Bug #17020: ObjectSpace.trace_object_allocations_stop raises if called before trace_object_allocations_start

• Status: Closed
• Assignee: byroot (Jean Boussier)
• Target version: -
• ruby -v: ruby 2.5.8p224 (2020-03-31 revision 67882) [x86_64-darwin19]
• Backport: 2.5: UNKNOWN, 2.6: UNKNOWN, 2.7: UNKNOWN

[byroot (Jean Boussier)]

The error is easy to reproduce:

e.g. on Ruby 2.3:

$ ruby -robjspace -e 'ObjectSpace.trace_object_allocations_stop'
-e:1:in `trace_object_allocations_stop': wrong argument type false (expected tracepoint) (TypeError)
	from -e:1:in `<main>'

Up to ruby 2.7.1:

$ ruby -robjspace -e 'ObjectSpace.trace_object_allocations_stop'
Traceback (most recent call last):
	1: from -e:1:in `<main>'
-e:1:in `trace_object_allocations_stop': wrong argument type false (expected tracepoint) (TypeError)

Patch: https://github.com/ruby/ruby/pull/3001

[nobu (Nobuyoshi Nakada)]

• Assignee set to byroot (Jean Boussier)

A possible alternative might be an exception, I thought first, but it isn’t a single condition.
Could you add the test?

[byroot (Jean Boussier)]

As explained in the PR, I have no idea how to write a regression test, because as long as ObjectSpace.trace_object_allocations_start was called at least once in the program, it no longer fails afterwards. That is what led me to think it's a bug.

I can write test that calls trace_object_allocations_stop but based on test ordering it might never fail on CI.

[byroot (Jean Boussier)]

I added a regression test using assert_separately.

[byroot (Jean Boussier)]

I can't close, but this is fixed in a74df67244199d1fd1f7a20b49dd5a096d2a13a2

[tenderlovemaking (Aaron Patterson)]

• Status changed from Open to Closed

I merged. Thanks for the patch