[Yt-dev] yt: moving pickled things

[Matthew Turk]

> Here's a thought I had about skirting this in the future-- Would it be
> possible to have yt alert the user when CTID is missing, and have a
> tool that inserts some other time as CTID straight into the file
> (current time at call I think would get you all the uniqueness of pf
> you desire)  This would get around the timestamp problem, which has
> bitten me more than once (esp. on machines that actually enforce their
> purge policy)

I think at one point I think we warned people; we don't right now, but
I'd be fine with doing so.  Additionally, adding in the CTID would be
fine with me, but I think other people might be better suited to
making that call; on the one hand, I'm sort of opposed to
non-interactively modifying things.

Although, I wonder sometimes -- how much effort should we be expending
to supporting edge cases of the code?  I mean, yours is a different
situation in the sense that you also know what's going on, and it's
sort of happening in the open now.  I am inclined to moving toward
just supporting LCA Enzo and maybe one or two other largely
non-invasive branches.  I had to write another unit parsing routine
about two weeks ago, and I am getting increasingly tired of inserting
corner cases.

-Matt