v0.10.3

• ▶ Up to here on every step of the diagram, as in tinyprobe-app 0.4.13: a button at the right of each step runs the workflow up to it and shows its response, without leaving the run view.
• The lens above each step in the editor, which was "Run to here", is "▶ Up to here" too, and its tooltip says what it does: it runs the steps before this one again, stops there and shows its response.

Works with tinyprobe 0.10.0 or newer, as before; ▶ Up to here needs 0.18.0.

What's Changed

• ▶ Up to here on the diagram's steps, and as the lens's name by @kikutano in #17

Full Changelog: v0.10.2...v0.10.3

v0.10.2

• The diagram above a run fits a narrow run view: its steps shrink down to 280 pixels before the view scrolls sideways (420 before), with the shared code of tinyprobe-app 0.4.11.

Works with tinyprobe 0.10.0 or newer, as before; Run to here needs 0.18.0.

What's Changed

• Bring the shared code up to tinyprobe-app 0.4.11 by @kikutano in #16

Full Changelog: v0.10.1...v0.10.2

v0.10.1

• Long runs are fast again in the run view. Since 0.10.0 a workflow of 500 steps took minutes to show its run: the diagram measured the panel on every event of the run, which made each event slower than the one before, and the results were rebuilt once per event. They are now drawn once for all the events that arrived in between, with the shared code of tinyprobe-app 0.4.9.

Works with tinyprobe 0.10.0 or newer, as before; Run to here needs 0.18.0.

What's Changed

• Bring the shared run view up to tinyprobe-app 0.4.9: fast long runs by @kikutano in #15

Full Changelog: v0.10.0...v0.10.1

v0.10.0

• The diagram of a workflow. The run view draws the workflow above its run: each step, what it calls, and the values it saves, colored the same where later steps use them. It lights up step by step as the workflow runs, and a click on a step shows its request and response. Diagram at the top of a workflow file (or tinyprobe: Show Workflow Diagram) shows it without running.
• A request is not a test. ▶ Call, and a call from the History, show the HTTP status of the response (401 Unauthorized in orange, 200 OK in green, 5xx and no response in red) instead of PASSED, with no "1/1 steps passed · 0 checks": just the request and the response.
• Request, then response. A step's detail shows what was sent before what came back, and the values it saved are folded under their names (Saved token, userId), opened with a click.
• The run view, the diagram, the editing of environment files and the request history are shared with tinyprobe-app, the tinyprobe app for the browser: the same code in both.
• License: GPL-3.0. From this version the extension is licensed under the GNU General Public License, version 3 (GPL-3.0-only): it stays free and open source, and a changed version can be shared only under the same license, with its source code. Releases up to 0.9.0 stay under the MIT License.

Works with tinyprobe 0.10.0 or newer, as before; Run to here needs 0.18.0.

What's Changed

• Share the run view with tinyprobe-app by @kikutano in #11
• Share the environment editing and the history with tinyprobe-app by @kikutano in #12
• Bring the shared run view up to tinyprobe-app 0.4.3; a call is not a test by @kikutano in #13
• The diagram of a workflow in the run view by @kikutano in #14

Full Changelog: v0.9.0...v0.10.0

v0.9.0

• Go to definition on the file of a schema check (tinyprobe 0.20): F12 or Ctrl+click on $: { schema: ../openapi.yaml#/components/schemas/User } opens openapi.yaml.
• Completion and hover of the new operators matches and schema come from tinyprobe schema: with tinyprobe 0.20.0 they appear with no update needed.

What's Changed

• Go to definition on the file of a schema check by @kikutano in #10

Full Changelog: v0.8.0...v0.9.0

v0.8.0

Sequence files, go to definition, and the items of for_each steps.

Sequences

• Files with sequence: (tinyprobe 0.18) are validated on save and get completion and hover. They have no Run link and no test of their own, since they run through the use steps of a workflow.
• In the with of a use, completion offers the sequence's params, required first. Hover on a use shows the sequence's name, description and params.

Go to definition (F12, Ctrl+click)

• On the endpoint of a call: its line in the API file, named imports included (billing.login).
• On a path: the sequence of a use, a file of import, the file of a for_each or of an upload.

for_each items

• The run view shows a row for each item of a for_each step, with its status and duration, and its own request, response and checks when selected. The step shows how many items passed, and a failed one opens on its first failed item.
• In the Test Explorer the items become children of their step once it has run, each passed or failed; the step's message lists every failed item.

Works with tinyprobe 0.10.0 or newer, like 0.7.0; sequences and for_each need 0.18.0.

What's Changed

• Sequence files: validation, completion of use, go to definition (M4-13) by @kikutano in #8
• for_each items in the run view and the Test Explorer (M4-14) by @kikutano in #9

Full Changelog: v0.7.0...v0.8.0

v0.7.0

A tinyprobe sidebar, with your request history and your environment.

History

• The last 50 calls made with ▶ Call, newest first, with their status, duration and time.
• Click one to see its request and response again in the run view.
• Send Again repeats it with the same arguments and the environment selected now; Open Endpoint goes to it in its API file.
• The history is kept for each workspace in VS Code's storage, not in your folder, with bodies cut at 64 KiB. Clear History empties it.

Environment

• The variables of the selected environment file, then those of the tinyprobe.variables setting, which win over the file.
• Edit a value, add a variable or delete one: only that variable's lines change, comments and order are kept, the file is saved, and Undo works. Values are typed like YAML, as in ▶ Call.
• Values whose name looks like a secret (password, apiKey, clientSecret, token...) are hidden until you click the eye.

Works with tinyprobe 0.10.0 or newer, like 0.6.0; Run to here still needs 0.18.0.

What's Changed

• History and Environment views (M4-7, M4-8) by @kikutano in #7

Full Changelog: v0.6.0...v0.7.0

v0.6.0

Run to here.

• A Run to here link on every step of a workflow runs the workflow up to that step and opens the run view on it, with its request and response. The step that needs the token and the ids of the earlier steps can be checked without running the rest.
• The steps before it run again every time, so nothing is kept between runs.
• For a use of a sequence, the link runs the whole sequence.

Completion and hover know the keys added by tinyprobe 0.17 and 0.18: form, multipart, cookies, basic, oauth2, for_each, use, sequence and named imports. The run view shows the Cookie and Authorization headers as they were sent.

Run to here needs tinyprobe 0.18.0 (tinyprobe update); with an older binary the link is not shown and everything else works as before.

What's Changed

• Run to here: run a workflow up to a step and show it by @kikutano in #6

Full Changelog: v0.5.0...v0.6.0

v0.5.0

Completion and hover in tinyprobe files.

• The keys you can write where the cursor is (a step, its expect, an endpoint, the conditions of a check), with their description, minus the ones already there.
• After call:, the endpoints of the workflow and of its imports, with method and path, also while the workflow is not valid yet.
• In with:, the arguments of the called endpoint: required params first, then the others, query parameters and the placeholders of the path.
• Hover on a key shows what it does; hover on the endpoint of a call shows its method, path, params and where it is defined.

Needs tinyprobe 0.14.0 (tinyprobe update); with an older binary everything else works as before.

What's Changed

• Add completion and hover in tinyprobe files by @kikutano in #5

Full Changelog: v0.4.0...v0.5.0

v0.4.0

The chat gets the tinyprobe tools. The extension registers the tinyprobe MCP server with VS Code, so Copilot and the other chat agents of the editor can call guide, inspect, call and run with nothing to configure. VS Code asks you before an agent sends requests with call or run. Needs tinyprobe 0.13.0 (tinyprobe update) and VS Code 1.101 or newer; with older versions everything else works as before.

tinyprobe: Generate Tests with AI. Opens the chat in agent mode with the prompt that asks the agent to read your backend's routes, authentication and seed data, then write and run the tinyprobe tests. It uses the model you already have, and checks its work with the tinyprobe tools. The prompt is also copied to the clipboard, for agents outside the editor. "Migrate with AI" offers the same as its first choice, "The backend's code".

What's Changed

• Give VS Code's chat agents the tinyprobe tools by @kikutano in #3
• Add tinyprobe: Generate Tests with AI by @kikutano in #4

Full Changelog: v0.3.0...v0.4.0