Re: Fix to long-standing crashes in GC

[Kim F. Storm]

Stefan Monnier <address@hidden> writes:

> > Yesterday, I did a partial test where I wrote a fun to check all
> > Lisp_Misc blocks for invalid misc types (except Lisp_Misc_Free) for
> > every call to Fcons -- and no such problems occurred, i.e. the bad
> > misc cells pointed to from some cons cells were NOT in the marker
> > block chain...
> 
> Hmm... was this run before, after, or during GC ?

It was on every entry to Fcons -- so pretty often.
I also call'ed the check function directly from gdb after it had
crashed, and there were still no invalid misc objects...

> Can you use the mem_find to try and figure out whether it's pointing at a
> GC-managed block or not?

IIRC, it was pointing to a MEM_CONS block.

> Can you also disable the block-release code (the one that calls xfree on
> blocks of cnos cells and lisp_misc objects) to see if it is maybe pointing
> to blocks that have been reclaimed?

I will try that.

-- 
Kim F. Storm <address@hidden> http://www.cua.dk