pitrou commented on a change in pull request #11220: URL: https://github.com/apache/arrow/pull/11220#discussion_r716743132
########## File path: go/arrow/cdata/cdata_exports.go ########## @@ -0,0 +1,390 @@ +// Licensed to the Apache Software Foundation (ASF) under one +// or more contributor license agreements. See the NOTICE file +// distributed with this work for additional information +// regarding copyright ownership. The ASF licenses this file +// to you under the Apache License, Version 2.0 (the +// "License"); you may not use this file except in compliance +// with the License. You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package cdata + +// #include <stdlib.h> +// #include "arrow/c/abi.h" +// #include "arrow/c/helpers.h" +// +// extern void releaseExportedSchema(struct ArrowSchema* schema); +// extern void releaseExportedArray(struct ArrowArray* array); +// +// void goReleaseArray(struct ArrowArray* array) { +// releaseExportedArray(array); +// } +// void goReleaseSchema(struct ArrowSchema* schema) { +// releaseExportedSchema(schema); +// } +import "C" + +import ( + "bytes" + "encoding/binary" + "fmt" + "reflect" + "strings" + "unsafe" + + "github.com/apache/arrow/go/arrow" + "github.com/apache/arrow/go/arrow/array" + "github.com/apache/arrow/go/arrow/ipc" +) + +func encodeCMetadata(keys, values []string) []byte { + if len(keys) != len(values) { + panic("unequal metadata key/values length") + } + npairs := len(keys) + + var b bytes.Buffer + totalSize := 4 + for i := range keys { + totalSize += 8 + len(keys[i]) + len(values[i]) + } + b.Grow(totalSize) + + binary.Write(&b, binary.LittleEndian, int32(npairs)) Review comment: Hmm, this should use native endianness. See examples here: https://github.com/apache/arrow/blob/master/cpp/src/arrow/c/bridge_test.cc#L133-L196 ########## File path: go/arrow/cdata/exports.go ########## @@ -0,0 +1,104 @@ +// Licensed to the Apache Software Foundation (ASF) under one +// or more contributor license agreements. See the NOTICE file +// distributed with this work for additional information +// regarding copyright ownership. The ASF licenses this file +// to you under the Apache License, Version 2.0 (the +// "License"); you may not use this file except in compliance +// with the License. You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package cdata + +import ( + "reflect" + "sync" + "sync/atomic" + "unsafe" + + "github.com/apache/arrow/go/arrow/array" +) + +// #include <stdlib.h> +// #include "arrow/c/helpers.h" +import "C" + +var ( + handles = sync.Map{} + handleIdx uintptr +) + +type dataHandle uintptr + +func storeData(d *array.Data) dataHandle { + h := atomic.AddUintptr(&handleIdx, 1) + if h == 0 { + panic("cgo: ran out of space") + } + d.Retain() + handles.Store(h, d) + return dataHandle(h) +} + +func (d dataHandle) releaseData() { + arrd, ok := handles.LoadAndDelete(uintptr(d)) + if !ok { + panic("cgo: invalid datahandle") + } + arrd.(*array.Data).Release() +} + +//export releaseExportedSchema +func releaseExportedSchema(schema *CArrowSchema) { + if C.ArrowSchemaIsReleased(schema) == 1 { + return + } + defer C.ArrowSchemaMarkReleased(schema) + + C.free(unsafe.Pointer(schema.name)) + C.free(unsafe.Pointer(schema.format)) + C.free(unsafe.Pointer(schema.metadata)) + + var children []*CArrowSchema + s := (*reflect.SliceHeader)(unsafe.Pointer(&children)) + s.Data = uintptr(unsafe.Pointer(schema.children)) + s.Len = int(schema.n_children) + s.Cap = int(schema.n_children) + + for _, c := range children { + C.ArrowSchemaRelease(c) + } Review comment: It seems you should also `C.free(schema.children[0])` and then `C.free(schema.children)`, no? ########## File path: go/arrow/cdata/test/test_export_to_cgo.py ########## @@ -91,5 +93,22 @@ def test_record_batch(self): cgotest.importRecordBatch(self.ptr_schema, self.ptr_array) +class TestGoToPython(BaseTestGoPython): + + def test_get_schema(self): + with self.assert_pyarrow_memory_released(): + cgotest.exportSchema(self.ptr_schema) + + sc = pa.Schema._import_from_c(self.ptr_schema) + assert sc == self.make_schema() + + def test_get_batch(self): + with self.assert_pyarrow_memory_released(): + cgotest.exportRecordBatch(self.ptr_schema, self.ptr_array) + arrnew = pa.RecordBatch._import_from_c(self.ptr_array, self.ptr_schema) + assert arrnew == self.make_batch() + + Review comment: I would also suggest to add round-tripping tests (for example Python to Go to Python). ########## File path: go/arrow/cdata/test/test_export_to_cgo.py ########## @@ -91,5 +93,22 @@ def test_record_batch(self): cgotest.importRecordBatch(self.ptr_schema, self.ptr_array) +class TestGoToPython(BaseTestGoPython): + + def test_get_schema(self): + with self.assert_pyarrow_memory_released(): + cgotest.exportSchema(self.ptr_schema) + + sc = pa.Schema._import_from_c(self.ptr_schema) + assert sc == self.make_schema() + + def test_get_batch(self): + with self.assert_pyarrow_memory_released(): Review comment: Is there also a way to check that all Go memory was released? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
