I saw a test failure in main due to the long poll timeout being set to 100 millisec in one of the tests, resulting in a division-by-zero error in ServeShapePlug:
1) test ServeShapePlug sends an up-to-date response after a timeout if no changes are observed (Electric.Plug.ServeShapePlugTest)
test/electric/plug/serve_shape_plug_test.exs:472
** (ArithmeticError) bad argument in arithmetic expression
code: |> ServeShapePlug.call([])
stacktrace:
(electric 0.7.2) lib/electric/plug/serve_shape_plug.ex:33: Electric.Plug.ServeShapePlug.TimeUtils.seconds_since_oct9th_2024_next_interval/1
(electric 0.7.2) lib/electric/plug/serve_shape_plug.ex:354: Electric.Plug.ServeShapePlug.put_resp_cache_headers/2
(electric 0.7.2) lib/electric/plug/serve_shape_plug.ex:1: Electric.Plug.ServeShapePlug.plug_builder_call/2
(electric 0.7.2) deps/plug/lib/plug/error_handler.ex:80: Electric.Plug.ServeShapePlug.call/2
test/electric/plug/serve_shape_plug_test.exs:495: (test)
The following output was logged:
17:04:52.501 [info] Query String: live=true&offset=100_0&root_table=public.users&shape_id=test-shape-id
I don't know what is the significance of 9 Oct 2024. I wonder if it could be better documented in the code to remove any guesswork.
I saw a test failure in
maindue to the long poll timeout being set to 100 millisec in one of the tests, resulting in a division-by-zero error in ServeShapePlug:I don't know what is the significance of 9 Oct 2024. I wonder if it could be better documented in the code to remove any guesswork.