From: Dave Wysochanski Date: Thu, 11 Dec 2008 22:22:58 -0500 Subject: [PATCH 12/12] Remove init_full_scan_done(0) call after create_toolcontext(). In-Reply-To: <20081211231005.GM24785@agk.fab.redhat.com> References: <1228965978-19132-1-git-send-email-dwysocha@redhat.com> <20081211231005.GM24785@agk.fab.redhat.com> Message-ID: <1229052178.20795.38.camel@localhost.localdomain> List-Id: To: lvm-devel@redhat.com MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit On Thu, 2008-12-11 at 23:10 +0000, Alasdair G Kergon wrote: > > Is this a bug? Why are we setting _full_scan_done = 0 here when we've > > already read the config file(s) for the 'scan' and other variables, then > > scanned the system for devices as shown above? > > We haven't scanned the system at that point. > > This scanning code is certainly messy, but it will become 'legacy' when the > udev integration work is done, so it's not worth spending much time on. > (It's a while now since we've had a bug report against this scanning code > so it's best not to change the behaviour of code paths.) > You are right. I don't know where I got lost in that code but I clearly did because I could not reproduce the scanning from the persistent filter file reading. A simpler argument would have just started with 0 or 1 for _full_scan_done and explored alternatives. As you say though, it's not worth it! ;-0