tolkira 0.1.1 → 0.2.0
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.
- package/DESIGN.md +12 -7
- package/GLOSSARY.md +15 -0
- package/README.md +30 -16
- package/dist/cli.js +531 -92
- package/dist/cli.js.map +3 -3
- package/dist/index.d.ts +1 -1
- package/dist/index.js +472 -88
- package/dist/index.js.map +3 -3
- package/dist/input/junit.d.ts +3 -1
- package/dist/input/load.d.ts +11 -0
- package/dist/render/escape.d.ts +1 -1
- package/dist/render/html.d.ts +1 -0
- package/dist/report/model.d.ts +4 -1
- package/dist/summary.d.ts +1 -0
- package/docs/COMPATIBILITY.md +14 -4
- package/package.json +2 -1
package/DESIGN.md
CHANGED
|
@@ -103,19 +103,22 @@ describe the viewport, not the font.
|
|
|
103
103
|
| Top bar / tabs | `.topbar`, `.view` | Report and Details; active tab is a text state, not a card |
|
|
104
104
|
| Run header | `.runHead`, `.health` | Title, metadata, pass rate, segmented status bar |
|
|
105
105
|
| Toolbar | `.toolbar`, `.search`, `.filter` | Search and status filters; search opens matching features and scenarios; filters only for present statuses |
|
|
106
|
-
|
|
|
106
|
+
| Triage index | `.triage` | Needs-attention rows with status text, counts and up to three scenario links; five rows visible, up to fifteen more behind a disclosure |
|
|
107
|
+
| Source strip | `.sources` | Per-input cards for multi-file runs with real provenance; status, counts and a proportion bar; hidden below 560 px, where the feature source badge remains |
|
|
108
|
+
| Feature | `details.feature` | Native disclosure; status, name, counts, proportion bar, source badge, URI |
|
|
107
109
|
| Scenario | `details.scenario` | Fixed thumbnail gutter, status, name, tags, steps, duration |
|
|
108
110
|
| Step | `.step` | Status glyph plus screen-reader label, keyword, text, error, doc string, focusable table, attachments |
|
|
109
111
|
| Attachment | `.shot`, `.fileText`, `.fileLink` | Images inline; text in `<pre>`; binary as download; alt names the scenario and step |
|
|
110
|
-
| Details | `.analysis` | Feature
|
|
112
|
+
| Details | `.analysis` | Feature-by-status matrix, status distribution, Run / Cucumber environment / Custom metadata blocks, agent summary |
|
|
111
113
|
|
|
112
114
|
## Interaction rules
|
|
113
115
|
|
|
114
116
|
- Feature and scenario summaries are the only click targets that toggle; a
|
|
115
117
|
click inside the detail does not close it.
|
|
116
118
|
- Chevron rotates on open; `prefers-reduced-motion` removes the transition.
|
|
117
|
-
- Search and filters: hide non-matches
|
|
118
|
-
update a polite live region, show an empty state
|
|
119
|
+
- Search and filters: hide non-matches and open matching features; search also
|
|
120
|
+
opens matching scenarios, update a polite live region, show an empty state
|
|
121
|
+
when nothing matches.
|
|
119
122
|
- Switching tabs scrolls to the top; the previous panel's scroll position is
|
|
120
123
|
not restored.
|
|
121
124
|
- The lightbox locks background scrolling, restores focus to its trigger on
|
|
@@ -124,7 +127,8 @@ describe the viewport, not the font.
|
|
|
124
127
|
- Status is never colour-only: scenario badges carry text, step markers carry a
|
|
125
128
|
shape glyph plus a screen-reader label, and forced-colours mode keeps the
|
|
126
129
|
glyph, marker border and disclosure chevron visible.
|
|
127
|
-
- The failure is open by default;
|
|
130
|
+
- The failure is open by default; screenshot-bearing scenarios and their
|
|
131
|
+
features also open; everything else is scannable when collapsed.
|
|
128
132
|
- No right-click-only actions; use visible buttons or a `…` menu. Keyboard and
|
|
129
133
|
touch must reach the same actions.
|
|
130
134
|
- No modal for a task that inline disclosure can solve.
|
|
@@ -159,6 +163,8 @@ Direction A is implemented in production:
|
|
|
159
163
|
- warm paper tokens, hairline sections and monospace measurements;
|
|
160
164
|
- fixed thumbnail gutter on every scenario row, also on mobile;
|
|
161
165
|
- chevrons, failed-first disclosures and search that opens matching features;
|
|
166
|
+
- exception-first Report order, a triage index and a source strip;
|
|
167
|
+
- feature counts/proportion bars and a feature-by-status matrix in Details;
|
|
162
168
|
- `undefined` is a distinct status, not a neutral one;
|
|
163
169
|
- Details tab with producer, runtime, OS, CPU, protocol and user metadata;
|
|
164
170
|
- empty state for search/filter results;
|
|
@@ -167,8 +173,7 @@ Direction A is implemented in production:
|
|
|
167
173
|
CSS lives in `src/render/report.css`; `scripts/build-styles.mjs` generates
|
|
168
174
|
`src/render/styles.ts` for the inline report.
|
|
169
175
|
|
|
170
|
-
Still open:
|
|
171
|
-
Advanced.
|
|
176
|
+
Still open: charts/trends in Details and optional map/compact lenses.
|
|
172
177
|
|
|
173
178
|
## Non-goals
|
|
174
179
|
|
package/GLOSSARY.md
CHANGED
|
@@ -29,6 +29,21 @@ component tree.
|
|
|
29
29
|
**Shell** — Tolkira's own wrapper around the renderer: title, metadata, summary,
|
|
30
30
|
theme, logo.
|
|
31
31
|
|
|
32
|
+
**Triage index** — the "Needs attention" block above the feature list. It shows
|
|
33
|
+
non-passing features with counts and up to three scenario links; five rows are
|
|
34
|
+
visible and up to fifteen more sit behind a disclosure.
|
|
35
|
+
|
|
36
|
+
**Source strip** — per-input cards above the feature list for multi-file runs
|
|
37
|
+
with real provenance. Each card shows the input's worst status, counts and a
|
|
38
|
+
proportion bar.
|
|
39
|
+
|
|
40
|
+
**Feature matrix** — the Details table with one row per feature and one count
|
|
41
|
+
column per status present in the run, plus total and duration.
|
|
42
|
+
|
|
43
|
+
**Provenance** — the input file a feature came from. Tolkira only uses it when
|
|
44
|
+
the feature URI belongs to exactly one input; shared or missing provenance
|
|
45
|
+
renders a flat report.
|
|
46
|
+
|
|
32
47
|
**Summary** — the stable, renderer-independent JSON document emitted by
|
|
33
48
|
`--json` / `summarizeReport` for agents and CI. It is versioned and is not the
|
|
34
49
|
HTML report.
|
package/README.md
CHANGED
|
@@ -1,19 +1,20 @@
|
|
|
1
1
|
# Tolkira
|
|
2
2
|
|
|
3
|
-
**
|
|
3
|
+
**Cucumber reports that stay readable: legacy JSON, Messages and JUnit XML in one self-contained HTML file.**
|
|
4
4
|
|
|
5
5
|
Tolkira turns test results into a single HTML file you can open offline, attach
|
|
6
6
|
to a ticket, or keep as a CI artifact. Screenshots, logs and other attachments
|
|
7
7
|
stay inside the file. There is no server, no CDN and no telemetry.
|
|
8
8
|
|
|
9
|
-
Current release: **v0.
|
|
9
|
+
Current release: **v0.2.0** · [npm](https://www.npmjs.com/package/tolkira) · [Changelog](https://github.com/Silasoa/tolkira/blob/main/docs/CHANGELOG.md)
|
|
10
10
|
|
|
11
11
|
> **How this is built:** Tolkira is human-directed and largely AI-generated.
|
|
12
|
-
> The maintainer sets the product direction
|
|
13
|
-
>
|
|
14
|
-
>
|
|
15
|
-
>
|
|
16
|
-
>
|
|
12
|
+
> The maintainer sets the product direction and approves releases. Most code,
|
|
13
|
+
> tests and documentation are drafted with large language models, then checked
|
|
14
|
+
> by automated tests and AI-assisted review passes. The maintainer reviews the
|
|
15
|
+
> release gate and tests the report from time to time. AI output can be wrong;
|
|
16
|
+
> please check the report against your own results and open an issue if
|
|
17
|
+
> something looks off. [Read more](#how-this-project-is-built).
|
|
17
18
|
|
|
18
19
|
## Install
|
|
19
20
|
|
|
@@ -63,16 +64,21 @@ tolkira test-results/ --json
|
|
|
63
64
|
|
|
64
65
|
- **One file that works offline.** CSS, JavaScript and attachments are inline.
|
|
65
66
|
No network requests, no remote fonts, no analytics.
|
|
66
|
-
- **The failure first.** Failed scenarios are open by default,
|
|
67
|
-
and
|
|
67
|
+
- **The failure first.** Failed scenarios are open by default, and a compact
|
|
68
|
+
triage index lists non-passing features and scenarios before passed ones,
|
|
69
|
+
with links that open the scenario detail.
|
|
70
|
+
- **Multi-file context.** A source strip shows per-input status, counts and a
|
|
71
|
+
proportion bar when the run has more than one input file and provenance is
|
|
72
|
+
unambiguous. On phones the strip is hidden and the feature row keeps the
|
|
73
|
+
source.
|
|
68
74
|
- **Screenshots and attachments that survive.** Images, text, video, audio and
|
|
69
75
|
files are preserved through the legacy JSON adapter.
|
|
70
76
|
- **Search and filters.** Search scenarios, features, steps and tags; filter by
|
|
71
77
|
status and tag; open matching features automatically and highlight matches.
|
|
72
78
|
- **Copy failure.** Copy a failed scenario's feature, failed step and error as
|
|
73
79
|
plain text for a ticket or an agent.
|
|
74
|
-
- **A Details view when you need it.** Status distribution, a feature
|
|
75
|
-
run metadata and the Cucumber environment (producer, runtime, OS, CPU,
|
|
80
|
+
- **A Details view when you need it.** Status distribution, a feature-by-status
|
|
81
|
+
matrix, run metadata and the Cucumber environment (producer, runtime, OS, CPU,
|
|
76
82
|
protocol).
|
|
77
83
|
- **A stable summary for agents.** `--json` and `summarizeReport()` return a
|
|
78
84
|
versioned schema with counts, durations, scenarios, steps and failure
|
|
@@ -113,6 +119,11 @@ Tolkira keeps the old input working and fixes the parts that hurt:
|
|
|
113
119
|
- a stable JSON summary for agents and CI;
|
|
114
120
|
- a small, reviewed dependency surface.
|
|
115
121
|
|
|
122
|
+
Single-file output alone is no longer unique: the official formatter and Allure
|
|
123
|
+
can produce one file. Tolkira's difference is the combination - both Cucumber
|
|
124
|
+
generations, JUnit as cheap breadth, an agent-readable summary, and no server,
|
|
125
|
+
CDN or telemetry.
|
|
126
|
+
|
|
116
127
|
The visual comparison uses the report setup from
|
|
117
128
|
[Seed-Test](https://github.com/adessoSE/Seed-Test). If you are curious about
|
|
118
129
|
Seed-Test, you are welcome to take a look; the comparison notes and screenshots
|
|
@@ -212,8 +223,9 @@ notes live in [`docs/PRODUCT.md`](https://github.com/Silasoa/tolkira/blob/main/d
|
|
|
212
223
|
|
|
213
224
|
Tolkira is an experiment in AI-assisted open source. The maintainer defines the
|
|
214
225
|
problem, the scope and the design direction; large language models draft most of
|
|
215
|
-
the code, tests and documentation;
|
|
216
|
-
|
|
226
|
+
the code, tests and documentation; automated tests and AI-assisted review passes
|
|
227
|
+
check changes; the maintainer reviews the release gate, approves versions and
|
|
228
|
+
tests the report from time to time.
|
|
217
229
|
|
|
218
230
|
The project does not name specific models in its files on purpose. The
|
|
219
231
|
maintainer configures the models and tooling. This keeps the project
|
|
@@ -229,6 +241,8 @@ version you use.
|
|
|
229
241
|
|
|
230
242
|
MIT. See [`LICENSE`](https://github.com/Silasoa/tolkira/blob/main/LICENSE).
|
|
231
243
|
|
|
232
|
-
Tolkira is not
|
|
233
|
-
|
|
234
|
-
uses
|
|
244
|
+
Tolkira is an independent project, not an official Cucumber, SmartBear or
|
|
245
|
+
Seed-Test product. The maintainer also contributes to Seed-Test; the visual
|
|
246
|
+
comparison uses Seed-Test's report setup. Cucumber and Gherkin are used
|
|
247
|
+
descriptively; the project uses the MIT-licensed Cucumber Messages format and
|
|
248
|
+
package.
|