Blair Zajac wrote:
> On Sep 5, 2007, at 5:34 PM, David Glasser wrote:
>> On 9/5/07, Malcolm Rowe <firstname.lastname@example.org> wrote:
>>> This is, of course, impossible. Format 3 FSFS filesystems always have a
>>> transaction-current file... unless you're perhaps talking about an
>>> older format 3 filesystem that might have existed only in a development
>>> version of 1.5.0? :-)
>>> While I'm -0 on anything that includes supporting code for unreleased
>>> versions of the filesystem, I'm -1 on anything that performs write
>>> operations on a filesystem that we aren't necessarily writing to -- FSFS
>>> filesystems should be readable from read-only media, and the below won't
>>> be (you'd need to ignore EACCESS, for one).
>> +1 to everything Malcolm said.
> Ditto here.
Noooooo problem. This was a 1.5-dev repository I was working with, but
haven't been tracking the various format changes to know if I was fixing a
real bug or just cleaning up a mess made by using unreleased code. Clearly,
it's the latter in this case. Thanks for the info, guys.
C. Michael Pilato <email@example.com>
CollabNet <> www.collab.net <> Distributed Development On Demand
Received on Thu Sep 6 14:16:34 2007