Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
The table of contents is too big for display.
Diff view
Diff view
  •  
  •  
  •  
6 changes: 2 additions & 4 deletions .claude/skills/build/SKILL.md
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
name: build
description: Build and locate the PerformanceMonitor executables
argument-hint: [lite|darling|viewer|dashboard|installer]
argument-hint: [lite|darling|viewer]
disable-model-invocation: false
---

Expand All @@ -18,11 +18,9 @@ Build a component of PerformanceMonitor and report the result. All paths are rel
| `lite` | `Lite/PerformanceMonitorLite.csproj` | `net10.0-windows` |
| `darling` or `service` | `Darling/PerformanceMonitor.Darling.Service/PerformanceMonitor.Darling.Service.csproj` | `net10.0` |
| `viewer` | `Darling/PerformanceMonitor.Darling.Viewer/PerformanceMonitor.Darling.Viewer.csproj` | `net10.0-windows` |
| `dashboard` or `full` | `deprecated/Dashboard/Dashboard.csproj` | DEPRECATED since v3.3.0 |
| `installer` | `deprecated/Installer/PerformanceMonitorInstaller.csproj` | DEPRECATED since v3.3.0 |
| (empty) | build whatever is in the current working directory, or ask | |

Lite, the Darling service, and the Darling Viewer are the shipping artifacts. The full Dashboard and the CLI Installer moved to `deprecated/` in v3.3.0. They still build and their tests still run in CI, but they ship no release artifacts and get bug-fix support only.
Lite, the Darling service, and the Darling Viewer are the shipping artifacts. The full Dashboard and the CLI Installer are frozen in `deprecated/dashboard-installer-source.zip`. Nothing builds, tests or releases them from this repository, so there is no `dashboard` or `installer` argument. `deprecated/README.md` says how to build them from the zip.

## Steps

Expand Down
6 changes: 3 additions & 3 deletions .claude/skills/release-checklist/SKILL.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ disable-model-invocation: false

Run through the full release prep checklist for PerformanceMonitor. `$ARGUMENTS` is the target version (e.g., `3.3.0`).

> **Scope (since v3.3.0):** the Full Dashboard and the CLI Installer are DEPRECATED and OUT of this process — they ship no release artifacts and get no live install/upgrade testing here. The Darling service zip and the Linux tarball DO get live install and upgrade testing, in section 8a. They still build and their test suites still run in CI on the release event, and they inherit the single `<Version>` with everything else in the tree (below). The release artifacts are: Lite (zip + Velopack Setup.exe), the Darling service zip, and the Darling Viewer Setup.exe.
> **Scope (since v3.3.0):** the Full Dashboard and the CLI Installer are DEPRECATED and OUT of this process — they ship no release artifacts and get no live install/upgrade testing here. The Darling service zip and the Linux tarball DO get live install and upgrade testing, in section 8a. Their source is frozen in `deprecated/dashboard-installer-source.zip`: nothing builds or tests it in CI, and it does not take the single `<Version>` (below). The release artifacts are: Lite (zip + Velopack Setup.exe), the Darling service zip, and the Darling Viewer Setup.exe.

## Checklist

Expand Down Expand Up @@ -68,7 +68,7 @@ Do NOT add internal/implementation changes — only user-facing features and beh

### 3. Upgrade Scripts — Schema Change Audit (conditional)

Applies ONLY when `install/` or `upgrades/` changed since the last release (the deprecated Dashboard edition remains on bug-fix support, so its schema discipline still holds when it IS touched):
Applies ONLY when `install/` or `upgrades/` changed since the last release. The frozen Dashboard in the zip reads this schema, so its schema discipline still holds when it IS touched:

1. All schema changes MUST go through the `upgrades/{from}-to-{to}/` folder — never ad hoc, never baked into base install scripts. View-only changes flow through `CREATE OR ALTER` re-runs and need no upgrade script.
2. **List all .sql files in the upgrade folder and compare against `upgrade.txt`.** A script that is not listed does not run.
Expand All @@ -84,7 +84,7 @@ All projects must succeed with **0 Warning(s) / 0 Error(s)** — the repo's bar

The statement-column census test fails while any entry or watched writer is pending, in every Darling test run, with or without the variable. Run `DARLING_RELEASE_CUT=1 Darling/Darling.Tests/bin/Debug/net10.0-windows/Darling.Tests.exe -method "*StatementCensus_PendingEntries_BlockARelease*"` to see the list early; the cut proceeds only when it passes. The release workflow (build.yml, the "Statement filter release gate" step) runs the same test with the variable set before it publishes or packages anything, so a release cannot ship while an entry is pending even if this step is skipped; run it here first to see the list early.

(The deprecated Installer/Dashboard test suites run in CI on the release event with the CI's own filter; they are no longer run from this checklist.)
(The frozen Installer and Dashboard test suites are not built or run anywhere. See `deprecated/README.md`.)

### 5. Cloud Platform Testing (shared collector layer — test via DARLING)

Expand Down
7 changes: 2 additions & 5 deletions .github/ISSUE_TEMPLATE/bug_report.yml
Original file line number Diff line number Diff line change
Expand Up @@ -12,8 +12,6 @@ body:
options:
- Lite
- Darling
- Full Dashboard (deprecated)
- Installer (deprecated)
- SQL collection scripts
validations:
required: true
Expand All @@ -40,7 +38,7 @@ body:
id: windows-version
attributes:
label: Windows Version
description: The OS where the Lite, Darling, or Dashboard app is running.
description: The OS where the Lite or Darling app is running.
placeholder: "e.g., Windows 11 23H2, Windows Server 2022"
validations:
required: true
Expand Down Expand Up @@ -85,7 +83,7 @@ body:
id: errors
attributes:
label: Error Messages / Log Output
description: Paste any error messages or relevant log entries. Dashboard logs are in %LOCALAPPDATA%\PerformanceMonitor\Logs\; Lite logs are in %LOCALAPPDATA%\PerformanceMonitorLite-Data\logs\ (on versions before the data-root move, %LOCALAPPDATA%\PerformanceMonitorLite\logs\); Darling logs are in %ProgramData%\PerformanceMonitorDarling\logs\.
description: Paste any error messages or relevant log entries. Lite logs are in %LOCALAPPDATA%\PerformanceMonitorLite-Data\logs\ (on versions before the data-root move, %LOCALAPPDATA%\PerformanceMonitorLite\logs\); Darling logs are in %ProgramData%\PerformanceMonitorDarling\logs\.
render: text
validations:
required: false
Expand Down Expand Up @@ -115,7 +113,6 @@ body:
description: |
Anything else that might help:
- Did this work before? If so, what changed?
- Is SQL Server Agent running? (deprecated Full Dashboard edition only)
- Are you monitoring Azure SQL DB, Azure MI, or AWS RDS?
validations:
required: false
1 change: 0 additions & 1 deletion .github/ISSUE_TEMPLATE/question.yml
Original file line number Diff line number Diff line change
Expand Up @@ -11,7 +11,6 @@ body:
options:
- Lite
- Darling
- Full Dashboard (deprecated)
- General / setup
validations:
required: true
Expand Down
2 changes: 0 additions & 2 deletions .github/PULL_REQUEST_TEMPLATE.md
Original file line number Diff line number Diff line change
Expand Up @@ -10,8 +10,6 @@ A clear description of the change and why it's being made.
- [ ] Darling Tests
- [ ] SQL collection scripts
- [ ] Documentation
- [ ] Full Dashboard (deprecated)
- [ ] CLI Installer (deprecated)

## How was this tested?

Expand Down
17 changes: 3 additions & 14 deletions .github/scripts/ci-plan.py
Original file line number Diff line number Diff line change
Expand Up @@ -21,8 +21,6 @@
actionlint true when a workflow file changed
check_web_js true when a dashboard JavaScript file changed
pg_old_targets true when a pull request changed a file the old-PostgreSQL-target job covers (#6985, PG_OLD_TARGET_PATHS); never on a push, the nightly runs it
dashboard_tests true when the build job builds and runs deprecated/Dashboard.Tests (#6955): ci-select.py's build_dashboard_tests
installer_tests true when the build job builds and runs deprecated/Installer.Tests (#6955): ci-select.py's build_installer_tests

The pattern sets are NOT copied here. The areas come from ci-select.py's Rules, which reads the docs allowlist and the
area filters from the same files dorny/paths-filter reads, and the Lite-reads classes from ci-select.py's
Expand Down Expand Up @@ -164,7 +162,7 @@ def changed_test_classes(rules, files: "list[str]", catalog: "dict | None") -> "

def classify(rules, event: str, base: str, files: "list[str]", catalog: "dict | None" = None) -> dict:
"""The areas of one change: {areas, darling_scope, lite_scope, why}."""
everything = {"areas": "all", "darling_scope": "full", "lite_scope": "full", "dashboard_tests": True, "installer_tests": True}
everything = {"areas": "all", "darling_scope": "full", "lite_scope": "full"}
if event != "pull_request":
return {**everything, "reason": f"a {event} runs every area."}
if base != "dev":
Expand All @@ -178,10 +176,9 @@ def classify(rules, event: str, base: str, files: "list[str]", catalog: "dict |
return {**everything, "reason": f"no area knows {odd[0]}, so every area runs."}

a = rules.areas(files)
area_names = ("root", "core", "installer_core", "dashboard", "lite", "installer", "darling")
area_names = ("root", "core", "lite", "darling")
if a["all_count"] == a["docs_count"] and not any(a[n] for n in area_names):
return {"areas": "docs-only", "darling_scope": "none", "lite_scope": "none",
"dashboard_tests": False, "installer_tests": False,
"reason": "only documentation changed."}

cs = _CS
Expand Down Expand Up @@ -318,7 +315,7 @@ def plan(event: str, base: str, head_ref: str, draft: bool, labels: "list[str]",
out = {"run": "true" if outcome == "run" else "false", "outcome": outcome, "why": why,
"areas": "none", "darling_scope": "none", "lite_scope": "none", "build": "false",
"darling_classes": "", "lite_classes": "", "test_minutes": "0", "legs": "0", "legs_json": "[]",
"actionlint": "false", "check_web_js": "false", "pg_old_targets": "false", "dashboard_tests": "false", "installer_tests": "false"}
"actionlint": "false", "check_web_js": "false", "pg_old_targets": "false"}
if outcome != "run":
return out

Expand All @@ -331,14 +328,6 @@ def plan(event: str, base: str, head_ref: str, draft: bool, labels: "list[str]",
out["actionlint"] = _bool(any(f.startswith(".github/workflows/") for f in files))
out["check_web_js"] = _bool(any(WEB_JS_MARKER in "/" + f for f in files))
out["pg_old_targets"] = _bool(event == "pull_request" and any(f.startswith(PG_OLD_TARGET_PATHS) for f in files))
# A run that does not narrow (a push, another base, an unknown path) builds both; docs-only builds neither. Otherwise
# ci-select.py's decide() answers from the same areas the workflow's own filter reads (#6955).
if "dashboard_tests" in c:
out.update(dashboard_tests=_bool(c["dashboard_tests"]), installer_tests=_bool(c["installer_tests"]))
else:
d = _CS.decide(files, event, rules)
out.update(dashboard_tests=_bool(d["build_dashboard_tests"]), installer_tests=_bool(d["build_installer_tests"]))

if out["build"] == "true":
catalog = catalog if catalog is not None else load_catalog()
picks = {"darling": c.get("darling_classes", []), "lite": c.get("lite_classes", [])}
Expand Down
16 changes: 2 additions & 14 deletions .github/scripts/ci-select.py
Original file line number Diff line number Diff line change
Expand Up @@ -271,12 +271,10 @@ def decide(files: list[str], event: str, rules: Rules) -> dict:
release = event == "release"
all_count, docs_count = a["all_count"], a["docs_count"]
docs_only = all_count > 0 and all_count == docs_count
area_names = ("root", "core", "installer_core", "dashboard", "lite", "installer", "darling")
area_names = ("root", "core", "lite", "darling")
fastpath = (not release) and docs_only and not any(a[n] for n in area_names)

build_lite = release or a["lite"] or a["core"] or a["root"]
build_installer = release or a["installer"] or a["installer_core"] or a["root"]
build_dashboard = release or a["dashboard"] or a["core"] or a["installer_core"] or a["root"]
build_darling = release or a["darling"] or a["core"] or a["root"]

pg_run = (not release) and a["gate_darling"]
Expand All @@ -290,8 +288,6 @@ def decide(files: list[str], event: str, rules: Rules) -> dict:
"guard_run": (not release) and not docs_only,
"docs_fastpath": fastpath,
"build_lite_tests": build_lite,
"build_installer_tests": build_installer,
"build_dashboard_tests": build_dashboard,
"build_darling_tests": build_darling,
"darling_scope": darling_scope(
event, build_darling, pg_run, release or a["darling_full"] or a["core"] or a["root"]),
Expand Down Expand Up @@ -377,16 +373,12 @@ def _holders_of(lines: list[str]) -> set[str]:
SUITE_DIRS = {
"darling": os.path.join("Darling", "Darling.Tests"),
"lite": "Lite.Tests",
"dashboard": os.path.join("deprecated", "Dashboard.Tests"),
"installer": os.path.join("deprecated", "Installer.Tests"),
}
SUITE_BY_PREFIX = {
"Darling.Tests": "darling",
"PerformanceMonitor.Darling.Tests": "darling",
"Lite.Tests": "lite",
"PerformanceMonitorLite.Tests": "lite",
"Dashboard.Tests": "dashboard",
"Installer.Tests": "installer",
}


Expand Down Expand Up @@ -557,10 +549,6 @@ def class_selected(suite: str, traits: set[str], d: dict) -> bool:
if d["lite_mode"] in ("full", "release"):
return True
return d["lite_mode"] == "reads" and "Reads=Darling" in traits
if suite == "dashboard":
return d["build_dashboard_tests"]
if suite == "installer":
return d["build_installer_tests"]
return True


Expand Down Expand Up @@ -1568,7 +1556,7 @@ def git_drift(map_sha: str, base: str, root: str = ROOT) -> "list[str] | None":
def selection_args(rules: Rules) -> dict:
"""The keep-full and documentation patterns, read from the same filters the workflow evaluates."""
keep = list(KEEP_FULL)
for name in ("root", "core", "installer_core"):
for name in ("root", "core"):
# #5951: the shared projects are judged by map_select()'s shared_change_reason(), not as keep-full wholesale.
keep.extend(p for p in rules.build.get(name, []) if not p.startswith(SHARED_PROJECT_DIRS))
return {"keep_full": tuple(dict.fromkeys(keep)), "doc_patterns": tuple(rules.build["docs"])}
Expand Down
20 changes: 0 additions & 20 deletions .github/scripts/tests/ci-plan-test.py
Original file line number Diff line number Diff line change
Expand Up @@ -37,14 +37,8 @@ def check(name, actual, expected):
- 'global.json'
core:
- 'PerformanceMonitor.Common/**/!(*.md)'
installer_core:
- 'PerformanceMonitor.Installer/**/!(*.md)'
dashboard:
- 'deprecated/Dashboard/**/!(*.md)'
lite:
- 'Lite/**/!(*.md)'
installer:
- 'deprecated/Installer/**/!(*.md)'
darling:
- 'Darling/**/!(*.md)'
- 'Lite/**/*.xaml'
Expand Down Expand Up @@ -172,20 +166,6 @@ def run(files, event="pull_request", base="dev", head="feature/x", draft=False,
r = run(["global.json"])
check("root build file: all", r["areas"], "all")

r = run(["deprecated/Installer/X.cs"])
check("deprecated only: no area", r["areas"], "none")
check("deprecated only: builds, no legs", (r["build"], r["legs"]), ("true", "0"))
check("deprecated installer file: installer tests only", (r["installer_tests"], r["dashboard_tests"]), ("true", "false"))
r = run(["deprecated/Dashboard/X.cs"])
check("deprecated dashboard file: dashboard tests only", (r["dashboard_tests"], r["installer_tests"]), ("true", "false"))
check("shared library: dashboard tests, not installer tests", (run(["PerformanceMonitor.Common/X.cs"])["dashboard_tests"],
run(["PerformanceMonitor.Common/X.cs"])["installer_tests"]), ("true", "false"))
check("darling file only: neither suite", (run(["Darling/X.cs"])["dashboard_tests"], run(["Darling/X.cs"])["installer_tests"]),
("false", "false"))
check("docs only: neither suite", (run(["README.md"])["dashboard_tests"], run(["README.md"])["installer_tests"]), ("false", "false"))
check("unknown path: both suites", (run(["tools/odd.txt"])["dashboard_tests"], run(["tools/odd.txt"])["installer_tests"]),
("true", "true"))

# --- all areas -------------------------------------------------------------------------------------------------
r = run(["tools/odd.txt"])
check("unknown path: all", (r["areas"], r["darling_scope"], r["lite_scope"]), ("all", "full", "full"))
Expand Down
Loading
Loading