Re #28153 Change how Bk2BkExpConvPV gets the width

Framework/CurveFitting/src/Functions/Bk2BkExpConvPV.cpp

@@ -69,11 +69,13 @@ double Bk2BkExpConvPV::fwhm() const {
 }
 
 /** Set FWHM
- * It is an illegal operation of this type of peak
+ * This cannot be set directly so we assume Gamma is 0 and set Sigma by doing
+ * the reverse of calHandEta
  */
 void Bk2BkExpConvPV::setFwhm(const double w) {
-  UNUSED_ARG(w);
-  throw std::invalid_argument("Bk2BkExpConvPV is not allowed to set FWHM.");
+  this->setParameter("Gamma", 0);
+  auto sigma2 = std::pow(w, 2) / (8.0 * M_LN2);
+  this->setParameter("Sigma2", sigma2);
 }
 
 /** Set peak center

docs/source/release/v5.0.0/mantidworkbench.rst

@@ -105,6 +105,7 @@ Bugfixes
 - The help button in fitting now finds the relevant page.
 - Fixed an issue with changing normalisation on single spectra plots done from a script
 - Fixed an issue where fitting a distribution workspace was normalised twice.
+- Fixed an issue where adding a Bk2BkExpConvPV function to the fit browser caused a crash
 - Overplots will be normalized by bin width if they are overplotting a curve from a workspace which is a distribution.
 - Several bugs in the way Python scripts were parsed and executed, including blank lines after a colon and tabs in strings, have been fixed.
 - Axes limits of a plot no longer automatically rescale when errorbars are on/off