Group the selective-CI test-selector comment by trigger and lead with what runs

[radical]

What this changes

The Tests selector PR comment posted by tools/SelectTests (the
selective-CI advisory comment) was hard to read. It listed every selected
test project on its own line with the selection reason appended, so when one
change fans out across the project graph the same reason repeats on dozens
of lines. A reviewer couldn't tell at a glance which projects and jobs run,
or which change drove the bulk of the selection.

Example — the old comment on #18332 had 46 near-identical lines like:

- `Aspire.Hosting.Azure.Kusto.Tests` — via graph from `tests/Aspire.Hosting.Tests/Dashboard/DashboardServiceHostTests.cs`
- `Aspire.Hosting.Blazor.Tests` — via graph from `tests/Aspire.Hosting.Tests/Dashboard/DashboardServiceHostTests.cs`
- … 30 more identical reasons …

The fix

Restructure the comment into what runs, then how it was chosen:

• Lead with what runs — a flat list of selected jobs, then a flat list of
  selected test projects. Full impact is visible up front, even with many
  changed files.
• Then how it was chosen — group the selected projects under each trigger
  (a changed file and its graph fan-out under one heading, an affected
  project, or a derived test). Each reason is stated once instead of per
  project; large fan-outs collapse into <details>; a per-job table names
  what triggered each job. A headline ⚠️ call-out names the single change
  driving a large share of a big selection.

Rendered on #18332's inputs, the 46 projects collapse to a handful of
trigger groups, with the headline surfacing that 33 of 46 come from one
changed test file.

Call-outs

• Full attribution is preserved — a project selected by several triggers
  appears under each group, so no "why" is dropped (the old "list every
  cause" guarantee still holds, just inverted).
• Updated the CLI and Layer 1 integration tests to assert the new structure,
  and docs/ci/test-trigger-selector-design.md to describe it.
• No change to the selection logic, the job booleans, the step-summary audit
  detail, or the one-comment-per-commit posting behavior — only the comment's
  body layout.

[github-actions]

🚀 Dogfood this PR with:

⚠️ WARNING: Do not do this without first carefully reviewing the code of this PR to satisfy yourself it is safe.

curl -fsSL https://github.com/microsoft/aspire/main/eng/scripts/get-aspire-cli-pr.sh | bash -s -- 18368

Or

• Run remotely in PowerShell:

iex "& { $(irm https://raw.githubusercontent.com/microsoft/aspire/main/eng/scripts/get-aspire-cli-pr.ps1) } 18368"

[Copilot]

Pull request overview

This PR restructures the Tests selector advisory PR comment produced by tools/SelectTests (the selective-CI tool). The old comment listed each selected test project on its own line with the selection reason appended, so a single change fanning out across the project graph repeated the same reason on dozens of lines. The new layout leads with what runs (flat lists of selected jobs and projects) and then explains how it was chosen by inverting the per-project causes into per-trigger groups, collapsing large fan-outs into <details>, adding a per-job reasons table, and surfacing a ⚠️ headline when one trigger drives a large share of a big selection. There is no change to selection logic, job booleans, the step-summary, or the one-comment-per-commit posting behavior — only the comment body layout.

Changes:

• Rewrote WriteSelectionComment and added grouping/rendering helpers (AppendSelectionRationale, CauseGroupKey/Header/Descriptor, RenderProjectList, MemberWithHops, JobCausesText, StripJobPrefix, WriteCommentFile); removed the now-unused AllCausesSuffix.
• Updated the CLI and Layer 1 integration tests to assert the new comment structure.
• Updated docs/ci/test-trigger-selector-design.md to describe the grouped, trigger-first layout.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
tools/SelectTests/Program.cs	Core rewrite of the PR comment renderer: flat "what runs" lists, per-trigger grouping with direct/graph/other buckets, headline call-out, and a job-reasons table.
tests/Infrastructure.Tests/TestTriggerMap/SelectTestsCliTests.cs	Updated assertions to match the new headings and multi-trigger grouping; comments revised accordingly.
tests/Infrastructure.Tests/TestTriggerMap/SelectTestsLayer1IntegrationTests.cs	Updated graph-fan-out assertions to the new "via the project graph" grouping output.
docs/ci/test-trigger-selector-design.md	Documentation updated to describe the trigger-first, grouped comment layout.

Notes from the review (both nit-level): the other bucket uses a fixed "tests" label that reads as "1 tests" for a single project (inconsistent with the careful pluralization elsewhere in the method), and the new headline call-out / <details> collapse thresholds are not covered by any test.

[github-actions]

Retrying the failed CI jobs for this pull request from the CI run attempt. The rerun is being tracked in the rerun attempt.

[github-actions]

Retrying the failed CI jobs for this pull request from the CI run attempt. The rerun is being tracked in the rerun attempt.

[radical]

Superseded by #18376, which is opened from a branch in this repo so the Tests selector selection comment can actually post (a fork PR runs the pull_request workflow with a read-only GITHUB_TOKEN, so it cannot comment on the base repo). Same change, plus fork-aware handling that surfaces the selection in the job step summary when posting is not permitted. Closing this one.