Ticket #3549 (closed: duplicate)
Remove Guassian1D warning
Reported by: | Anders Markvardsen | Owned by: | Anders Markvardsen |
---|---|---|---|
Priority: | minor | Milestone: | Iteration 30 |
Component: | Mantid | Keywords: | |
Cc: | taylorrj@… | Blocked By: | |
Blocking: | Tester: | Nick Draper |
Description
Russell writes:
I’m trying to nail all of the compiler warnings on the mac. There should be just 1 left now, and it’s this one:
Gaussian1DTest.h(33): warning #1125: function "Mantid::CurveFitting::Fit1D::functionDeriv(const double *, Mantid::API::Jacobian *, const double *, size_t={darwin_size_t={unsigned long}})" is hidden by "ToyAlgorithm::functionDeriv" -- virtual function override intended?
void functionDeriv(const double* in, Mantid::CurveFitting::Jacobian* out, const double* xValues, const double* yValues, const double* yErrors, const size_t nData)
This is in the ‘ToyAlgorithm’ class in Gaussian1DTest. Basically, the signature of functionDeriv in Fit1D/Gaussian1D was changed at some point, but it wasn’t updated here. What this presumably means is that the ToyAlgorithm method doesn’t get called. In fact, if I alter the signature of ToyAlgorithm::functionDeriv to match the base class, then the test fails. If I comment out the method altogether, or just change the place in the test that uses ToyAlgorithm to use Gaussian1D, then the test passes.
So I suspect it isn’t doing what it’s supposed to be doing.
Roman already fixed this. Russell writes:
Thanks, but Roman already fixed this: http://trac.mantidproject.org/mantid/changeset/13917