@taqwright/lime-cli 0.30.0 → 0.32.0
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/README.md +44 -2
- package/dist/index.js +355 -265
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -87,6 +87,30 @@ With repository access, you can also run straight from a checkout — see
|
|
|
87
87
|
plaintext `http://` so your `LIME_CI_TOKEN` is never sent in the clear.
|
|
88
88
|
`http://localhost` / `127.0.0.1` is allowed for local development.
|
|
89
89
|
|
|
90
|
+
### Set up a repo: `lime-cli init`
|
|
91
|
+
|
|
92
|
+
In your test repository, with `LIME_CI_TOKEN` (and your device-cloud keys, if you use one) in a
|
|
93
|
+
git-ignored `.env`:
|
|
94
|
+
|
|
95
|
+
```sh
|
|
96
|
+
npx @taqwright/lime-cli init
|
|
97
|
+
```
|
|
98
|
+
|
|
99
|
+
It runs doctor's checks, then checks that you can push to the repository (`gh`) and that your
|
|
100
|
+
BrowserStack or TestMu AI credentials and app ids work (free, nothing is billed). Then it offers
|
|
101
|
+
each fix it can make, and makes it only with a yes:
|
|
102
|
+
|
|
103
|
+
- saves your LIME project in `lime.json`, so no command has to name it again
|
|
104
|
+
- adds `.env` to `.gitignore` if your token there would be committed
|
|
105
|
+
- installs the taqwright skill at `.claude/skills/taqwright/SKILL.md`, where Claude Code loads it
|
|
106
|
+
by itself (commit it, so the whole team has it)
|
|
107
|
+
- adds a short Taqwright section to `AGENTS.md`, for Codex, Copilot and Cursor
|
|
108
|
+
|
|
109
|
+
It ends with **Ready** and the sentence to type in your coding agent, for example
|
|
110
|
+
*"Automate TC-12 on Android and iOS"*. Run it again any time: it changes nothing that is already
|
|
111
|
+
right, and offers an updated skill when the server's has moved on. A coding agent running it
|
|
112
|
+
(no terminal) gets a list of what would be written; `--yes` writes it.
|
|
113
|
+
|
|
90
114
|
### Check your setup: `lime-cli doctor`
|
|
91
115
|
|
|
92
116
|
One command checks all of the above, plus your test repo, before you spend a credit. It needs only
|
|
@@ -312,6 +336,19 @@ test fails it; `--diff <file>` takes a unified diff for CI.
|
|
|
312
336
|
flagging direct device use. It is the repo owner's declaration, committed and reviewed; the
|
|
313
337
|
recording rule and the assertion rule still apply.
|
|
314
338
|
|
|
339
|
+
**Your own locator helpers** are read too. A one-line wrapper over an Appium call
|
|
340
|
+
(`By hint(String v) { return AppiumBy.xpath("//*[@hint='" + v + "']"); }`) is picked up from the
|
|
341
|
+
repo's Java/Kotlin, so `hint("Username")` is checked as the xpath it builds, and the brief tells
|
|
342
|
+
the agent to write locators the same way. A helper lime-cli cannot read is declared in `lime.json`:
|
|
343
|
+
|
|
344
|
+
```json
|
|
345
|
+
{ "locatorHelpers": { "byLabel": "accessibility id",
|
|
346
|
+
"byHint": { "using": "xpath", "template": "//*[@hint='{0}']" } } }
|
|
347
|
+
```
|
|
348
|
+
|
|
349
|
+
A new locator written through a helper that is neither inferred nor declared **fails** `check`:
|
|
350
|
+
a locator that could not be read was not checked, and `check` never says it was.
|
|
351
|
+
|
|
315
352
|
`--lime-dir` writes `.lime/` next to your project: the recorded steps the agent converts
|
|
316
353
|
from, a `run.json` of what was recorded (goal, platform, cloud device), and the briefs.
|
|
317
354
|
**No credential is ever written there** — cloud credentials are re-read from
|
|
@@ -328,8 +365,8 @@ make that a failure instead, or `--agent none` to skip adaptation deliberately.
|
|
|
328
365
|
|
|
329
366
|
### Your suite is not taqwright
|
|
330
367
|
|
|
331
|
-
`lime-cli export` handles any of the frameworks LIME supports — pytest, JUnit 5,
|
|
332
|
-
WebdriverIO, C# NUnit, taqwright — and works the rest out from the repo itself:
|
|
368
|
+
`lime-cli export` handles any of the frameworks LIME supports — pytest, JUnit 5, Java +
|
|
369
|
+
Cucumber, RSpec, WebdriverIO, C# NUnit, taqwright — and works the rest out from the repo itself:
|
|
333
370
|
|
|
334
371
|
```sh
|
|
335
372
|
cd your-test-repo
|
|
@@ -348,6 +385,11 @@ claude -p "Read .lime/ADAPT.md and follow it exactly." \
|
|
|
348
385
|
--output-format stream-json --verbose --permission-mode acceptEdits
|
|
349
386
|
```
|
|
350
387
|
|
|
388
|
+
On a **Cucumber** suite the test is a scenario in a `.feature` file: the brief lists every step
|
|
389
|
+
the project already defines so the agent reuses them, puts new steps beside the existing step
|
|
390
|
+
classes and keeps locators in page objects. `lime-cli check` treats the scenario and its step
|
|
391
|
+
definitions as the test, and `lime-cli verify` runs only the new scenario, by name.
|
|
392
|
+
|
|
351
393
|
Override anything it guessed with `--framework` and `--dest`. If you would rather lime-cli drive
|
|
352
394
|
the agent for you, that is `lime-cli adapt`.
|
|
353
395
|
|