[Mailman-Developers] Reconstructing config.pck
Thu, 8 Aug 2002 11:31:48 -0700
On Thu, 8 Aug 2002 14:11:22 -0400 email@example.com (Barry A. Warsaw) wrote:
>I'm not sure what config.pck.save is -- did you mean config.pck.last?
>Also, do you have a config.pck.tmp.* file laying around? Mailman
>tries really really hard not to leave you with no usable fall backs.
Yes, I meant config.pck.last. And there aren't any
Just to be clear, this wasn't a Mailman problem AFAICT; the files
got corrupted on the disk.
On Thu, 8 Aug 2002 11:18:48 -0700 "Donn Cave" <firstname.lastname@example.org> wrote:
>Quoth email@example.com (Barry A. Warsaw):
>| Here's the sad part: by default, pickles are written in binary format
>| so you're going to have a more fun plucking apart the pickle. Pickles
>| are essentially little state machines, and the `language' is defined
>| in Python's pickle.py module. UTSL is probably going to be your best
>Stuff accumulates in Unpickler().stack, so assuming there's nothing
>of interest after the start of invalid data, one could catch the
>exception and examine "stack".
That would help, not having to do all the parsing by hand at least.
I'll give it a shot.
>Beats vi, maybe.
Emacs, man. Someone's probably even written python-pickle-mode.
Thanks, Barry and Donn.