fix(doclang): keep a single element-head layer token in merged inline runs

[eren-YU]

Fixes #778.

What changed

DocLangInlineSerializer merges an inline group's children and then wraps the merged text in a single <text> element. Each child is serialized independently beforehand, and every child whose content_layer is not BODY emits its own <layer/> element-head token. The merged wrapper therefore carried one <layer/> per child, while the DocLang element_head sequence permits exactly one — so the exported document failed XSD validation:

Element '{https://www.doclang.ai/ns/v0}layer': This element is not expected.

The merged text now keeps the first leading <layer/> and drops the rest.

This is the shape produced by a Word page footer, where a group of adjacent runs is classified as furniture.

Why not the other approach

The issue lists a second option — compute one <layer/> for the group and tell children not to emit their own. That changes the child serializers' contract and touches more code. Dropping the duplicates at the merge point is the smaller change and keeps each child's output intact.

Only leading tokens are touched, so a text run that happens to contain a literal <layer value="..."/> string is left alone.

How I checked

doclang 0.7.3's bundled doclang.xsd, validating the serializer output directly:

case	before	after
group parent, two furniture runs	2 <layer/>, invalid	1 <layer/>, valid
furniture + BODY + furniture	2 <layer/>, invalid	1 <layer/>, valid
all BODY	no <layer/>	no <layer/>, unchanged

Three tests added to tests/test_serialization_doclang.py; two of them fail without the fix and all three pass with it. The third is a guard against over-correcting the all-BODY case.

uv run pytest tests/test_serialization_doclang.py -q -k "inline_runs"

ruff check and ruff format --check are clean.

Scope note

DocLangInlineSerializer also produces invalid output when the parent TextItem carries text of its own and should_wrap is False — the runs then land outside any wrapper. That is a separate defect with its own report and is deliberately not touched here.

[github-actions]

✅ DCO Check Passed

Thanks @eren-YU, all your commits are properly signed off. 🎉

[mergify]

Merge Protections

🔴 1 of 2 protections blocking · waiting on 👀 reviews

	Protection	Waiting on
🔴	Require two reviewer for test updates	👀 reviews
🟢	Enforce conventional commit	—

🔴 Require two reviewer for test updates

Waiting for

• #approved-reviews-by >= 2

This rule is failing.

When test data is updated, we require two reviewers

• #approved-reviews-by >= 2

---

Show 1 satisfied protection

🟢 Enforce conventional commit

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?: