Create doctor command to fix repo_units broken by dumps from 1.14.3-1.14.6 (#17136)

There was a serious issue with the `gitea dump` command in 1.14.3-1.14.6 which led to corruption of the `config` field of the `repo_unit` table. 

This PR adds a doctor command to attempt to fix the broken repo_units. Users affected by #16961 should run:

```
gitea doctor --fix --run fix-broken-repo-units
```

Fix #16961

Signed-off-by: Andrew Thornton <art27@cantab.net>

4 files changed, 596 insertions, 1 deletions

diff --git a/models/helper.go b/models/helper.go
index c499b5512d..710c15a978 100644
--- a/models/helper.go
+++ b/models/helper.go
@@ -51,7 +51,7 @@ func JSONUnmarshalHandleDoubleEncode(bs []byte, v interface{}) error {
 			rs = append(rs, temp...)
 		}
 		if ok {
-			if rs[0] == 0xff && rs[1] == 0xfe {
+			if len(rs) > 1 && rs[0] == 0xff && rs[1] == 0xfe {
 				rs = rs[2:]
 			}
 			err = json.Unmarshal(rs, v)
diff --git a/models/repo_unit.go b/models/repo_unit.go
index 7061119bd8..474f65bf03 100644
--- a/models/repo_unit.go
+++ b/models/repo_unit.go
@@ -220,3 +220,9 @@ func getUnitsByRepoID(e db.Engine, repoID int64) (units []*RepoUnit, err error)
 
 	return units, nil
 }
+
+// UpdateRepoUnit updates the provided repo unit
+func UpdateRepoUnit(unit *RepoUnit) error {
+	_, err := db.GetEngine(db.DefaultContext).ID(unit.ID).Update(unit)
+	return err
+}
diff --git a/modules/doctor/fix16961.go b/modules/doctor/fix16961.go
new file mode 100644
index 0000000000..60cc5ffe2f
--- /dev/null
+++ b/modules/doctor/fix16961.go
@@ -0,0 +1,318 @@
+// Copyright 2021 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package doctor
+
+import (
+	"bytes"
+	"fmt"
+
+	"code.gitea.io/gitea/models"
+	"code.gitea.io/gitea/models/db"
+	"code.gitea.io/gitea/modules/log"
+	"code.gitea.io/gitea/modules/timeutil"
+	"xorm.io/builder"
+)
+
+// #16831 revealed that the dump command that was broken in 1.14.3-1.14.6 and 1.15.0 (#15885).
+// This led to repo_unit and login_source cfg not being converted to JSON in the dump
+// Unfortunately although it was hoped that there were only a few users affected it
+// appears that many users are affected.
+
+// We therefore need to provide a doctor command to fix this repeated issue #16961
+
+func parseBool16961(bs []byte) (bool, error) {
+	if bytes.EqualFold(bs, []byte("%!s(bool=false)")) {
+		return false, nil
+	}
+
+	if bytes.EqualFold(bs, []byte("%!s(bool=true)")) {
+		return true, nil
+	}
+
+	return false, fmt.Errorf("unexpected bool format: %s", string(bs))
+}
+
+func fixUnitConfig16961(bs []byte, cfg *models.UnitConfig) (fixed bool, err error) {
+	err = models.JSONUnmarshalHandleDoubleEncode(bs, &cfg)
+	if err == nil {
+		return
+	}
+
+	// Handle #16961
+	if string(bs) != "&{}" && len(bs) != 0 {
+		return
+	}
+
+	return true, nil
+}
+
+func fixExternalWikiConfig16961(bs []byte, cfg *models.ExternalWikiConfig) (fixed bool, err error) {
+	err = models.JSONUnmarshalHandleDoubleEncode(bs, &cfg)
+	if err == nil {
+		return
+	}
+
+	if len(bs) < 3 {
+		return
+	}
+	if bs[0] != '&' || bs[1] != '{' || bs[len(bs)-1] != '}' {
+		return
+	}
+	cfg.ExternalWikiURL = string(bs[2 : len(bs)-1])
+	return true, nil
+}
+
+func fixExternalTrackerConfig16961(bs []byte, cfg *models.ExternalTrackerConfig) (fixed bool, err error) {
+	err = models.JSONUnmarshalHandleDoubleEncode(bs, &cfg)
+	if err == nil {
+		return
+	}
+	// Handle #16961
+	if len(bs) < 3 {
+		return
+	}
+
+	if bs[0] != '&' || bs[1] != '{' || bs[len(bs)-1] != '}' {
+		return
+	}
+
+	parts := bytes.Split(bs[2:len(bs)-1], []byte{' '})
+	if len(parts) != 3 {
+		return
+	}
+
+	cfg.ExternalTrackerURL = string(bytes.Join(parts[:len(parts)-2], []byte{' '}))
+	cfg.ExternalTrackerFormat = string(parts[len(parts)-2])
+	cfg.ExternalTrackerStyle = string(parts[len(parts)-1])
+	return true, nil
+}
+
+func fixPullRequestsConfig16961(bs []byte, cfg *models.PullRequestsConfig) (fixed bool, err error) {
+	err = models.JSONUnmarshalHandleDoubleEncode(bs, &cfg)
+	if err == nil {
+		return
+	}
+
+	// Handle #16961
+	if len(bs) < 3 {
+		return
+	}
+
+	if bs[0] != '&' || bs[1] != '{' || bs[len(bs)-1] != '}' {
+		return
+	}
+
+	// PullRequestsConfig was the following in 1.14
+	// type PullRequestsConfig struct {
+	// 	IgnoreWhitespaceConflicts bool
+	// 	AllowMerge                bool
+	// 	AllowRebase               bool
+	// 	AllowRebaseMerge          bool
+	// 	AllowSquash               bool
+	// 	AllowManualMerge          bool
+	// 	AutodetectManualMerge     bool
+	// }
+	//
+	// 1.15 added in addition:
+	// DefaultDeleteBranchAfterMerge bool
+	// DefaultMergeStyle             MergeStyle
+	parts := bytes.Split(bs[2:len(bs)-1], []byte{' '})
+	if len(parts) < 7 {
+		return
+	}
+
+	var parseErr error
+	cfg.IgnoreWhitespaceConflicts, parseErr = parseBool16961(parts[0])
+	if parseErr != nil {
+		return
+	}
+	cfg.AllowMerge, parseErr = parseBool16961(parts[1])
+	if parseErr != nil {
+		return
+	}
+	cfg.AllowRebase, parseErr = parseBool16961(parts[2])
+	if parseErr != nil {
+		return
+	}
+	cfg.AllowRebaseMerge, parseErr = parseBool16961(parts[3])
+	if parseErr != nil {
+		return
+	}
+	cfg.AllowSquash, parseErr = parseBool16961(parts[4])
+	if parseErr != nil {
+		return
+	}
+	cfg.AllowManualMerge, parseErr = parseBool16961(parts[5])
+	if parseErr != nil {
+		return
+	}
+	cfg.AutodetectManualMerge, parseErr = parseBool16961(parts[6])
+	if parseErr != nil {
+		return
+	}
+
+	// 1.14 unit
+	if len(parts) == 7 {
+		return true, nil
+	}
+
+	if len(parts) < 9 {
+		return
+	}
+
+	cfg.DefaultDeleteBranchAfterMerge, parseErr = parseBool16961(parts[7])
+	if parseErr != nil {
+		return
+	}
+
+	cfg.DefaultMergeStyle = models.MergeStyle(string(bytes.Join(parts[8:], []byte{' '})))
+	return true, nil
+}
+
+func fixIssuesConfig16961(bs []byte, cfg *models.IssuesConfig) (fixed bool, err error) {
+	err = models.JSONUnmarshalHandleDoubleEncode(bs, &cfg)
+	if err == nil {
+		return
+	}
+
+	// Handle #16961
+	if len(bs) < 3 {
+		return
+	}
+
+	if bs[0] != '&' || bs[1] != '{' || bs[len(bs)-1] != '}' {
+		return
+	}
+
+	parts := bytes.Split(bs[2:len(bs)-1], []byte{' '})
+	if len(parts) != 3 {
+		return
+	}
+	var parseErr error
+	cfg.EnableTimetracker, parseErr = parseBool16961(parts[0])
+	if parseErr != nil {
+		return
+	}
+	cfg.AllowOnlyContributorsToTrackTime, parseErr = parseBool16961(parts[1])
+	if parseErr != nil {
+		return
+	}
+	cfg.EnableDependencies, parseErr = parseBool16961(parts[2])
+	if parseErr != nil {
+		return
+	}
+	return true, nil
+}
+
+func fixBrokenRepoUnit16961(repoUnit *models.RepoUnit, bs []byte) (fixed bool, err error) {
+	// Shortcut empty or null values
+	if len(bs) == 0 {
+		return false, nil
+	}
+
+	switch models.UnitType(repoUnit.Type) {
+	case models.UnitTypeCode, models.UnitTypeReleases, models.UnitTypeWiki, models.UnitTypeProjects:
+		cfg := &models.UnitConfig{}
+		repoUnit.Config = cfg
+		if fixed, err := fixUnitConfig16961(bs, cfg); !fixed {
+			return false, err
+		}
+	case models.UnitTypeExternalWiki:
+		cfg := &models.ExternalWikiConfig{}
+		repoUnit.Config = cfg
+
+		if fixed, err := fixExternalWikiConfig16961(bs, cfg); !fixed {
+			return false, err
+		}
+	case models.UnitTypeExternalTracker:
+		cfg := &models.ExternalTrackerConfig{}
+		repoUnit.Config = cfg
+		if fixed, err := fixExternalTrackerConfig16961(bs, cfg); !fixed {
+			return false, err
+		}
+	case models.UnitTypePullRequests:
+		cfg := &models.PullRequestsConfig{}
+		repoUnit.Config = cfg
+
+		if fixed, err := fixPullRequestsConfig16961(bs, cfg); !fixed {
+			return false, err
+		}
+	case models.UnitTypeIssues:
+		cfg := &models.IssuesConfig{}
+		repoUnit.Config = cfg
+		if fixed, err := fixIssuesConfig16961(bs, cfg); !fixed {
+			return false, err
+		}
+	default:
+		panic(fmt.Sprintf("unrecognized repo unit type: %v", repoUnit.Type))
+	}
+	return true, nil
+}
+
+func fixBrokenRepoUnits16961(logger log.Logger, autofix bool) error {
+	// RepoUnit describes all units of a repository
+	type RepoUnit struct {
+		ID          int64
+		RepoID      int64
+		Type        models.UnitType
+		Config      []byte
+		CreatedUnix timeutil.TimeStamp `xorm:"INDEX CREATED"`
+	}
+
+	count := 0
+
+	err := db.Iterate(
+		db.DefaultContext,
+		new(RepoUnit),
+		builder.Gt{
+			"id": 0,
+		},
+		func(idx int, bean interface{}) error {
+			unit := bean.(*RepoUnit)
+
+			bs := unit.Config
+			repoUnit := &models.RepoUnit{
+				ID:          unit.ID,
+				RepoID:      unit.RepoID,
+				Type:        unit.Type,
+				CreatedUnix: unit.CreatedUnix,
+			}
+
+			if fixed, err := fixBrokenRepoUnit16961(repoUnit, bs); !fixed {
+				return err
+			}
+
+			count++
+			if !autofix {
+				return nil
+			}
+
+			return models.UpdateRepoUnit(repoUnit)
+		},
+	)
+
+	if err != nil {
+		logger.Critical("Unable to iterate acrosss repounits to fix the broken units: Error %v", err)
+		return err
+	}
+
+	if !autofix {
+		logger.Warn("Found %d broken repo_units", count)
+		return nil
+	}
+	logger.Info("Fixed %d broken repo_units", count)
+
+	return nil
+}
+
+func init() {
+	Register(&Check{
+		Title:     "Check for incorrectly dumped repo_units (See #16961)",
+		Name:      "fix-broken-repo-units",
+		IsDefault: false,
+		Run:       fixBrokenRepoUnits16961,
+		Priority:  7,
+	})
+}
diff --git a/modules/doctor/fix16961_test.go b/modules/doctor/fix16961_test.go
new file mode 100644
index 0000000000..017f585335
--- /dev/null
+++ b/modules/doctor/fix16961_test.go
@@ -0,0 +1,271 @@
+// Copyright 2021 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package doctor
+
+import (
+	"testing"
+
+	"code.gitea.io/gitea/models"
+	"github.com/stretchr/testify/assert"
+)
+
+func Test_fixUnitConfig_16961(t *testing.T) {
+	tests := []struct {
+		name      string
+		bs        string
+		wantFixed bool
+		wantErr   bool
+	}{
+		{
+			name:      "empty",
+			bs:        "",
+			wantFixed: true,
+			wantErr:   false,
+		},
+		{
+			name:      "normal: {}",
+			bs:        "{}",
+			wantFixed: false,
+			wantErr:   false,
+		},
+		{
+			name:      "broken but fixable: &{}",
+			bs:        "&{}",
+			wantFixed: true,
+			wantErr:   false,
+		},
+		{
+			name:      "broken but unfixable: &{asdasd}",
+			bs:        "&{asdasd}",
+			wantFixed: false,
+			wantErr:   true,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			gotFixed, err := fixUnitConfig16961([]byte(tt.bs), &models.UnitConfig{})
+			if (err != nil) != tt.wantErr {
+				t.Errorf("fixUnitConfig_16961() error = %v, wantErr %v", err, tt.wantErr)
+				return
+			}
+			if gotFixed != tt.wantFixed {
+				t.Errorf("fixUnitConfig_16961() = %v, want %v", gotFixed, tt.wantFixed)
+			}
+		})
+	}
+}
+
+func Test_fixExternalWikiConfig_16961(t *testing.T) {
+	tests := []struct {
+		name      string
+		bs        string
+		expected  string
+		wantFixed bool
+		wantErr   bool
+	}{
+		{
+			name:      "normal: {\"ExternalWikiURL\":\"http://someurl\"}",
+			bs:        "{\"ExternalWikiURL\":\"http://someurl\"}",
+			expected:  "http://someurl",
+			wantFixed: false,
+			wantErr:   false,
+		},
+		{
+			name:      "broken: &{http://someurl}",
+			bs:        "&{http://someurl}",
+			expected:  "http://someurl",
+			wantFixed: true,
+			wantErr:   false,
+		},
+		{
+			name:      "broken but unfixable: http://someurl",
+			bs:        "http://someurl",
+			wantFixed: false,
+			wantErr:   true,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			cfg := &models.ExternalWikiConfig{}
+			gotFixed, err := fixExternalWikiConfig16961([]byte(tt.bs), cfg)
+			if (err != nil) != tt.wantErr {
+				t.Errorf("fixExternalWikiConfig_16961() error = %v, wantErr %v", err, tt.wantErr)
+				return
+			}
+			if gotFixed != tt.wantFixed {
+				t.Errorf("fixExternalWikiConfig_16961() = %v, want %v", gotFixed, tt.wantFixed)
+			}
+			if cfg.ExternalWikiURL != tt.expected {
+				t.Errorf("fixExternalWikiConfig_16961().ExternalWikiURL = %v, want %v", cfg.ExternalWikiURL, tt.expected)
+			}
+		})
+	}
+}
+
+func Test_fixExternalTrackerConfig_16961(t *testing.T) {
+	tests := []struct {
+		name      string
+		bs        string
+		expected  models.ExternalTrackerConfig
+		wantFixed bool
+		wantErr   bool
+	}{
+		{
+			name: "normal",
+			bs:   `{"ExternalTrackerURL":"a","ExternalTrackerFormat":"b","ExternalTrackerStyle":"c"}`,
+			expected: models.ExternalTrackerConfig{
+				ExternalTrackerURL:    "a",
+				ExternalTrackerFormat: "b",
+				ExternalTrackerStyle:  "c",
+			},
+			wantFixed: false,
+			wantErr:   false,
+		},
+		{
+			name: "broken",
+			bs:   "&{a b c}",
+			expected: models.ExternalTrackerConfig{
+				ExternalTrackerURL:    "a",
+				ExternalTrackerFormat: "b",
+				ExternalTrackerStyle:  "c",
+			},
+			wantFixed: true,
+			wantErr:   false,
+		},
+		{
+			name:      "broken - too many fields",
+			bs:        "&{a b c d}",
+			wantFixed: false,
+			wantErr:   true,
+		},
+		{
+			name:      "broken - wrong format",
+			bs:        "a b c d}",
+			wantFixed: false,
+			wantErr:   true,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			cfg := &models.ExternalTrackerConfig{}
+			gotFixed, err := fixExternalTrackerConfig16961([]byte(tt.bs), cfg)
+			if (err != nil) != tt.wantErr {
+				t.Errorf("fixExternalTrackerConfig_16961() error = %v, wantErr %v", err, tt.wantErr)
+				return
+			}
+			if gotFixed != tt.wantFixed {
+				t.Errorf("fixExternalTrackerConfig_16961() = %v, want %v", gotFixed, tt.wantFixed)
+			}
+			if cfg.ExternalTrackerFormat != tt.expected.ExternalTrackerFormat {
+				t.Errorf("fixExternalTrackerConfig_16961().ExternalTrackerFormat = %v, want %v", tt.expected.ExternalTrackerFormat, cfg.ExternalTrackerFormat)
+			}
+			if cfg.ExternalTrackerStyle != tt.expected.ExternalTrackerStyle {
+				t.Errorf("fixExternalTrackerConfig_16961().ExternalTrackerStyle = %v, want %v", tt.expected.ExternalTrackerStyle, cfg.ExternalTrackerStyle)
+			}
+			if cfg.ExternalTrackerURL != tt.expected.ExternalTrackerURL {
+				t.Errorf("fixExternalTrackerConfig_16961().ExternalTrackerURL = %v, want %v", tt.expected.ExternalTrackerURL, cfg.ExternalTrackerURL)
+			}
+		})
+	}
+}
+
+func Test_fixPullRequestsConfig_16961(t *testing.T) {
+	tests := []struct {
+		name      string
+		bs        string
+		expected  models.PullRequestsConfig
+		wantFixed bool
+		wantErr   bool
+	}{
+		{
+			name: "normal",
+			bs:   `{"IgnoreWhitespaceConflicts":false,"AllowMerge":false,"AllowRebase":false,"AllowRebaseMerge":false,"AllowSquash":false,"AllowManualMerge":false,"AutodetectManualMerge":false,"DefaultDeleteBranchAfterMerge":false,"DefaultMergeStyle":""}`,
+		},
+		{
+			name: "broken - 1.14",
+			bs:   `&{%!s(bool=false) %!s(bool=true) %!s(bool=true) %!s(bool=true) %!s(bool=true) %!s(bool=false) %!s(bool=false)}`,
+			expected: models.PullRequestsConfig{
+				IgnoreWhitespaceConflicts: false,
+				AllowMerge:                true,
+				AllowRebase:               true,
+				AllowRebaseMerge:          true,
+				AllowSquash:               true,
+				AllowManualMerge:          false,
+				AutodetectManualMerge:     false,
+			},
+			wantFixed: true,
+		},
+		{
+			name: "broken - 1.15",
+			bs:   `&{%!s(bool=false) %!s(bool=true) %!s(bool=true) %!s(bool=true) %!s(bool=true) %!s(bool=false) %!s(bool=false) %!s(bool=false) merge}`,
+			expected: models.PullRequestsConfig{
+				AllowMerge:        true,
+				AllowRebase:       true,
+				AllowRebaseMerge:  true,
+				AllowSquash:       true,
+				DefaultMergeStyle: models.MergeStyleMerge,
+			},
+			wantFixed: true,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			cfg := &models.PullRequestsConfig{}
+			gotFixed, err := fixPullRequestsConfig16961([]byte(tt.bs), cfg)
+			if (err != nil) != tt.wantErr {
+				t.Errorf("fixPullRequestsConfig_16961() error = %v, wantErr %v", err, tt.wantErr)
+				return
+			}
+			if gotFixed != tt.wantFixed {
+				t.Errorf("fixPullRequestsConfig_16961() = %v, want %v", gotFixed, tt.wantFixed)
+			}
+			assert.EqualValues(t, &tt.expected, cfg)
+		})
+	}
+}
+
+func Test_fixIssuesConfig_16961(t *testing.T) {
+	tests := []struct {
+		name      string
+		bs        string
+		expected  models.IssuesConfig
+		wantFixed bool
+		wantErr   bool
+	}{
+		{
+			name: "normal",
+			bs:   `{"EnableTimetracker":true,"AllowOnlyContributorsToTrackTime":true,"EnableDependencies":true}`,
+			expected: models.IssuesConfig{
+				EnableTimetracker:                true,
+				AllowOnlyContributorsToTrackTime: true,
+				EnableDependencies:               true,
+			},
+		},
+		{
+			name: "broken",
+			bs:   `&{%!s(bool=true) %!s(bool=true) %!s(bool=true)}`,
+			expected: models.IssuesConfig{
+				EnableTimetracker:                true,
+				AllowOnlyContributorsToTrackTime: true,
+				EnableDependencies:               true,
+			},
+			wantFixed: true,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			cfg := &models.IssuesConfig{}
+			gotFixed, err := fixIssuesConfig16961([]byte(tt.bs), cfg)
+			if (err != nil) != tt.wantErr {
+				t.Errorf("fixIssuesConfig_16961() error = %v, wantErr %v", err, tt.wantErr)
+				return
+			}
+			if gotFixed != tt.wantFixed {
+				t.Errorf("fixIssuesConfig_16961() = %v, want %v", gotFixed, tt.wantFixed)
+			}
+			assert.EqualValues(t, &tt.expected, cfg)
+		})
+	}
+}