[ntuple] Use dynamic platform-dependent width rather than hard-coded width of long

[ferdymercury]

This Pull request:

Changes or fixes:

Fixes #11616

Checklist:

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

[github-actions]

Test Results

    21 files      21 suites   3d 9h 36m 20s ⏱️
 3 217 tests  3 217 ✅ 0 💤 0 ❌
65 834 runs  65 834 ✅ 0 💤 0 ❌

Results for commit d6a99b5.

♻️ This comment has been updated with latest results.

[hahnjo]

I think we should generally replace all instances of this pattern (ie also for kInt_t, kUInt_t etc.) and fully rely on RField to do the mapping.

[ferdymercury]

I think we should generally replace all instances of this pattern (ie also for kInt_t, kUInt_t etc.) and fully rely on RField to do the mapping.

Do you mean eg changing

case kUShort_t: Attach(std::make_unique<RField<uint16_t>>("_0")); break;

With

case kUShort_t: Attach(std::make_unique<RField<UShort_t>>("_0")); break;

?