Author pablogsal
Recipients brandtbucher, nascheme, pablogsal, tim.peters
Date 2020-10-09.19:46:26
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1602272786.82.0.951617252047.issue41984@roundup.psfhosted.org>
In-reply-to
Content
Thanks for noticing this, Brandt!

> The fix is simple: track all instances of user-defined classes, no exceptions

I have the fear that this may increase the time of collections substantially because even if many of them won't have references at all, they still net to be visited and they count towards the time complexity of the algorithm that detects isolated sub-cycles. I mention this because this could affect an unbounded number of instances per type. For example, some of the types that are untracked.

Maybe I am missing something but we could mark them as having GC support unconditionally but still leave them untracking and unconditionally add a tracking call on setattribute.
History
Date User Action Args
2020-10-09 19:46:26pablogsalsetrecipients: + pablogsal, tim.peters, nascheme, brandtbucher
2020-10-09 19:46:26pablogsalsetmessageid: <1602272786.82.0.951617252047.issue41984@roundup.psfhosted.org>
2020-10-09 19:46:26pablogsallinkissue41984 messages
2020-10-09 19:46:26pablogsalcreate