[Gnu-arch-users] Re: Avoiding ancestor scan during get with revision library

[Stefan Monnier]

> Something that would be quite valuable in this would be non-volatile
> ancestry data in the tree.  Patchlogs are almost right (and I've got tools
> that use 'em), but since patchlogs can be removed (e.g. by replay --reverse,
> or by your good self), we can't rely on them in every case.

Huh?  That's what patch-logs are for, isn't it?
If the user dicks around with it, it's his problem.
Of course, we should probably be careful so as to degrade gracefully when
a patch-log is missing, but it seems to be pretty naturally the case for
all the examples you give.


        Stefan