@gadicc/loom 2.3.0 → 2.3.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"version": 1,
|
|
3
|
-
"supportDigest": "
|
|
3
|
+
"supportDigest": "8a037d6466febb97f71148bc0f9cdc764bc996bc26671e620ecfcf50aa496e88",
|
|
4
4
|
"supportSources": [
|
|
5
5
|
"src/apiKeys/betterAuth.ts",
|
|
6
6
|
"src/apiKeys/permissions.ts",
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@gadicc/loom",
|
|
3
|
-
"version": "2.3.
|
|
3
|
+
"version": "2.3.1",
|
|
4
4
|
"description": "Shared helpers and verification tooling for Next.js applications.",
|
|
5
5
|
"homepage": "https://github.com/gadicc/loom#readme",
|
|
6
6
|
"repository": {
|
|
@@ -11,7 +11,7 @@
|
|
|
11
11
|
"url": "https://github.com/gadicc/loom/issues"
|
|
12
12
|
},
|
|
13
13
|
"loom": {
|
|
14
|
-
"gitCommit": "
|
|
14
|
+
"gitCommit": "aefcdc95762cffdae8af715ce57f6344f7abc45f"
|
|
15
15
|
},
|
|
16
16
|
"type": "module",
|
|
17
17
|
"main": "./esm/mod.js",
|
package/skills/loom/SKILL.md
CHANGED
|
@@ -85,6 +85,16 @@ from the consumer's working directory.
|
|
|
85
85
|
|
|
86
86
|
## High-value code improvements
|
|
87
87
|
|
|
88
|
+
When adding or changing tests, choose the lightest tier that proves their
|
|
89
|
+
assertions. For database-related tests, use `loom-db` and its
|
|
90
|
+
[test boundaries](../loom-db/references/test-boundaries.md) guidance. Inspect
|
|
91
|
+
tests touched by ordinary feature work for avoidable database initialization;
|
|
92
|
+
the user does not need to request a separate performance review. Implement
|
|
93
|
+
small, behavior-preserving splits that fit the task and verify the retained
|
|
94
|
+
integration coverage. Record broader opportunities with their scope and
|
|
95
|
+
evidence instead of starting a fleet rewrite or changing runner/backends as a
|
|
96
|
+
side effect.
|
|
97
|
+
|
|
88
98
|
While doing requested work, look for high-value improvements in code you
|
|
89
99
|
inspect, including existing code that works but is unnecessarily verbose or
|
|
90
100
|
difficult to maintain. Behavior-preserving cleanup is valuable when it makes
|
package/skills/loom-db/SKILL.md
CHANGED
|
@@ -83,7 +83,10 @@ Use this skill when working on the shared DB stack from a consuming app.
|
|
|
83
83
|
## Validation and tests
|
|
84
84
|
|
|
85
85
|
Read [test boundaries and local resources](references/test-boundaries.md) before
|
|
86
|
-
|
|
86
|
+
adding or changing database-related tests, retargeting existing tests, changing
|
|
87
|
+
test backends or adopting resource controls. Apply the criteria during normal
|
|
88
|
+
feature work, including small opportunities to remove unnecessary DB startup
|
|
89
|
+
from tests already being touched; do not wait for a separate optimization request.
|
|
87
90
|
|
|
88
91
|
- Use `loom db schema-gen` after schema changes that affect generated Valibot
|
|
89
92
|
output.
|
|
@@ -1,5 +1,11 @@
|
|
|
1
1
|
# Database test boundaries and local resources
|
|
2
2
|
|
|
3
|
+
Apply these criteria while adding or editing tests during ordinary feature work.
|
|
4
|
+
Inspect the touched tests' runtime imports and assertion claims; make a small
|
|
5
|
+
verified boundary improvement when it fits the task. Propose broader work with
|
|
6
|
+
concrete candidates and evidence. No periodic agent, automatic fleet sweep or
|
|
7
|
+
target percentage is implied.
|
|
8
|
+
|
|
3
9
|
Retarget tests by what their assertions prove. Pure decisions use plain values
|
|
4
10
|
and a module with no runtime DB/schema import. Feature orchestration may use a
|
|
5
11
|
small repository fake only when the fake and real implementation share a
|
|
@@ -70,6 +76,17 @@ unattended performance experiments. Other users and unwrapped jobs remain
|
|
|
70
76
|
outside this policy. Do not promise OS responsiveness or speed from affinity.
|
|
71
77
|
|
|
72
78
|
Keep new controls opt-in until representative comparisons justify adoption.
|
|
79
|
+
|
|
80
|
+
A smaller dedicated local PGlite worker budget is a separate experiment from engine
|
|
81
|
+
reuse. Restricting workers can reduce simultaneous engines but still creates a
|
|
82
|
+
fresh harness per isolated file. Reusing an engine requires independent proof
|
|
83
|
+
of reset correctness, fixture/schema ownership, transaction behavior and leaked
|
|
84
|
+
state/resource cleanup. Pure and SQL jobs must share an explicit total budget
|
|
85
|
+
when run together; independent per-job caps do not provide aggregate admission.
|
|
86
|
+
The current launcher supports one finite single-project run, not a composed
|
|
87
|
+
multi-project scheduler. Do not disable isolation or share mutable databases
|
|
88
|
+
merely to make worker processes longer-lived.
|
|
89
|
+
|
|
73
90
|
The initial busy-desktop pilot found two workers slower than the unchanged
|
|
74
91
|
three-file selection, and per-run prepared SQL was slower after preparation.
|
|
75
92
|
Neither result determines the best setting for another app/device. No worker
|