Many thanks for taking care of all these extensions to the K-matrix @johndan !
Will land this now.

Thanks for checking through this, @tlatham !

• Remove unnecessary error message

• Add catch-all return and finish B-W updates

Hi @johndan, thanks for this. Few comments inline.

Sounds good to me, many thanks @johndan.

I've just written this up in an LHCb ANA. Can I use this, and @jback's initial documentation (from Mattermost), to take a first-pass at this?

Just to say thanks very much to @jback and @tlatham for taking time to do a really thorough review. Much appreciated!

Code can be merged, although it would be good if the "DK P-wave" comment for calcGamma in LauKMatrixPropagator.hh is removed before this revision is landed.

Thanks, @jback, done!

• Correct comment

Thanks @johndan for implementing my suggestion.

• Compilation fixes

• Factorise retrieval of the phase-space density and use to obtain barrier-factor matrix for all channels

All looks good to me thanks @johndan. If @jback can do a final check over as well, since there have been a few changes, then hopefully we can get this landed soon.

• Reset haveCalled... variable in initialiseMatrices()

• Return propagator pointer in LauIsobarDynamics::defineKMatrixPropagator
• Change proposed by Tom to ensure that K-matrix is always treated as non-narrow for integration
• Fix: set spin type correctly to Legendre for the K-matrix

• Fix 5 compile-time syntax errors

• Remove noexcept
• Merge release notes
• Restore getFloatingParameters doxygen comment
• Fix LauKMatrixProdSVP whitespace issues
• Fix parameter naming in resAmp function declaration for ProdPole and SVP
• Further whitespace fix to SVP
• Uniform default variable
• Add const variables
• Make verbose flag static constexpr (nice, thanks Tom\!)
• Add braces for Tom
• Move setting of propagator denominator parameter to setParameters(...) routine
• Correct a initialisation and move it to a local variable in storeParameters()
• Update setParameters(...) text to describe new non-zero-spin-related input lines in configuration file for K-matrix
• Simplify default assignation for L_ and radii_ vectors
• Reinstate D39 name for pole mass sq param
• Reinstate D39 name for scatt constant param
• Reverse change to error message
• Use atoi for angular momenta not atof
• Add switch on phasespaceindex inside calcGamma(...) to catch undefined calls to that function

Do you have an example of using this generalisation of the K-matrix that could go into the examples dir?

Not yet, but very happy to add one when this implementation has been tested a bit more in my fits. Can I do this later, or shall I pause D41 landing until I've gained a bit more experience with the K-matrix in D0 K+?

• Remove noexcept
• Update release notes
• Restore getFloatingParameters doxygen comment
• Fix LauKMatrixProdSVP whitespace issues
• Fix parameter naming in resAmp function declaration for ProdPole and SVP
• Further whitespace fix to SVP
• Uniform default variable
• Add const variables
• Make verbose flag static constexpr (nice, thanks Tom\!)
• Add braces for Tom
• Move setting of propagator denominator parameter to setParameters(...) routine
• Correct a initialisation and move it to a local variable in storeParameters()
• Update setParameters(...) text to describe new non-zero-spin-related input lines in configuration file for K-matrix
• Simplify default assignation for L_ and radii_ vectors
• Reinstate D39 name for pole mass sq param
• Reinstate D39 name for scatt constant param
• Reverse change to error message
• Use atoi for angular momenta not atof
• Add switch on phasespaceindex inside calcGamma(...) to catch undefined calls to that function

Thanks for all this @johndan it looks really nice. I've not really gone through the physics since I think you and @jback have gone through that thoroughly already. So these comments are all more technical.
There are various inline comments (mostly minor). Plus, here are a few general comments:

• Do you have an example of using this generalisation of the K-matrix that could go into the examples dir?
• Please remove the noexcept specifications on functions
  • I would prefer to do a systematic sweep through the package as part of T34 and I still need to fully firm up my understanding of when it is necessary/helpful to do this and what are the prerequisites for it to be valid to do so
  • Having said that, I'm pretty certain that some of those that you've specified noexcept are actually potentially throwing because they call functions that are themselves potentially throwing, e.g. TMatrixD::operator[]
• Please update the release notes

• minor change cherry-picking earlier BelleNR commit

• Typo in LauKMatrixPropagator.cc
• Updates to K-matrix propagator to use refined barrier factor
• Clean up indentation
• Parallel updates to KMatrixPropagator.hh
• Cleanup indentation
• Revert "Allow slope of exponential non-resonant shape to vary positive (for low-stats fits)"
• rebase conflict
• rebase conflict
• D*K system should be in P-wave\!
• Changes after round of code-review with @jback
• Remove whitespace
• Remove whitespace
• Remove extraneous comment
• Fix compilation errors
• rebase conflict
• Simplify setting of denominator parameter in barrier factor
• Cache the a/R^2 calculation
• Neater access to D*0 mass information from LauResonanceMaker - thanks to Tom
• Add definition of angular momentum term for spin-2 (max spin) to K-matrix
• Remove done TODO comment and make const several function arguments
• rebase conflict
• Add barrier factors to numerator of K-matrix amplitude in ProdPole and ProdSVP, which need to know the spin of the K-matrix using the new getL() function
• Fix typo in ProdPole.cc and ProdSVP.cc
• - Changes to allow different orbital angular momenta for different channels inside the K-matrix
• Correctly initialise a_ vector. Remove calculation of angular factor from K-matrix propagator. Remove any need to update the propagator with the LauKinematics object
• Remove angular parts of K-matrix header, and add extra function to indicate the final-state channel of the various coupled channels
• Update ProdPole and SVP to use resAmp (and leave amplitude to LauAbsResonance)
• Update ProdPole and SVP headers
• rebase conflict
• Changes to LauResonanceMaker to allow retrieval of parent Blatt-Weisskopf factor without using resInfo that is not defined for K-matrix
• Changes to LauBlattWeisskopfFactor to accommodate tangential use by K-matrix to get parent factor without access to a resInfo
• Add K-matrix name to gCoupling parameters to avoid clashes in models with >1 K-matrix
• Resolve merge conflict
• Allow other parameters of the SVP part to be floated
• Fix merge conflict
• rebase conflict
• Remove duplicate par_mSq0_
• rebase conflict
• Protect all parameters of calls to resAmp for prodpole / svp with const
• resolve conflict
• Missing ;
• Movement of one comment and correct access to one floating parameter
• Remove unnecessary whitespace
• Allow floating of s0scatt and mSq0
• Cherry pick commit from 19th Jan that moved enum class KMatrixChannels and resolve associated conflicts
• Turn off auto-verbose-printout in propagator
• Move a comment - trivial
• Increase precision of verbose printout
• Clean whitespace

Many thanks @johndan, this looks great. I understand from @jback that he's also happy with this, so I'll go ahead and land this one onto master.

• Movement of one comment
• Allow floating of s0scatt and mSq0
• Cherry pick commit from 19th Jan that moved enum class KMatrixChannels and resolve associated conflicts
• Turn off auto-verbose-printout in propagator

• Allow floating of s0scatt and mSq0

• Movement of one comment and correct access to one floating parameter
• trivial reset of changes caught up in earlier commit

• Minor, aesthetic, improvements

The exact amplitude has been reproduced with another fitter. Anything left for me to do before we land?

Can we 'land' this so that we can start merging D41? Thanks!

• Protect all parameters of calls to resAmp for prodpole / svp with const
• Clean up K-matrix propagator default values. Add lots of const protection to function calls. Noexcept throughout

• Use default values at declaration in LauAbsResonance to significantly simplify the constructors

• Cleanup LauAbsResonance constructor which involved a lot of duplication for K-matrix

• Remove duplicate par_mSq0_

Thanks, @jback. A few answers to your responses, with just one that doesn't (yet) make sense to me...

• Add K-matrix name to gCoupling parameters to avoid clashes in models with >1 K-matrix
• Resolve merge conflict
• Allow other parameters of the SVP part to be floated
• Fix merge conflict

• Remove whitespace at end of file
• Fix compilation errors

I think only 1 parameter needed a changed name - @jback let me know if I missed something here though.

• Allow other parameters of the SVP to be floated

• Add K-matrix name to gCoupling parameters to avoid clashes in models with >1 K-matrix

• Correctly initialise a_ vector. Remove calculation of angular factor from K-matrix propagator. Remove any need to update the propagator with the LauKinematics object
• Remove angular parts of K-matrix header, and add extra function to indicate the final-state channel of the various coupled channels
• Update ProdPole and SVP to use resAmp (and leave amplitude to LauAbsResonance)
• Update LauAbsResonance to apply spin to constructed objects that don't use the 'normal' constructor
• Changes to LauResonanceMaker to allow retrieval of parent Blatt-Weisskopf factor without using resInfo that is not defined for K-matrix
• Changes to LauBlattWeisskopfFactor to accommodate tangential use by K-matrix to get parent factor without access to a resInfo

I've included the changes to tell the K-matrix about the orbital angular momentum of each channel, as discussed on Mattermost. I've also made 3 comments in the code for you, @jback. Thanks!

• Changes to allow different orbital angular momenta for different channels inside the K-matrix

• Changes to allow different orbital angular momenta for different channels inside the K-matrix

From further discussion elsewhere, it looks like the code needs further edits to allow channel-dependent L values (for a given total angular momentum J).

Need to apply an upper bound = 0 for s_0^scatt if this is allowed to float. Some parameters still need "names" added, as per Tom's suggestion.

• Remove done TODO comment and make const several function arguments
• Add function to LauKMatrixPropagator that returns K-matrix spin, and make various function arguments const
• Add barrier factors to numerator of K-matrix amplitude in ProdPole and ProdSVP, which need to know the spin of the K-matrix using the new getL() function
• Fix typo in ProdPole.cc and ProdSVP.cc

• Correct displaced doxygen comment
• Simplify setting of denominator parameter in barrier factor from file
• Cache the a/R^2 calculation - thanks John!
• Neater access to D*0 mass information from LauResonanceMaker - thanks Tom!
• Add definition of angular momentum term for spin-2 (max spin) to K-matrix

• Correct displaced doxygen comment
• Simplify setting of denominator parameter in barrier factor from file
• Cache the a/R^2 calculation - thanks John!
• Neater access to D*0 mass information from LauResonanceMaker - thanks Tom!
• Add definition of angular momentum term for spin-2 (max spin) to K-matrix

Thanks @jback and @tlatham, I've handled all the remaining comments. There are some deeper issues to resolve (see the Laura++ mattermost channel) before we land this. D39 needs to come first, anyway, and that one is ready to go.

• Correct Doxygen comments

In D41#1074, @johndan wrote:

General question, @tlatham : should I cherry-pick commits to bring this differential in line with the master (e.g. c38fa2a34f4a772754d1e8f610def10747d95104 - LauFormulaPar fix)?

Thanks @johndan, the changes you have made so far look good and makes the code more general. I have replied to your questions, and have also suggested that we calculate the barrier radii-squared terms at initialisation. Not sure what we can do for avoiding hard-coding the D* mass at initialisation (regarding access to LauResonanceMaker info).

• D*K system should be in P-wave, not just D0K !
• Changes after round of code-review with @jback
• Remove whitespace and make minor bugfixes found during compilation tests

Thank you for the comments @jback . I've implemented them in most cases, or left a reply in the few where further input from you is needed.

I've had a first look at the code and have provided some suggestions for changes that I think are needed.

• Add K-matrix name to parameter names, to allow multiple K-M's per model

• Add braces for Tom
• Allow pole mass-squared to vary

• Add doxygen lines for one function

Thanks very much for this @johndan. It looks good.
I think it would just be good to take the opportunity while we're at it to allow any of the parameters to float.
We should also include the name of the propagator in the name of the parameters so that if we have more than one K-matrix in a given fit we don't get identically named parameters appearing - this will cause problems with the sim-fit for example.
See the inline comments for details.

I tested the patch and it does the job.