Can anyone tell me why the object created with finializer is not garbage
collected during GC.start executed in line 10 please ?
Robert Pankowecki
On Wed, 25 May 2011 20:11:20 +0900, Robert Pankowecki wrote:
Can anyone tell me why the object created with finializer is not
garbage
collected during GC.start executed in line 10 please ?Robert Pankowecki
Because exceptions thrown in finalizers are silently ignored:
http://rxr.whitequark.org/mri/source/gc.c#2862
(note the last rb_protect).
But ‘puts 12345’ is executed at the end of the program instead in the
middle of the test.
On Wed, May 25, 2011 at 1:11 PM, Robert Pankowecki
[email protected] wrote:
Can anyone tell me why the object created with finializer is not garbage
collected during GC.start executed in line 10 please ?
There are no guarantees - neither about when GC will occur nor about
which objects it will collect. AFAIK Ruby MRI’s GC doesn’t even
bother to do something until your memory usage goes beyond a
particular point. It’s unlikely that you reach that point with your
test.
Kind regards
robert
Robert Pankowecki wrote in post #1000863:
Can anyone tell me why the object created with finializer is not garbage
collected during GC.start executed in line 10 please ?
The GC puts “objects with finalizers” on a queue to be GC’ed later,
because (I guess) it doesn’t want to run any ruby code during the middle
of the GC run.
-roger-