From d9a60619679e7918a71f12975a732c4975a8087c Mon Sep 17 00:00:00 2001
From: Minh Vu <vuhoangminh97@gmail.com>
Date: Sat, 27 Jun 2026 16:47:53 +0200
Subject: [PATCH 1/2] fix(table): keep version hint URIs intact in orphan
 cleanup

---
 table/orphan_cleanup.go      | 12 +++++++++++-
 table/orphan_cleanup_test.go | 33 +++++++++++++++++++++++++++++++++
 2 files changed, 44 insertions(+), 1 deletion(-)

diff --git a/table/orphan_cleanup.go b/table/orphan_cleanup.go
index 49d323e1d..e6f5cdc4d 100644
--- a/table/orphan_cleanup.go
+++ b/table/orphan_cleanup.go
@@ -304,7 +304,7 @@ func (t Table) getReferencedFiles(ctx context.Context, fs iceio.IO, maxConcurren
 
 	// Add version hint file (for Hadoop-style tables)
 	// Following Java's ReachableFileUtil.versionHintLocation() logic:
-	versionHintPath := filepath.Join(metadata.Location(), "metadata", "version-hint.text")
+	versionHintPath := versionHintLocation(metadata.Location())
 	referenced[normalizeFilePath(versionHintPath)] = false
 
 	for sf := range metadata.Statistics() {
@@ -597,6 +597,16 @@ func normalizeFilePathWithConfig(path string, cfg *orphanCleanupConfig) string {
 	return normalizeNonURLPath(path)
 }
 
+func versionHintLocation(tableLocation string) string {
+	if strings.Contains(tableLocation, "://") || strings.HasPrefix(tableLocation, "file:") {
+		if joined, err := url.JoinPath(tableLocation, "metadata", "version-hint.text"); err == nil {
+			return joined
+		}
+	}
+
+	return filepath.Join(tableLocation, "metadata", "version-hint.text")
+}
+
 // normalizeURLPath normalizes URL-based file paths with scheme/authority equivalence.
 //
 // This function handles the complexities of cloud storage URIs where the same file
diff --git a/table/orphan_cleanup_test.go b/table/orphan_cleanup_test.go
index 1bf6a7a5d..7957461a7 100644
--- a/table/orphan_cleanup_test.go
+++ b/table/orphan_cleanup_test.go
@@ -22,6 +22,7 @@ import (
 	"errors"
 	"fmt"
 	stdfs "io/fs"
+	"path/filepath"
 	"strings"
 	"testing"
 	"time"
@@ -157,6 +158,36 @@ func TestNormalizeNonURLPath(t *testing.T) {
 	}
 }
 
+func TestVersionHintLocation(t *testing.T) {
+	tests := []struct {
+		name     string
+		location string
+		expected string
+	}{
+		{
+			name:     "s3_uri",
+			location: "s3://bucket/table",
+			expected: "s3://bucket/table/metadata/version-hint.text",
+		},
+		{
+			name:     "file_uri",
+			location: "file:///tmp/table",
+			expected: "file:///tmp/table/metadata/version-hint.text",
+		},
+		{
+			name:     "local_path",
+			location: filepath.Join("local", "table"),
+			expected: filepath.Join("local", "table", "metadata", "version-hint.text"),
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			assert.Equal(t, tt.expected, versionHintLocation(tt.location))
+		})
+	}
+}
+
 func TestApplySchemeEquivalence(t *testing.T) {
 	equalSchemes := map[string]string{
 		"s3,s3a,s3n": "s3",
@@ -532,6 +563,8 @@ func TestGetReferencedFiles_IncludesStatisticsFiles(t *testing.T) {
 	assert.Contains(t, refs, normalizeFilePath("s3://bucket/stats/table-stats.puffin"))
 	assert.Contains(t, refs, normalizeFilePath("s3://bucket/stats/part-stats.puffin"))
 	assert.Contains(t, refs, normalizeFilePath(tbl.metadataLocation))
+	assert.Contains(t, refs, normalizeFilePath("s3://bucket/test/location/metadata/version-hint.text"))
+	assert.NotContains(t, refs, normalizeFilePath("s3:/bucket/test/location/metadata/version-hint.text"))
 	assert.NotContains(t, refs, normalizeFilePath("s3://bucket/stats/not-referenced.puffin"))
 	assert.NotContains(t, refs, "")
 }

From 9bcc275a6d6b49e075b4abbacac1d489aeac29e3 Mon Sep 17 00:00:00 2001
From: Minh Vu <vuhoangminh97@gmail.com>
Date: Sat, 27 Jun 2026 21:51:41 +0200
Subject: [PATCH 2/2] test(table): cover file URI version hint path

---
 table/orphan_cleanup_test.go | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/table/orphan_cleanup_test.go b/table/orphan_cleanup_test.go
index 7957461a7..1bea46804 100644
--- a/table/orphan_cleanup_test.go
+++ b/table/orphan_cleanup_test.go
@@ -174,6 +174,11 @@ func TestVersionHintLocation(t *testing.T) {
 			location: "file:///tmp/table",
 			expected: "file:///tmp/table/metadata/version-hint.text",
 		},
+		{
+			name:     "file_uri_opaque",
+			location: "file:/tmp/table",
+			expected: "file:/tmp/table/metadata/version-hint.text",
+		},
 		{
 			name:     "local_path",
 			location: filepath.Join("local", "table"),