fix: render true booleans as "true" in XML output

[gennaroprota]

The XML writer emitted true boolean members as <x>1</x> (and omitted the element when false). It now emits the xsd:boolean primary lexical form, <x>true</x>. The omit-when-false convention is preserved, so the golden XMLs change only in the body of already-present boolean elements. The RELAX NG schema is updated accordingly.

---

Summary

This fixes a long-standing readability quirk in the XML output. The XML writer emitted true boolean members as <x>1</x> (and omitted the element when false). Reading such output meant the consumer had to know in advance that the member was actually a boolean, as 1 looks indistinguishable from an integer.

The writer now emits the xsd:boolean primary lexical form: <x>true</x>. The omit-when-false convention is preserved, so the golden XML fixtures change only in the body of already-present boolean elements. The RELAX NG schema is updated to match.

Changes

• Source: One-line fix in src/lib/Gen/xml/XMLWriter.cpp; the boolean-serialization branch now writes "true" instead of "1".
• Golden tests: 207 XML fixtures across test-files/golden-tests/ updated to reflect the new <x>true</x> form.
• Build: mrdocs.rnc updated so the schema accepts the new lexical form on boolean elements.
• Breaking changes: Downstream tooling that parses mrdocs XML output and matches boolean fields exactly against "1" will need to update to handle "true". The omit-when-false convention is unchanged, so the handling of absent elements is unaffected.

Testing

• The golden XML fixtures are themselves the regression coverage: any future regression that re-emits "1" for booleans (or omits an element when true) fails the golden tests.
• No CI workflow changes needed.

Documentation

The schema (mrdocs.rnc) is the canonical reference for the XML output format and is updated. No prose documentation describes the previous "1" encoding, so nothing else needs updating.

[github-actions]

✨ Highlights

• 🧪 Existing golden tests changed (behavior likely shifted)

🧾 Changes by Scope

Scope	Lines Δ%	Lines Δ	Lines +	Lines -	Files Δ	Files +	Files ~	Files ↔	Files -
🥇 Golden Tests	100%	4610	2305	2305	207	-	207	-	-
🛠️ Source	<1%	20	11	9	1	-	1	-	-
🏗️ Build	<1%	2	1	1	1	-	1	-	-
Total	100%	4632	2317	2315	209	-	209	-	-

Legend: Files + (added), Files ~ (modified), Files ↔ (renamed), Files - (removed)

🔝 Top Files

• test-files/golden-tests/symbols/variable/function-objects.xml (Golden Tests): 370 lines Δ (+185 / -185)
• test-files/golden-tests/symbols/record/class-template-specializations-1.xml (Golden Tests): 312 lines Δ (+156 / -156)
• test-files/golden-tests/config/inherit-base-members/copy.xml (Golden Tests): 204 lines Δ (+102 / -102)

Generated by 🚫 dangerJS against 721e020

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.16%. Comparing base (f942a24) to head (721e020).

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #1200      +/-   ##
===========================================
+ Coverage    78.66%   83.16%   +4.49%     
===========================================
  Files          344       35     -309     
  Lines        35681     3658   -32023     
  Branches      7199      843    -6356     
===========================================
- Hits         28067     3042   -25025     
+ Misses        4984      409    -4575     
+ Partials      2630      207    -2423     

Flag	Coverage Δ
bootstrap	83.16% <ø> (ø)
cpp	?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[cppalliance-bot]

An automated preview of the documentation is available at https://1200.mrdocs.prtest2.cppalliance.org/index.html

If more commits are pushed to the pull request, the docs will rebuild at the same URL.

2026-06-18 07:35:46 UTC

[alandefreitas]

The XML writer emitted true boolean members as 1 (and omitted the element when false). It now emits the xsd:boolean primary lexical form, true. The omit-when-false convention is preserved

I'm no XML expert, but that seems like a weird convention to make. If we omit when false, this is technically a flag. And flags don't have to say "true". I thought booleans were typically represented as attributes in XML.

[gennaroprota]

The XML writer emitted true boolean members as 1 (and omitted the element when false). It now emits the xsd:boolean primary lexical form, true. The omit-when-false convention is preserved

I'm no XML expert, but that seems like a weird convention to make. If we omit when false, this is technically a flag. And flags don't have to say "true".

I was going to point out the same. true is currently redundant.

I thought booleans were typically represented as attributes in XML.

Yes. For us, the form would be something like:

<function name="foo" is-const="true" is-variadic="false" ...>
  <param>...</param>
  ...
</function>

The MrDocs schema currently has almost no data attributes, so booleans-as-elements is at least consistent. I'd suggest switching to "empty element (e.g. <is-variadic/>) when true, nothing when false", as that avoids the enormous diff that would result from moving all booleans to attributes.

[alandefreitas]

The MrDocs schema currently has almost no data attributes

Yes. We used to have many.

But I think it's going to be more like is-const than is-const="true" because =true is also redundant here.

For name="foo", then I'm not sure. We need research on typical usage and trade-offs.

[alandefreitas]

Hi, checking in. We agreed on "empty element when true (<is-variadic/>), omitted when false" on 2026-05-22, but the branch still has the original <x>true</x> from 2026-05-13. Is this PR blocked by something else?