Re: [Monotone-devel] Encountered a segfault on monotone pull 8-O

[Nathaniel Smith]

On Fri, Feb 11, 2005 at 11:22:15AM +0100, Christof Petig wrote:
> I just wanted to tell you that after 81 minutes (on a 3GHz HT P4!) a
> pull repeatably causes a segfault (I didn't believe it the first time):
[snip]
> this certainly looks like a far too deep recursion to me, though it
> finally ends before consuming all available memory.
[snip]
> I will try to rerun it with debug output enabled, perhaps this gives a
> better clue. Do you, Graydon or Nathaniel, have another command to throw
> at gdb?

Ah, nope, it seems perfectly clear -- it's just the recursion.  One of
the sad things about C/C++ (in all the usual implementations, anyway)
is that deep recursion will segfault when it exhausts the stack, not
when it exhausts the heap, and the stack is much smaller.

So we just need to change the delayed packet writing code to be
non-recursive.

-- Nathaniel

-- 
So let us espouse a less contested notion of truth and falsehood, even
if it is philosophically debatable (if we listen to philosophers, we
must debate everything, and there would be no end to the discussion).
  -- Serendipities, Umberto Eco

This email may be read aloud.