Skip to content

fix(ipc): handle missing flatbuffer tables and dictionary flags #960

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
paleolimbot merged 3 commits into from
Oct 13, 2026
Merged

fix(ipc): handle missing flatbuffer tables and dictionary flags #960

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
15ff47c
fix(ipc): validate optional flatbuffer tables
paleolimbot Oct 9, 2026
7a547a5
fix(ipc): avoid narrowing message type
paleolimbot Oct 10, 2026
37ccebf
fix(ipc): validate dictionary batch vector counts
paleolimbot Oct 11, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
66 changes: 45 additions & 21 deletions src/nanoarrow/ipc/decoder.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1139,26 +1139,30 @@ static int ArrowIpcDecoderSetTypeUnion(struct ArrowSchema* schema,

static int ArrowIpcDecoderSetType(struct ArrowSchema* schema, ns(Field_table_t) field,
int64_t n_children, struct ArrowError* error) {
int type_type = ns(Field_type_type(field));
ns(Type_union_type_t) type_type = ns(Field_type_type(field));
flatbuffers_generic_t type = ns(Field_type_get(field));
if (type == NULL) {
ArrowErrorSet(error, "Field type '%s' has no value", ns(Type_type_name(type_type)));
return EINVAL;
}

switch (type_type) {
case ns(Type_Null):
return ArrowIpcDecoderSetTypeSimple(schema, NANOARROW_TYPE_NA, error);
case ns(Type_Bool):
return ArrowIpcDecoderSetTypeSimple(schema, NANOARROW_TYPE_BOOL, error);
case ns(Type_Int):
return ArrowIpcDecoderSetTypeInt(schema, ns(Field_type_get(field)), error);
return ArrowIpcDecoderSetTypeInt(schema, type, error);
case ns(Type_FloatingPoint):
return ArrowIpcDecoderSetTypeFloatingPoint(schema, ns(Field_type_get(field)),
error);
return ArrowIpcDecoderSetTypeFloatingPoint(schema, type, error);
case ns(Type_Decimal):
return ArrowIpcDecoderSetTypeDecimal(schema, ns(Field_type_get(field)), error);
return ArrowIpcDecoderSetTypeDecimal(schema, type, error);
case ns(Type_Binary):
return ArrowIpcDecoderSetTypeSimple(schema, NANOARROW_TYPE_BINARY, error);
case ns(Type_LargeBinary):
return ArrowIpcDecoderSetTypeSimple(schema, NANOARROW_TYPE_LARGE_BINARY, error);
case ns(Type_FixedSizeBinary):
return ArrowIpcDecoderSetTypeFixedSizeBinary(schema, ns(Field_type_get(field)),
error);
return ArrowIpcDecoderSetTypeFixedSizeBinary(schema, type, error);
case ns(Type_BinaryView):
ArrowErrorSet(error, "BinaryView not yet supported in IPC reader");
return ENOTSUP;
Expand All @@ -1170,33 +1174,31 @@ static int ArrowIpcDecoderSetType(struct ArrowSchema* schema, ns(Field_table_t)
ArrowErrorSet(error, "Utf8View not yet supported in IPC reader");
return ENOTSUP;
case ns(Type_Date):
return ArrowIpcDecoderSetTypeDate(schema, ns(Field_type_get(field)), error);
return ArrowIpcDecoderSetTypeDate(schema, type, error);
case ns(Type_Time):
return ArrowIpcDecoderSetTypeTime(schema, ns(Field_type_get(field)), error);
return ArrowIpcDecoderSetTypeTime(schema, type, error);
case ns(Type_Timestamp):
return ArrowIpcDecoderSetTypeTimestamp(schema, ns(Field_type_get(field)), error);
return ArrowIpcDecoderSetTypeTimestamp(schema, type, error);
case ns(Type_Duration):
return ArrowIpcDecoderSetTypeDuration(schema, ns(Field_type_get(field)), error);
return ArrowIpcDecoderSetTypeDuration(schema, type, error);
case ns(Type_Interval):
return ArrowIpcDecoderSetTypeInterval(schema, ns(Field_type_get(field)), error);
return ArrowIpcDecoderSetTypeInterval(schema, type, error);
case ns(Type_Struct_):
return ArrowIpcDecoderSetTypeSimpleNested(schema, "+s", error);
case ns(Type_List):
return ArrowIpcDecoderSetTypeSimpleNested(schema, "+l", error);
case ns(Type_LargeList):
return ArrowIpcDecoderSetTypeSimpleNested(schema, "+L", error);
case ns(Type_FixedSizeList):
return ArrowIpcDecoderSetTypeFixedSizeList(schema, ns(Field_type_get(field)),
error);
return ArrowIpcDecoderSetTypeFixedSizeList(schema, type, error);
case ns(Type_ListView):
case ns(Type_LargeListView):
ArrowErrorSet(error, "ListView/LargeListView not yet supported in IPC reader");
return ENOTSUP;
case ns(Type_Map):
return ArrowIpcDecoderSetTypeMap(schema, ns(Field_type_get(field)), error);
return ArrowIpcDecoderSetTypeMap(schema, type, error);
case ns(Type_Union):
return ArrowIpcDecoderSetTypeUnion(schema, ns(Field_type_get(field)), n_children,
error);
return ArrowIpcDecoderSetTypeUnion(schema, type, n_children, error);
case ns(Type_RunEndEncoded):
ArrowErrorSet(error, "RunEndEncoded not yet supported in IPC reader");
return ENOTSUP;
Expand Down Expand Up @@ -1229,6 +1231,9 @@ static int ArrowIpcSetDictionaryEncoding(
return EINVAL;
}

int64_t field_flags = schema->flags;
schema->flags |= ARROW_FLAG_NULLABLE;

struct ArrowSchema tmp;
ArrowSchemaMove(schema, &tmp);

Expand Down Expand Up @@ -1259,6 +1264,10 @@ static int ArrowIpcSetDictionaryEncoding(
schema->flags |= ARROW_FLAG_DICTIONARY_ORDERED;
}

if (!(field_flags & ARROW_FLAG_NULLABLE)) {
schema->flags &= ~ARROW_FLAG_NULLABLE;
}

// Sort out field metadata between the schema and the dictionary member
NANOARROW_RETURN_NOT_OK_WITH_ERROR(ArrowIpcMoveDictionaryMetadataBackToField(schema),
error);
Expand Down Expand Up @@ -1424,6 +1433,11 @@ static int ArrowIpcDecoderDecodeDictionaryBatchHeader(
(struct ArrowIpcDecoderPrivate*)decoder->private_data;

ns(DictionaryBatch_table_t) dictionary = (ns(DictionaryBatch_table_t))message_header;
if (!ns(DictionaryBatch_data_is_present(dictionary))) {
ArrowErrorSet(error, "DictionaryBatch has no data");
return EINVAL;
}

private_data->dictionary.id = ns(DictionaryBatch_id(dictionary));
private_data->dictionary.is_delta = ns(DictionaryBatch_isDelta(dictionary));

Expand Down Expand Up @@ -1710,7 +1724,8 @@ ArrowErrorCode ArrowIpcDecoderDecodeHeader(struct ArrowIpcDecoder* decoder,

// Read some basic information from the message
decoder->metadata_version = ns(Message_version(message));
decoder->message_type = ns(Message_header_type(message));
ns(MessageHeader_union_type_t) message_type = ns(Message_header_type(message));
decoder->message_type = message_type;
decoder->body_size_bytes = ns(Message_bodyLength(message));

switch (decoder->metadata_version) {
Expand All @@ -1730,7 +1745,13 @@ ArrowErrorCode ArrowIpcDecoderDecodeHeader(struct ArrowIpcDecoder* decoder,
}

flatbuffers_generic_t message_header = ns(Message_header_get(message));
switch (decoder->message_type) {
if (message_header == NULL) {
ArrowErrorSet(error, "Message header '%s' has no value",
ns(MessageHeader_type_name(message_type)));
return EINVAL;
}

switch (message_type) {
case ns(MessageHeader_Schema):
NANOARROW_RETURN_NOT_OK(
ArrowIpcDecoderDecodeSchemaHeader(decoder, message_header, error));
Expand Down Expand Up @@ -2776,11 +2797,14 @@ static ArrowErrorCode ArrowIpcDecoderDecodeDictionaryInternal(
struct ArrowIpcDecoderPrivate* dictionary_decoder_private_data =
(struct ArrowIpcDecoderPrivate*)dictionary->decoder.private_data;
dictionary->decoder.message_type = NANOARROW_IPC_MESSAGE_TYPE_RECORD_BATCH;
dictionary_decoder_private_data->last_message = record_batch;
dictionary->decoder.metadata_version = decoder->metadata_version;
// Transfer the endianness and compression settings so that buffers are byte-swapped
// and decompressed if needed (the nested decoder uses a default decompressor)
dictionary_decoder_private_data->endianness = private_data->endianness;
dictionary->decoder.codec = decoder->codec;
// Validate the nested batch's node and buffer counts before accessing its vectors.
NANOARROW_RETURN_NOT_OK(
ArrowIpcDecoderDecodeRecordBatchHeader(&dictionary->decoder, record_batch, error));
dictionary_decoder_private_data->last_message = record_batch;

struct ArrowArrayView* array_view;
NANOARROW_RETURN_NOT_OK(ArrowIpcDecoderDecodeArrayViewInternal(
Expand Down
Loading
Loading