[#74190] [Ruby trunk Feature#12134] Comparison between `true` and `false` — duerst@...
SXNzdWUgIzEyMTM0IGhhcyBiZWVuIHVwZGF0ZWQgYnkgTWFydGluIETDvHJzdC4KCgpUc3V5b3No
3 messages
2016/03/07
[#74269] Type systems for Ruby — Rob Blanco <ml@...>
Dear ruby-core,
5 messages
2016/03/10
[#74395] [Ruby trunk Feature#12142] Hash tables with open addressing — shyouhei@...
Issue #12142 has been updated by Shyouhei Urabe.
3 messages
2016/03/17
[ruby-core:74310] [Ruby trunk Bug#12175] rb_gc_mark prevents the value from being garbage collected if called outside of a GC cycle
From:
mlumish@...
Date:
2016-03-14 23:09:14 UTC
List:
ruby-core #74310
Issue #12175 has been reported by Michael Lumish. ---------------------------------------- Bug #12175: rb_gc_mark prevents the value from being garbage collected if called outside of a GC cycle https://bugs.ruby-lang.org/issues/12175 * Author: Michael Lumish * Status: Open * Priority: Normal * Assignee: * ruby -v: ruby 2.1.2p95 (2014-05-08 revision 45877) [x86_64-linux] * Backport: 2.1: UNKNOWN, 2.2: UNKNOWN, 2.3: UNKNOWN ---------------------------------------- I had some code that called rb_gc_mark in an initializer function instead of in the corresponding mark function, and the value never got garbage collected, causing a memory leak. This seems like an easy mistake to make, and there does not appear to be a corresponding function to make the value eligible for garbage collection again. If the intention of the API when calling it in that situation is to make it a root object, then there should be a function to remove that indicator once the object no longer needs to be live. And if the intention is for the function to only work properly during a GC mark phase, then it should be an error to call it at any other time. -- https://bugs.ruby-lang.org/ Unsubscribe: <mailto:[email protected]?subject=unsubscribe> <http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>