feat(template): add commit_groups filter for ordered group rendering

[ChrisJr404]

Closes #9.

Background

commits | group_by(attribute="group") is Tera's built-in filter, and it backs the grouped result with a BTreeMap. That means groups always come out in alphabetical (or emoji-codepoint) order, regardless of the order in which they were declared in commit_parsers. The thread on #9 covers this and the workaround that's spread across the ecosystem: prefix group names with <!-- 0 -->, <!-- 1 -->, ..., then strip the comments with striptags. The project's own cliff.toml uses that workaround.

What this PR does

It leaves group_by alone (so existing templates keep working byte-for-byte) and adds a new filter, commit_groups, that returns an array of { name, commits } records instead of a map. Because the result is an array, iteration order is deterministic and template-defined.

The filter has two modes:

• No groups argument: groups appear in the order they first occur in the commit list, which matches commit chronology.
• With a groups=... argument (an array of group names): groups are sorted to match that order. Any group not in groups is appended afterward in first-appearance order.

To make the common case ergonomic, Changelog::build now injects the ordered list of group names from commit_parsers into the template context as commit_parsers_groups. So a user template can simply write:

{% for entry in commits | commit_groups(groups=commit_parsers_groups) %}
### {{ entry.name }}
{% for commit in entry.commits %}- {{ commit.message }}
{% endfor %}{% endfor %}

and get groups in the order they wrote them in cliff.toml, with no <!-- N --> hack and no striptags filter. With the same input, switching to the existing group_by filter still produces alphabetical order, so this is purely additive.

Reproducer from the issue

Given the configuration from arthrarnld's comment:

commit_parsers = [
  { message = "^feat", group = ":rocket: New features" },
  { message = "^fix",  group = ":bug: Bug fixes" },
  { message = "^perf", group = ":zap: Performance" },
  { message = "^chore", group = ":gear: Miscellaneous" },
]

and four commits in chore -> fix -> perf -> feat order, the existing group_by template renders:

### :bug: Bug fixes
### :gear: Miscellaneous
### :rocket: New features
### :zap: Performance

After this PR, switching the template to commit_groups(groups=commit_parsers_groups) renders:

### :rocket: New features
### :bug: Bug fixes
### :zap: Performance
### :gear: Miscellaneous

Tests

• template::test::test_commit_groups_filter_preserves_first_appearance_when_no_groups
• template::test::test_commit_groups_filter_uses_groups_argument
• template::test::test_commit_groups_filter_appends_unknown_groups
• template::test::test_commit_groups_filter_skips_null_groups
• changelog::test::changelog_group_order_matches_commit_parsers (regression test using the issue reproducer end-to-end through Changelog::generate)

Notes

• group_by is unchanged. Default templates are unchanged. Adoption is opt-in via the new filter.
• The commit_parsers_groups context key is set by Changelog::build before any user-supplied context is added, so add_context("commit_parsers_groups", ...) still wins if a caller wants to override it.
• This builds on the design discussion from the closed feat: add tera filter for ordered commit groups #339 PR (orhun's commit-groups-filter-improved branch) but keeps the state on the Template instance instead of a global static, and uses tera's filter-arg mechanism to plumb the order list through.

Test plan

• cargo test --workspace --lib (all new tests pass; pre-existing repo::test::* failures reproduce on main and are env-dependent)
• cargo clippy --all-targets --workspace introduces no new lints
• cargo fmt --check introduces no new diffs against the existing baseline
• End-to-end smoke test with the binary against a 4-commit repo confirms three modes: explicit-order (matches commit_parsers), no-args (chronology), and existing group_by (alphabetical, unchanged)

[codecov-commenter]

Codecov Report

❌ Patch coverage is 92.30769% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 49.88%. Comparing base (5d3a670) to head (9f0f501).

Files with missing lines	Patch %	Lines
git-cliff-core/src/changelog.rs	81.82%	2 Missing ⚠️
git-cliff-core/src/template.rs	96.43%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1518      +/-   ##
==========================================
+ Coverage   48.90%   49.88%   +0.98%     
==========================================
  Files          26       26              
  Lines        2272     2310      +38     
==========================================
+ Hits         1111     1152      +41     
+ Misses       1161     1158       -3     

Flag	Coverage Δ
unit-tests	49.88% <92.31%> (+0.98%)	⬆️

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

☔ View full report in Codecov by Sentry.
📢 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.