Add AllowUnknownEnumValues setting to allow disabling enum validation

[konradbloor]

Hi - I saw this issue and thought it would be interesting to try to submit a PR for it.

Resolves #1290

AllowUnknownEnumValues session setting added; this then skips validation of field values against the data dictionary enumerated values. This means they won't be rejected with IncorrectTagValue. The other validation checks are left unaffected. Fields without enumerated values behave as they did before. The per-field allowOtherValues dictionary attribute remains the targeted alternative for individual fields.

Mirrors the equivalent QuickFIX/N setting (connamara/quickfixn#929) in name, default (N), and semantics.

Unlike QuickFIX/N, the flag lives in ValidationSettings rather than in DataDictionary, since DataDictionary instances are cached/shared amongst sessions here.

Testing (tests written first and confirmed failing before the implementation):

• Unknown enum value rejected by default; accepted with the setting enabled — including a multiple-value field (ExecInst) with one invalid token, mirroring the QuickFIX/N test
• ValidationSettings copy constructor retains the new flag
• DefaultSessionFactory wiring: setting Y propagates to the session; absent setting defaults to false
• Full module suites pass: quickfixj-base (167 tests), quickfixj-core (1677 tests, incl. the acceptance suite)
• Built/tested with the minimal-fix-latest,skipBundlePlugin profiles per AGENTS.md
• Docs entry added to configuration.md

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 70.42%. Comparing base (237d2bc) to head (6b63228).

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #1303      +/-   ##
============================================
+ Coverage     70.35%   70.42%   +0.07%     
- Complexity     2241     2247       +6     
============================================
  Files           159      159              
  Lines          9056     9065       +9     
  Branches       1192     1192              
============================================
+ Hits           6371     6384      +13     
+ Misses         2220     2217       -3     
+ Partials        465      464       -1     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[the-thing]

Looks ok, but comments like these are not required. SCM or issue tracking is for history.

 /**
     * https://github.com/quickfix-j/quickfixj/issues/1290
     * */

[konradbloor]

Thanks @the-thing for reviewing - good point! I removed them.

[the-thing]

I missed this one. Injecting directly boolean flag instead of full object is probably better.

[konradbloor]

Thanks - I agree, is it worth my doing follow up PR so we have consistency amongst the neighbouring calls?

[the-thing]

Separate PR? - yes.

It might be worth keeping settings passed as an object if more than two flags (or two and more) are being read in a single method?

[konradbloor]

Cool - once I'm back later I'll do this separately. Agreed too many on one method won't help clarity, I'll have a look through and see what would be clearest 👍 Thanks again for the review.

[the-thing]

LGTM

[konradbloor]

Thanks for the review on this. Just confirming my plan - to make sure that plan works for anyone interested:

Once this has merged I'll then make a followup PR from master to make all the neighbouring calls inside DataDictionary consistent (so ValidationSettings aren't being passed when there's just, say, a single boolean that is relevant).

[chrjohn]

@konradbloor thanks for the PR.
@the-thing thanks for your review. I am OOO for the next weeks so will try to keep up as good as possible.