[hist] Add option I to TH1::Chisquare for bin integral chi2

[JasMehta08]

This Pull Request

Changes or fixes

TH1::Fit supports option "I" which uses the integral of the fitted function over each bin instead of the value at the bin center. TH1::Chisquare did not parse this option, so TF1::GetChisquare() and TH1::Chisquare(f, "I") returned different results after a fit with option "I".
Changes across 3 files:

• TH1.cxx: parse option "I" into useIntegral, pass it to ROOT::Fit::Chisquare, and
  document the new option alongside the existing "R", "L", "P" options.

• HFitImpl.cxx: add bool useIntegral = false to ROOT::Fit::Chisquare(TH1) and
  HFit::ComputeChi2. Set DataOptions::fIntegral = useIntegral before BinData is filled.

• HFitInterface.h: update the public declaration with the new parameter, defaulted to
  false for full backward compatibility.

Before this fix:

h.Fit(f, "QI")
f.GetChisquare() # 7.26965
h.Chisquare(f, "I") # 7.27083 <-- "I" was silently ignored

After this fix:

h.Fit(f, "QI")
f.GetChisquare() # 7.26965
h.Chisquare(f, "I") # 7.26965 <-- correct

Checklist

• tested changes locally
• updated the docs (if necessary)

This PR fixes #21695

[github-actions]

Test Results

    21 files      21 suites   2d 23h 31m 5s ⏱️
 3 833 tests  3 830 ✅  1 💤 2 ❌
72 199 runs  72 179 ✅ 18 💤 2 ❌

For more details on these failures, see this check.

Results for commit e01ba9e.

♻️ This comment has been updated with latest results.