If so, do you have any idea what caused this (manual intervention in the archive, baz, the library relink script itself :-^, ...)?
It's a bit mysterious, but I think what happend is that I did a
library-add while the script was running, and this reused an
inode number that the script already had in its contents cache. So when the script got around to the newly-added revision, it
misidentified the contents of a file, and relinked it to the
wrong thing.
Ah, nasty. Did you already implement whatever is needed to prevent this from happening
Not yet. Hmm. I think we just need to include size, mode and mtime in the id. Cool. (Technically possible to circumvent, but this is a chance all Arch clients take.)
or should we ensure that no one accesses the library while running the script?
For now, yeah. I'll see what I can do.
Aaron
-- Aaron Bentley Director of Technology Panometrics, Inc.
_______________________________________________ Gnu-arch-users mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnu-arch-users
GNU arch home page: http://savannah.gnu.org/projects/gnu-arch/
