Hi Lukáš, On July 26, 2011 08:10:28 AM Lukáš Jirkovský wrote: > I think this is one of the most serious bugs hugin ever had.
What makes you reach this judgment? > In my opinion it's some threading issue, because even though it > happens very often, it doesn't happen every time. You know more about this than me. I've added the question to the bug report, if we get enough observations to correlate the number of threads in the system with the observed behavior we'll verify your hypothesis. > > With such a binary strategy, we should zero in on the bug in a maximum of > > eight iterations. > > > > 224, 112, 56, 28, 14, 7, 4, 2, 1 > > Good idea! I didn't thought about bisecting hugin before. My guess is > that it's somehow related to the panorama overview. BTW there's bisect > extension [1] which does this automatically. > > [1] http://mercurial.selenic.com/wiki/BisectExtension Oh, nice to see that somebody implemented the binary strategy in Mercurial (copying it from Git). OT: The binary strategy is applicable to many practical problems. I first encountered it when researching financial valuation models [1]. [0] https://bugs.launchpad.net/hugin/+bug/792896 [1] http://en.wikipedia.org/wiki/Binomial_options_pricing_model
signature.asc
Description: This is a digitally signed message part.
