|
From: | graydon hoare |
Subject: | [Monotone-devel] Re: db init oddity |
Date: | Sat, 29 Nov 2003 13:00:03 -0500 |
User-agent: | Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031115 Thunderbird/0.3 |
Julian Seward wrote:
Same program on same data gives different results on different runs? Can you run it on valgrind and see if there are any uses of uninitialised values reported?
yes, there are several. I will patch and report these upstream to the sqlite author. though, in this case I suspect what tromey *might* be seeing is a residual journal file (from the previous run) with the same name stem as the db he's initializing. I've seen this upset sqlite before (transiently).
in any event, it would probably be a good use of time to make the auto-tester tromey's working on run the testsuite under valgrind. thanks for the tip.
-graydon
[Prev in Thread] | Current Thread | [Next in Thread] |