when demarshalling an object method call fails, the error is weird
Submitted by Simon McVittie
Assigned to Simon McVittie
Description
I wrote a regression test for Bug #80557. That test revealed that when something goes wrong while we demarshal a message (either excessive variant recursion depth or an unsupported type), we report it as o.fd.DBus.GLib.ErrorError.
I think this is unintended: I would expect ErrorError to be reserved for "assertion failure: C code is doing wrong things with GError". However, it's been like this since 2005.