fix: remove spaces in adjacency matrix expression

[wdconinc]

Briefly, what does this PR introduce?

When there are spaces in the adjacency matrix string expression's default value, JANA2 chops at the first space. This results in many warnings, i.e. https://github.com/eic/EICrecon/actions/runs/4121127389/jobs/7116665443.

The issue seems to be in the SetDefaultParameter call, since modifying u_adjacencyMatrix in the algorithm didn't have any impact. At that point the string has already been chopped.

What kind of change does this PR introduce?

• Bug fix (issue #__)
• New feature (issue #__)
• Documentation update
• Other: __

Please check if this PR fulfills the following:

• Tests for the changes have been added
• Documentation has been added / updated
• Changes have been communicated to collaborators

Does this PR introduce breaking changes? What changes might users need to make to their code?

No.

Does this PR change default behavior?

No.

[wdconinc]

@faustus123 @nathanwbrei I am very surprised that JANA2 changes default values set in the C++ in a SetDefaultParameter call. Can't we even trust default values anymore?

[2023-02-08 21:29:43.387] [EcalBarrelSciGlassProtoClusters] [error] (abs(tower_1 - tower_2) + (abs((sector_1 - sector_2) * 5 + row_1 - row_2) == 1) + (abs((sector_1 - sector_2) * 5 + row_1 - row_2) == (24 * 5 - 1))) == 1
[WARN] Parameter 'BEMC:EcalBarrelSciGlassProtoClusters:adjacencyMatrix' loses equality with itself after stringification
[2023-02-08 21:29:43.387] [EcalBarrelSciGlassProtoClusters] [error] (abs(tower_1

[faustus123]

@faustus123 @nathanwbrei I am very surprised that JANA2 changes default values set in the C++ in a SetDefaultParameter call. Can't we even trust default values anymore?

Ha! You're right. This is a bug where we need to handle strings differently than other data types. Spaces in strings are not being handled properly.

The issue is due to parameters being passed through std::stringstream to convert them into string form first and then back into the native type of the parameter (e.g. double). This is actually done on purpose so one can run a process using default values from the C++ code and have it dump the full configuration to a file. That configuration file should then be able to reproduce the job exactly. If there are round-off errors in the stringification, those will need to be included in the original run.

We will need to add a template specialization for strings types that skips the conversion using std::stringstream. We'll take care of that on the JANA2 end.

[wdconinc]

I'd like to merge this soon. Anyone can review and approve?

• Before: https://github.com/eic/EICrecon/actions/runs/4121127389/jobs/7116665443
• After: https://github.com/eic/EICrecon/actions/runs/4130669007/jobs/7137702389

[faustus123]

This looks like a reasonable and simple workaround.

[AyanRoy16]

Hello @wdconinc ,
We have added a test case for JParameter to show that a c++ command that is parsed as a string remains the same after it is parsed, the white spaces are correctly parsed. Let me know if we are missing out on something.
PR: JeffersonLab/JANA2#248
cc. @faustus123 @nathanwbrei