Bug #12685
closed
malloc error: pointer being freed was not allocated
Description
ruby crashed
- never seen that before
- should never happen
- looks serious to me
I'm using ruby 2.3.1p112 (2016-04-26 revision 54768) [x86_64-darwin15] from macports
I was running rails in dev mode.
The bug happened while reloading a page (like we all do a thousand times a day)
Error message in the shell was:
(12416,0x70000071a000) malloc: *** error for object 0x1c7fdde49ddcc0: pointer being freed was not allocated
See the dump attached.
This might either be a very serious bug,
or a bug in my hardware.
I must admit that I did see a sign that might point to me having a faulty RAM.
(a wrong rendering in a window's backing store)
But that was in EyeTV (a wrong rendering in a window's backing store)
(It is well known that EyeTV is a very buggy C++ software)
So I assume that having these two bugs within two days
might be more of an conincidence (and not caused by faulty ram)
So this might be a false alarm, but maybe it is not:
Please look at the backtrace, it crashed in:
6 libruby.2.3.0.dylib 0x0000000109da0b23 obj_free + 684
7 libruby.2.3.0.dylib 0x0000000109da049a gc_sweep_step + 473
8 libruby.2.3.0.dylib 0x0000000109d9f61a newobj_slowpath + 374
9 libruby.2.3.0.dylib 0x0000000109d9f484 newobj_slowpath_wb_protected + 20
10 libruby.2.3.0.dylib 0x0000000109d53177 ary_new + 66
11 libruby.2.3.0.dylib 0x0000000109d5d7fe flatten + 77
12 libruby.2.3.0.dylib 0x0000000109d5afd4 rb_ary_flatten_bang +
While a faulty bit could crash it anywhere
(and this must be a heavily exercised routine)
Please forget about faulty bits and alpha particles
(my machine is running perfectly stable)
It might actually be a software bug in the gc.
I fear this will be hard to reproduce,
but maybe the stack trace might help you.
I believe this to be a software bug in the gc (which is very serious)
(please read the trace)
Files
Updated by 
Updated by