fix(models): chunk commit status queries to avoid SQLite expression depth limit (#39611) (#39638)

Backport #39611 by @Kshot3000

* Fixes #39606
* Fixes #39329

`GetLatestCommitStatusForRepoCommitIDs` built one `OR` condition per
commit SHA. Rendering a commit list of more than ~1000 commits (for
example a large pull request, via `processGitCommits`) therefore failed
on SQLite with `SQL logic error: Expression tree is too large (maximum
depth 1000)`.

The max-index lookup now queries SHAs in chunks of 500 with `IN`, and
the follow-up status fetch runs in chunks of 100 `(index, sha)` pairs,
so the expression depth stays bounded regardless of the commit count.
Results are unchanged.

Verified with a new regression test that calls the function with 2001
SHAs: it fails with the reported error on `main` and passes with this
change. `go test ./models/git/` passes.

—
Fix offered freely; tips welcome: PayPal kyleblake0659@gmail.com · BTC
3GnR7TWBXAB3pPztBWpNF4LMNEX5yX8vZK

Co-authored-by: KShot <kshot9000@gmail.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>

models/git/commit_status.go

@@ -9,6 +9,7 @@ import (
 	"errors"
 	"fmt"
 	"net/url"
+	"slices"
 	"strconv"
 	"strings"
 	"time"
@@ -35,11 +36,11 @@ import (
 // CommitStatus holds a single Status of a single Commit
 type CommitStatus struct {
 	ID     int64                          `xorm:"pk autoincr"`
+	RepoID int64                          `xorm:"UNIQUE(repo_sha_index)"`
+	SHA    string                         `xorm:"VARCHAR(64) NOT NULL INDEX UNIQUE(repo_sha_index)"`
 	Index  int64                          `xorm:"INDEX UNIQUE(repo_sha_index)"`
-	RepoID int64                          `xorm:"INDEX UNIQUE(repo_sha_index)"`
 	Repo   *repo_model.Repository         `xorm:"-"`
 	State  commitstatus.CommitStatusState `xorm:"VARCHAR(7) NOT NULL"`
-	SHA    string                         `xorm:"VARCHAR(64) NOT NULL INDEX UNIQUE(repo_sha_index)"`
 
 	// TargetURL points to the commit status page reported by a CI system
 	// If Gitea Actions is used, it is a relative link like "{RepoLink}/actions/runs/{RunID}/jobs{JobID}"
@@ -420,46 +421,48 @@ func GetLatestCommitStatusForPairs(ctx context.Context, repoSHAs []RepoSHA) (map
 	return repoStatuses, nil
 }
 
-// GetLatestCommitStatusForRepoCommitIDs returns all statuses with a unique context for a given list of repo-sha pairs
-func GetLatestCommitStatusForRepoCommitIDs(ctx context.Context, repoID int64, commitIDs []string) (map[string][]*CommitStatus, error) {
-	type result struct {
+// GetLatestCommitStatusForRepoCommitIDs returns the commit statuses with a unique context for a given list of repo-sha pairs
+// If the provided commit IDs are too many, only the first part and the last part of the commit IDs will be queried.
+func GetLatestCommitStatusForRepoCommitIDs(ctx context.Context, repoID int64, allCommitIDs []string) (map[string][]*CommitStatus, error) {
+	const maxCommitIDs = 500
+	const maxBatchSize = 200
+	queryCommitIDs := allCommitIDs
+	if len(allCommitIDs) > maxCommitIDs {
+		// The commit IDs are usually from "commits list" or "compare" page (create a PR or compare commits), nobody can read so many commits at once.
+		// The commit IDs are usually sorted by time, so we can take the first half and the last half of the commit IDs to get the latest statuses.
+		log.Warn("GetLatestCommitStatusForRepoCommitIDs: too many commit IDs (%d) for repo %d, truncating to %d", len(allCommitIDs), repoID, maxCommitIDs)
+		queryCommitIDs = allCommitIDs[:maxCommitIDs/2]
+		queryCommitIDs = append(queryCommitIDs, allCommitIDs[len(allCommitIDs)-maxCommitIDs/2:]...)
+	}
+
+	baseSql := func() db.Session {
+		return db.GetEngine(ctx).Table(&CommitStatus{}).Where("repo_id = ?", repoID)
+	}
+
+	type shaMaxIndexResult struct {
 		Index int64
 		SHA   string
 	}
-
-	getBase := func() db.Session {
-		return db.GetEngine(ctx).Table(&CommitStatus{}).Where("repo_id = ?", repoID)
-	}
-	results := make([]result, 0, len(commitIDs))
-
-	conds := make([]builder.Cond, 0, len(commitIDs))
-	for _, sha := range commitIDs {
-		conds = append(conds, builder.Eq{"sha": sha})
-	}
-	sess := getBase().And(builder.Or(conds...)).
-		Select("max( `index` ) as `index`, sha").
-		GroupBy("context_hash, sha").OrderBy("max( `index` ) desc")
-
-	err := sess.Find(&results)
+	shaMaxIndexResults := make([]*shaMaxIndexResult, 0, len(allCommitIDs))
+	err := baseSql().And(builder.In("sha", queryCommitIDs)).
+		Select("max(`index`) as `index`, sha").
+		GroupBy("context_hash, sha").
+		Find(&shaMaxIndexResults)
 	if err != nil {
 		return nil, err
 	}
 
 	repoStatuses := make(map[string][]*CommitStatus)
-
-	if len(results) > 0 {
-		statuses := make([]*CommitStatus, 0, len(results))
-
-		conds = make([]builder.Cond, 0, len(results))
-		for _, result := range results {
-			conds = append(conds, builder.Eq{"`index`": result.Index, "sha": result.SHA})
+	for chunk := range slices.Chunk(shaMaxIndexResults, maxBatchSize) {
+		statuses := make([]*CommitStatus, 0, len(chunk))
+		condIndexSha := make([]builder.Cond, 0, len(chunk))
+		for _, res := range chunk {
+			condIndexSha = append(condIndexSha, builder.Eq{"`index`": res.Index, "sha": res.SHA})
 		}
-		err = getBase().And(builder.Or(conds...)).Find(&statuses)
+		err = baseSql().And(builder.Or(condIndexSha...)).Find(&statuses)
 		if err != nil {
 			return nil, err
 		}
-
-		// Group the statuses by commit
 		for _, status := range statuses {
 			repoStatuses[status.SHA] = append(repoStatuses[status.SHA], status)
 		}