Speed up the application of robust loss functions.
Since we added special handling for the case for rho[2] < 0,
the bulk of CorrectJacobian is pointless in the common case.
So add a simple one dimensional loop which rescales the Jacobian.
This speeds up this method immensely.
The robustification of a Jacobian gets speeded up by > 50%.
Change-Id: I97c4e897ccbb5521c053e1fb931c5d0d32f542c7
diff --git a/include/ceres/jet.h b/include/ceres/jet.h
index c78364d..55caa05 100644
--- a/include/ceres/jet.h
+++ b/include/ceres/jet.h
@@ -106,8 +106,8 @@
// Jet<double, 2> y(1); // Pick the 1st dual number for y.
// Jet<double, 2> z = f(x, y);
//
-// LG << "df/dx = " << z.a[0]
-// << "df/dy = " << z.a[1];
+// LOG(INFO) << "df/dx = " << z.a[0]
+// << "df/dy = " << z.a[1];
//
// Most users should not use Jet objects directly; a wrapper around Jet objects,
// which makes computing the derivative, gradient, or jacobian of templated
