Repository navigation
Support the new database/sql scanning mechanism in Go 1.27. #35
+633
−17
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,153 @@ | ||
| //go:build go1.27 | ||
|
|
||
| package stdlib | ||
|
|
||
| import ( | ||
| "database/sql" | ||
| "database/sql/driver" | ||
| "errors" | ||
| "io" | ||
| "reflect" | ||
| "time" | ||
|
|
||
| "github.com/firebirdsql/fbx/fbtype" | ||
| ) | ||
|
|
||
| // [Rows] implements [driver.RowsColumnScanner] as of Go 1.27. | ||
| var _ driver.RowsColumnScanner = (*Rows)(nil) | ||
|
|
||
| // NextRow implements the [driver.RowsColumnScanner] interface. It advances to the | ||
| // next row of data and returns [io.EOF] when there are no more rows. | ||
| func (r *Rows) NextRow() error { | ||
| var more bool | ||
| if r.skipNext { | ||
| more = r.skipNextMore | ||
| r.skipNext = false | ||
| } else { | ||
| more = r.rows.Next() | ||
| } | ||
|
|
||
| if !more { | ||
| if err := r.rows.Err(); err != nil { | ||
| return err | ||
| } | ||
| return io.EOF | ||
| } | ||
|
|
||
| r.initValueFuncs() | ||
|
|
||
| var err error | ||
| if r.hasBlobCols { | ||
| r.blobValues, err = scanBlobValues(r.rows, r.blobColumns) | ||
| if err != nil { | ||
| return err | ||
| } | ||
| } | ||
|
|
||
| return nil | ||
| } | ||
|
|
||
| // ScanColumn implements the [driver.RowsColumnScanner] interface. It preserves | ||
| // [database/sql] conversions for scalar destinations and [sql.Scanner] implementations. | ||
| // Other destinations, such as Go slices, fbtype.Array, are | ||
| // scanned directly using the fbx type map. | ||
| func (r *Rows) ScanColumn(scanCtx driver.ScanContext, index int, dest any) error { | ||
| if dest == nil { | ||
| return errors.New("destination not a pointer") | ||
| } | ||
| rv := reflect.ValueOf(dest) | ||
| if rv.Kind() == reflect.Pointer && rv.IsNil() { | ||
|
Contributor
There was a problem hiding this comment. This changes the existing database/sql behavior for typed-nil sql.Scanner values: their Scan method used to be called and could handle a nil receiver, but this check rejects them before that call. Is this intentional, or should we preserve the previous behavior? A test would help make the intended behavior explicit. |
||
| return errors.New("destination pointer is nil") | ||
| } | ||
|
|
||
| fd := r.rows.FieldDescriptions()[index] | ||
| typeMetadata := fbtype.NewTypeMetadata( | ||
| fd.Type, | ||
| fd.Subtype, | ||
| fd.Length, | ||
| fd.Scale, | ||
| ) | ||
|
|
||
| arr := fd.ArrayData | ||
| if arr != nil { | ||
| arrayData, err := fbtype.NewArrayData(arr.TypeId, arr.SubType, arr.Length, arr.Scale, arr.Dimensions) | ||
| if err != nil { | ||
| return err | ||
| } | ||
| typeMetadata.SetArrayMetadata(arrayData) | ||
| } | ||
| src := r.rows.RawValues()[index] | ||
|
|
||
| if fd.IsBlob() { | ||
| if src == nil { | ||
| return sql.ConvertAssign(scanCtx, dest, nil) | ||
| } | ||
| blobValue, err := buildBlobValue(r.blobValues[index], &r.rows.FieldDescriptions()[index]) | ||
| if err != nil { | ||
| return err | ||
| } | ||
|
|
||
| return sql.ConvertAssign(scanCtx, dest, blobValue) | ||
| } | ||
|
|
||
| if isSQLScanDestination(rv.Type()) { | ||
| var value driver.Value | ||
| if src != nil { | ||
| var err error | ||
| value, err = r.valueFuncs[index](src) | ||
| if err != nil { | ||
| return err | ||
| } | ||
| } | ||
| return sql.ConvertAssign(scanCtx, dest, value) | ||
| } | ||
|
|
||
| m := r.conn.conn.TypeMap() | ||
|
|
||
| return m.Scan(typeMetadata, src, dest) | ||
| } | ||
|
|
||
| // isSQLScanDestination includes named scalar types and nullable pointers to | ||
| // [database/sql] destinations. Select the conversion before scanning: retrying a | ||
| // failed scan could call user code twice or discard its error. | ||
| func isSQLScanDestination(t reflect.Type) bool { | ||
| var namedPointers map[reflect.Type]bool | ||
| for { | ||
| if t.Implements(reflect.TypeFor[sql.Scanner]()) { | ||
| return true | ||
| } | ||
| // These interfaces distinguish native byte slices, such as | ||
| // FlatArray[byte] and PreallocBytes, from ordinary []byte destinations. | ||
| if t.Implements(reflect.TypeFor[fbtype.ArraySetter]()) || t.Implements(reflect.TypeFor[fbtype.BytesScanner]()) { | ||
| return false | ||
| } | ||
| if t.Kind() != reflect.Pointer { | ||
| break | ||
| } | ||
| // Only defined pointer types can form a cycle without reaching a | ||
| // non-pointer type. Leave those to fbx's bounded scan planning. | ||
| if t.Name() != "" { | ||
| if namedPointers[t] { | ||
| return false | ||
| } | ||
| if namedPointers == nil { | ||
| namedPointers = make(map[reflect.Type]bool) | ||
| } | ||
| namedPointers[t] = true | ||
| } | ||
| t = t.Elem() | ||
| } | ||
|
|
||
| switch t.Kind() { | ||
| case reflect.Bool, reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64, | ||
| reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, | ||
| reflect.Float32, reflect.Float64, reflect.String, reflect.Interface: | ||
| return true | ||
| case reflect.Slice: | ||
| return t.Elem().Kind() == reflect.Uint8 | ||
| case reflect.Struct: | ||
| return t.ConvertibleTo(reflect.TypeFor[time.Time]()) | ||
| default: | ||
| return false | ||
| } | ||
| } | ||
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This branch means the handler was called more than once, which is already a test failure regardless of the error. Why do we expect net.ErrClosed here when the test only passes firstErr and secondErr? Seems we can include err in t.Fatalf instead, so the failure reports the unexpected extra call rather than an unrelated error mismatch. The same applies to the early-return checks in manager_internal_test.go, compression_close_test.go, and client_compression_test.go.