The un-checked seteuid(0) call is apparently a feature. It was copied as is out of the xserver code, and, particularly with the spice server, I assume that (a) the task to be traced is owned by the same user and so (b) uid 0 is not actually necessary.
Furthermore, as Jamie pointed out there is no gstack in the archive :)
So I've pushed a new merge proposal which removes the change to seteuid(0) failure handling.
I will look a bit more into the tests and see if any can be usefully run without snappy.
The un-checked seteuid(0) call is apparently a feature. It was copied as is out of the xserver code, and, particularly with the spice server, I assume that (a) the task to be traced is owned by the same user and so (b) uid 0 is not actually necessary.
Furthermore, as Jamie pointed out there is no gstack in the archive :)
So I've pushed a new merge proposal which removes the change to seteuid(0) failure handling.
I will look a bit more into the tests and see if any can be usefully run without snappy.