NoError(t, writer.Close())
+
+		t.Logf("original size: %d, compressed size: %d, rate: %.2f%%", len(testData), result.Len(), float64(result.Len())/float64(len(testData))*100)
+
+		reader, err := NewSeekableReader(bytes.NewReader(result.Bytes()))
+		require.NoError(t, err)
+
+		data, err := io.ReadAll(reader)
+		require.NoError(t, err)
+		require.NoError(t, reader.Close())
+
+		assert.Equal(t, testData, data)
+	})
+
+	t.Run("seek read", func(t *testing.T) {
+		result.Reset()
+		blockSize := 100_000
+
+		writer, err := NewSeekableWriter(result, blockSize)
+		require.NoError(t, err)
+
+		_, err = io.Copy(writer, bytes.NewReader(testData))
+		require.NoError(t, err)
+		require.NoError(t, writer.Close())
+
+		t.Logf("original size: %d, compressed size: %d, rate: %.2f%%", len(testData), result.Len(), float64(result.Len())/float64(len(testData))*100)
+
+		assertReader := &assertReadSeeker{r: bytes.NewReader(result.Bytes())}
+
+		reader, err := NewSeekableReader(assertReader)
+		require.NoError(t, err)
+
+		_, err = reader.Seek(10_000_000, io.SeekStart)
+		require.NoError(t, err)
+
+		data := make([]byte, 1000)
+		_, err = io.ReadFull(reader, data)
+		require.NoError(t, err)
+		require.NoError(t, reader.Close())
+
+		assert.Equal(t, testData[10_000_000:10_000_000+1000], data)
+
+		// Should seek 3 times,
+		// the first two times are for getting the index,
+		// and the third time is for reading the data.
+		assert.Equal(t, 3, assertReader.SeekTimes)
+		// Should read less than 2 blocks,
+		// even if the compression ratio is not good and the data is not in the same block.
+		assert.Less(t, assertReader.ReadBytes, blockSize*2)
+		// Should close the underlying reader if it is Closer.
+		assert.True(t, assertReader.Closed)
+	})
+
+	t.Run("tidy data", func(t *testing.T) {
+		testData := prepareTestData(t, 1000) // data size is less than a block
+
+		result.Reset()
+		blockSize := 100_000
+
+		writer, err := NewSeekableWriter(result, blockSize)
+		require.NoError(t, err)
+
+		_, err = io.Copy(writer, bytes.NewReader(testData))
+		require.NoError(t, err)
+		require.NoError(t, writer.Close())
+
+		t.Logf("original size: %d, compressed size: %d, rate: %.2f%%", len(testData), result.Len(), float64(result.Len())/float64(len(testData))*100)
+
+		reader, err := NewSeekableReader(bytes.NewReader(result.Bytes()))
+		require.NoError(t, err)
+
+		data, err := io.ReadAll(reader)
+		require.NoError(t, err)
+		require.NoError(t, reader.Close())
+
+		assert.Equal(t, testData, data)
+	})
+
+	t.Run("tidy block", func(t *testing.T) {
+		result.Reset()
+		blockSize := 100
+
+		writer, err := NewSeekableWriter(result, blockSize)
+		require.NoError(t, err)
+
+		_, err = io.Copy(writer, bytes.NewReader(testData))
+		require.NoError(t, err)
+		require.NoError(t, writer.Close())
+
+		t.Logf("original size: %d, compressed size: %d, rate: %.2f%%", len(testData), result.Len(), float64(result.Len())/float64(len(testData))*100)
+		// A too small block size will cause a bad compression rate,
+		// even the compressed data is larger than the original data.
+		assert.Greater(t, result.Len(), len(testData))
+
+		reader, err := NewSeekableReader(bytes.NewReader(result.Bytes()))
+		require.NoError(t, err)
+
+		data, err := io.ReadAll(reader)
+		require.NoError(t, err)
+		require.NoError(t, reader.Close())
+
+		assert.Equal(t, testData, data)
+	})
+
+	t.Run("compatible reader", func(t *testing.T) {
+		result.Reset()
+		blockSize := 100_000
+
+		writer, err := NewSeekableWriter(result, blockSize)
+		require.NoError(t, err)
+
+		_, err = io.Copy(writer, bytes.NewReader(testData))
+		require.NoError(t, err)
+		require.NoError(t, writer.Close())
+
+		t.Logf("original size: %d, compressed size: %d, rate: %.2f%%", len(testData), result.Len(), float64(result.Len())/float64(len(testData))*100)
+
+		// It should be able to read the data with a regular reader.
+		reader, err := NewReader(bytes.NewReader(result.Bytes()))
+		require.NoError(t, err)
+
+		data, err := io.ReadAll(reader)
+		require.NoError(t, err)
+		require.NoError(t, reader.Close())
+
+		assert.Equal(t, testData, data)
+	})
+
+	t.Run("wrong reader", func(t *testing.T) {
+		result.Reset()
+
+		// Use a regular writer to compress the data.
+		writer, err := NewWriter(result)
+		require.NoError(t, err)
+
+		_, err = io.Copy(writer, bytes.NewReader(testData))
+		require.NoError(t, err)
+		require.NoError(t, writer.Close())
+
+		t.Logf("original size: %d, compressed size: %d, rate: %.2f%%", len(testData), result.Len(), float64(result.Len())/float64(len(testData))*100)
+
+		// But use a seekable reader to read the data, it should fail.
+		_, err = NewSeekableReader(bytes.NewReader(result.Bytes()))
+		require.Error(t, err)
+	})
+}
+
+// prepareTestData prepares test data to test compression.
+// Random data is not suitable for testing compression,
+// so it collects code files from the project to get enough data.
+func prepareTestData(t *testing.T, size int) []byte {
+	// .../gitea/modules/zstd
+	dir, err := os.Getwd()
+	require.NoError(t, err)
+	// .../gitea/
+	dir = filepath.Join(dir, "../../")
+
+	textExt := []string{".go", ".tmpl", ".ts", ".yml", ".css"} // add more if not enough data collected
+	isText := func(info os.FileInfo) bool {
+		if info.Size() == 0 {
+			return false
+		}
+		for _, ext := range textExt {
+			if strings.HasSuffix(info.Name(), ext) {
+				return true
+			}
+		}
+		return false
+	}
+
+	ret := make([]byte, size)
+	n := 0
+	count := 0
+
+	queue := []string{dir}
+	for len(queue) > 0 && n < size {
+		file := queue[0]
+		queue = queue[1:]
+		info, err := os.Stat(file)
+		require.NoError(t, err)
+		if info.IsDir() {
+			entries, err := os.ReadDir(file)
+			require.NoError(t, err)
+			for _, entry := range entries {
+				queue = append(queue, filepath.Join(file, entry.Name()))
+			}
+			continue
+		}
+		if !isText(info) { // text file only
+			continue
+		}
+		data, err := os.ReadFile(file)
+		require.NoError(t, err)
+		n += copy(ret[n:], data)
+		count++
+	}
+
+	if n < size {
+		require.Failf(t, "Not enough data", "Only %d bytes collected from %d files", n, count)
+	}
+	return ret
+}
+
+type assertReadSeeker struct {
+	r         io.ReadSeeker
+	SeekTimes int
+	ReadBytes int
+	Closed    bool
+}
+
+func (a *assertReadSeeker) Read(p []byte) (int, error) {
+	n, err := a.r.Read(p)
+	a.ReadBytes += n
+	return n, err
+}
+
+func (a *assertReadSeeker) Seek(offset int64, whence int) (int64, error) {
+	a.SeekTimes++
+	return a.r.Seek(offset, whence)
+}
+
+func (a *assertReadSeeker) Close() error {
+	a.Closed = true
+	return nil
+}

release-notes/4924.md

@@ -0,0 +1,2 @@
+fix: [commit](https://codeberg.org/forgejo/forgejo/commit/9812b7af91b69386c5d4c08982aece7bd8f9a174) /repos/{owner}/{repo}/pulls/{index} [requested_reviewers contains null for teams](https://codeberg.org/forgejo/forgejo/issues/4108).
+feat: [commit](https://codeberg.org/forgejo/forgejo/commit/bf7373a2520ae56a1dc00416efa02de9749b63d3) Forgejo Actions logs are compressed by default. It can be disabled by setting `[actions].LOG_COMPRESSION=none`.

services/convert/pull.go

@@ -106,10 +106,25 @@ func ToAPIPullRequest(ctx context.Context, pr *issues_model.PullRequest, doer *u
 		log.Error("LoadRequestedReviewers[%d]: %v", pr.ID, err)
 		return nil
 	}
+	if err = pr.LoadRequestedReviewersTeams(ctx); err != nil {
+		log.Error("LoadRequestedReviewersTeams[%d]: %v", pr.ID, err)
+		return nil
+	}
+
 	for _, reviewer := range pr.RequestedReviewers {
 		apiPullRequest.RequestedReviewers = append(apiPullRequest.RequestedReviewers, ToUser(ctx, reviewer, nil))
 	}
 
+	for _, reviewerTeam := range pr.RequestedReviewersTeams {
+		convertedTeam, err := ToTeam(ctx, reviewerTeam, true)
+		if err != nil {
+			log.Error("LoadRequestedReviewersTeams[%d]: %v", pr.ID, err)
+			return nil
+		}
+
+		apiPullRequest.RequestedReviewersTeams = append(apiPullRequest.RequestedReviewersTeams, convertedTeam)
+	}
+
 	if pr.Issue.ClosedUnix != 0 {
 		apiPullRequest.Closed = pr.Issue.ClosedUnix.AsTimePtr()
 	}

templates/swagger/v1_json.tmpl

@@ -25088,6 +25088,13 @@
           },
           "x-go-name": "RequestedReviewers"
         },
+        "requested_reviewers_teams": {
+          "type": "array",
+          "items": {
+            "$ref": "#/definitions/Team"
+          },
+          "x-go-name": "RequestedReviewersTeams"
+        },
         "review_comments": {
           "description": "number of review comments made on the diff of a PR review (not including comments on commits or issues in a PR)",
           "type": "integer",

tests/integration/api_packages_conda_test.go

@@ -17,10 +17,10 @@ import (
 	"code.gitea.io/gitea/models/unittest"
 	user_model "code.gitea.io/gitea/models/user"
 	conda_module "code.gitea.io/gitea/modules/packages/conda"
+	"code.gitea.io/gitea/modules/zstd"
 	"code.gitea.io/gitea/tests"
 
 	"github.com/dsnet/compress/bzip2"
-	"github.com/klauspost/compress/zstd"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )