This is an automated email from the ASF dual-hosted git repository.
kou pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow.git
The following commit(s) were added to refs/heads/main by this push:
new 38c12865ae GH-44337: [CI][GLib] Fix a flaky StreamDecoder and Buffer
test (#44341)
38c12865ae is described below
commit 38c12865ae17b4b4f1c9d5f0789acbf3ca38a243
Author: Sutou Kouhei <[email protected]>
AuthorDate: Wed Oct 9 09:25:44 2024 +0900
GH-44337: [CI][GLib] Fix a flaky StreamDecoder and Buffer test (#44341)
### Rationale for this change
It's related to GC.
StreamDecoder accepts incomplete data. They are kept until enough data are
provided. A caller must not release the incomplete data before they are
processed. If they are released, StreamDecoder may touch unexpected data.
### What changes are included in this PR?
Refer unprocessed data until they are processed.
### Are these changes tested?
Yes.
### Are there any user-facing changes?
No.
* GitHub Issue: #44337
Authored-by: Sutou Kouhei <[email protected]>
Signed-off-by: Sutou Kouhei <[email protected]>
---
c_glib/test/test-stream-decoder.rb | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/c_glib/test/test-stream-decoder.rb
b/c_glib/test/test-stream-decoder.rb
index 108e687e3a..ef669a61f0 100644
--- a/c_glib/test/test-stream-decoder.rb
+++ b/c_glib/test/test-stream-decoder.rb
@@ -79,8 +79,15 @@ class TestStreamDecoder < Test::Unit::TestCase
end
def test_consume_buffer
+ # We need to keep data that aren't processed yet.
+ data = []
@buffer.data.to_s.each_byte do |byte|
- @decoder.consume_buffer(Arrow::Buffer.new(byte.chr))
+ data << byte.chr
+ can_clear = (@decoder.next_required_size == 1)
+ @decoder.consume_buffer(Arrow::Buffer.new(data.last))
+ # We can release a reference for kept data after they are
+ # processed.
+ data.clear if can_clear
end
assert_equal([
[:schema_decoded, @schema, @schema],
