Re: [eigen] non-linear optimization test summary

[ Thread Index | Date Index | More Archives ]

You didn't cause any trouble --- this has be bugging a few of us for a
while, already. A couple of weeks ago, we exchanged a dozen emails
with Thomas and Gael.

I didn't change anything. If tests are passing again, I would just
suspect plain luck / something that made the bug no longer be
triggered, just like it appeared someday for no good reason, but it
must still be latent. Indeed, if, as I suspect, it's just a bad FP
compare failing or succeeding depending on the desiderata of extended
FP precision, then it can be extremely capricious!!


2010/6/15 Hauke Heibel <hauke.heibel@xxxxxxxxxxxxxx>:
> Thanks for all the effort. I did not really mean to cause so much
> trouble before the release.
> Just out of curiosity. Did you change anything? I am just asking since
> almost all tests in the dashboards are suddenly passing again.
> - Hauke

Mail converted by MHonArc 2.6.19+