mirror of
https://github.com/seaweedfs/seaweedfs.git
synced 2026-10-12 01:07:35 +02:00
filer: keep S3 list order byte-lexicographic regardless of SQL name column collation (#9824)
* mysql: keep S3 list order byte-lexicographic regardless of name column collation ORDER BY name and the name > ? pagination predicate follow the column collation, so a case-insensitive filemeta.name (e.g. utf8mb3_general_ci) returns S3 keys out of byte order and breaks clients that merge two sorted listings. Detect the live name collation at startup; only when it isn't binary, wrap the list comparison, prefix, and ORDER BY in BINARY name so order and pagination stay consistent. Correctly configured utf8mb4_bin tables keep their indexed range scan unchanged, and the operator gets a warning to convert the column. * postgres: keep S3 list order byte-lexicographic regardless of name column collation ORDER BY name and the name > $n pagination predicate follow the column or database collation, so a locale-aware filemeta.name (e.g. the en_US.UTF-8 database default) returns S3 keys out of byte order and breaks clients that merge two sorted listings. Detect the live name collation at startup; only when it isn't byte-ordered, wrap the list comparison, prefix, and ORDER BY in COLLATE "C" so order and pagination stay consistent. A byte-ordered (C/POSIX/C.UTF-8) column keeps its indexed range scan unchanged, and the operator gets a warning to declare the column COLLATE "C".
This commit is contained in:
1 parent
8c2d9f466f
commit
a24f4844d3
10 files changed
+222
-8
No files matched your search
@@ -23,3 +23,45 @@ func TestEmptyUpsertTemplateFallsBackToPlainInsert(t *testing.T) {
|
||||
t.Fatalf("plain INSERT path should not contain ON CONFLICT, got: %s", got)
|
||||
}
|
||||
}
|
||||
|
||||
func TestListSqlDefaultOrderingFollowsColumn(t *testing.T) {
|
||||
gen := &SqlGenPostgres{}
|
||||
for _, got := range []string{gen.GetSqlListExclusive("filemeta"), gen.GetSqlListInclusive("filemeta")} {
|
||||
if strings.Contains(got, "COLLATE") {
|
||||
t.Fatalf("default list query should not force a collation, got: %s", got)
|
||||
}
|
||||
if !strings.Contains(got, "ORDER BY name ASC") {
|
||||
t.Fatalf("expected plain name ordering, got: %s", got)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func TestListSqlBinaryOrderingOnNonBinaryColumn(t *testing.T) {
|
||||
gen := &SqlGenPostgres{ForceBinaryCollation: true}
|
||||
for _, got := range []string{gen.GetSqlListExclusive("filemeta"), gen.GetSqlListInclusive("filemeta")} {
|
||||
if !strings.Contains(got, `ORDER BY name COLLATE "C" ASC`) {
|
||||
t.Fatalf(`expected COLLATE "C" ordering, got: %s`, got)
|
||||
}
|
||||
if !strings.Contains(got, `name COLLATE "C" like $4`) {
|
||||
t.Fatalf(`expected COLLATE "C" prefix filter, got: %s`, got)
|
||||
}
|
||||
if strings.Contains(got, "AND name>$2") || strings.Contains(got, "AND name>=$2") {
|
||||
t.Fatalf("pagination comparison must also be byte-ordered, got: %s", got)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func TestIsByteOrderedCollation(t *testing.T) {
|
||||
byteOrdered := []string{"C", "POSIX", "c", "C.UTF-8", "C.utf8"}
|
||||
for _, c := range byteOrdered {
|
||||
if !isByteOrderedCollation(c) {
|
||||
t.Fatalf("expected %q to be treated as byte-ordered", c)
|
||||
}
|
||||
}
|
||||
locale := []string{"en_US.UTF-8", "en_US.utf8", "en-US-x-icu", "und-x-icu", ""}
|
||||
for _, c := range locale {
|
||||
if isByteOrderedCollation(c) {
|
||||
t.Fatalf("expected %q to be treated as locale-aware", c)
|
||||
}
|
||||
}
|
||||
}
|
||||
Reference in new issue
Block a user