This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author al3x3i.k
Recipients al3x3i.k
Date 2022-03-03.14:48:34
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <>

Maybe it is just my misunderstanding of the mechanism, so please excuse me if it is so. Looking at the GC module code, I think that once there are X objects in the oldest gen, a full collection blocked due to the condition being always True.

So, if youngest gen is collected each 700 objects, the middle gen will have maximum 7000 objects. Full collection will be triggered each ~70K objects. X/4 = 70000 ===> X=280000. If oldest gen has >280K objects in it, the condition for full collection (long_lived_pending < gcstate->long_lived_total / 4) is always true since to my understanding, long_lived_pending is 70K at most, and got zeroed each full collection (
Date User Action Args
2022-03-03 14:48:34al3x3i.ksetrecipients: + al3x3i.k
2022-03-03 14:48:34al3x3i.ksetmessageid: <>
2022-03-03 14:48:34al3x3i.klinkissue46912 messages
2022-03-03 14:48:34al3x3i.kcreate