Pablo Galindo Salgado

👋 Hey, `memray` author and Python core dev here. I run your example with native symbols and debug info and this is what I get: ![Screenshot 2024-02-27 at 17 19...

If you want to give the refcycle theory a go you can use https://docs.python.org/3/library/gc.html#gc.set_debug with https://docs.python.org/3/library/gc.html#gc.DEBUG_LEAK to confirm what are cycles.

The problem with this option is that `memray` creates different result dumps for every forked process and expects the user to analyse every one separately. This feature will also require...

@gaborbernat any pointers on what is the best way to test with multiple `memray` versions?

I think this is enough to do the trick: ```diff diff --git a/tox.ini b/tox.ini index e41e2c8..16ff519 100644 --- a/tox.ini +++ b/tox.ini @@ -1,9 +1,10 @@ [tox] envlist = py310-cov -...

Oh, interesting. Could you give us a link to your project or the tests you are running so we can try to reproduce it?

I'm closing this issue as we don't have a reproducer. Please, feel free to reopen if you manage to get a simple example we can run to check the error.

Not really a fan of having `pretty_thread_name` and `thread_name`. Also, I think this PR makes the output nicer but it doesn't really solve that is requested in https://github.com/bloomberg/memray/issues/561 yet, so...

Do we know where these 7 minute are being spent? The dependencies are not that many so I am a bit surprised

Thanks @pitrou for bringing this to us. This is a very interesting problem indeed. The key to either method is that we need: * A specific name for a symbol....