New user self-registration is disabled due to spam. Please email eigen-core-team @ lists.tuxfamily.org if you need an account.
Before reporting a bug, please make sure that your Eigen version is up-to-date!

Bug 1156

Summary:	isApprox methods of certain classes should take const RealScalar& instead of RealScalar as precision argument
Product:	Eigen	Reporter:	twan_koolen
Component:	Geometry	Assignee:	Nobody <eigen.nobody>
Status:	RESOLVED FIXED
Severity:	Unknown	CC:	chtz, gael.guennebaud, hauke.heibel, jacob.benoit.1
Priority:	Normal
Version:	3.2
Hardware:	All
OS:	All
Whiteboard:

Description twan_koolen 2016-01-26 22:34:45 UTC

While DenseBase::isApprox already takes a const RealScalar& argument as the prec argument, some classes take a RealScalar by value. This can cause alignment issues similar to http://eigen.tuxfamily.org/bz/show_bug.cgi?id=1006, as well as being potentially expensive for certain custom scalar types.

Specifically, it looks like the following classes are affected:
./Eigen/src/Geometry/ParametrizedLine.h:  bool isApprox(const ParametrizedLine& other, typename NumTraits<Scalar>::Real prec = NumTraits<Scalar>::dummy_precision()) const
./Eigen/src/Geometry/Translation.h:  bool isApprox(const Translation& other, typename NumTraits<Scalar>::Real prec = NumTraits<Scalar>::dummy_precision()) const
./unsupported/Eigen/AlignedVector3:    inline bool isApprox(const MatrixBase<Derived>& other, RealScalar eps=NumTraits<Scalar>::dummy_precision()) const

Comment 1 Gael Guennebaud 2016-01-27 17:37:57 UTC

Thanks for the report, I found a few more:
devel: https://bitbucket.org/eigen/eigen/commits/7dfca64a4f83/
3.2:   https://bitbucket.org/eigen/eigen/commits/a0f1f2a625d2/

Comment 2 twan_koolen 2016-01-27 23:17:12 UTC

Awesome, thank you!