Ingrid Solberg
Took over a suite nobody had run in a year.
Member · 241 reputation · here since 17 Feb 2026
Asked (13)
Linting only the staged feature files in a pre-commit hook — is --json stable enough?
Trying to keep the hook under a second on a repo with about four hundred feature files, so linting the whole project every commit is out. This is what I have: Two things I cannot…
asked by Ingrid Solberg241 ·VS Code wrote .vscode/mcp.json but the chat still has no sdods tools
sdods mcp install vscode -p demo-shop reported that it created .vscode/mcp.json and the file looks right to me. Reloaded the window twice. The chat has the built-in tools and…
asked by Ingrid Solberg241 ·docker run fails on an M-series Mac: no matching manifest for linux/arm64/v8
Wanted the server and web UI without installing anything, so I went for the image: I thought Docker Desktop emulated whatever was missing on Apple silicon. It seems not, or I…
asked by Ingrid Solberg241 ·Pool leased an account whose username is the literal ${TEST_MEMBER_EMAIL} and auth rejected it
Local runs only. CI has been green throughout, which is why this took a day. projects/mybotbox/data/staging/users.csv: The failure is not from SDODS. It is from the auth…
asked by Ingrid Solberg241 ·Where do I actually look at expected vs actual after a @visual failure on CI?
The nightly went red on a baseline scenario. I can see the scenario failed and I can read the pixel count in the log, which tells me nothing about whether somebody changed a…
asked by Ingrid Solberg241 ·OAuth popup during a recording: everything I did in the popup is missing from the spec
Recording a login that goes through our identity provider. The provider opens in a popup, you finish there, the popup closes and the app is signed in. The recorded spec has the…
asked by Ingrid Solberg241 ·Lint passes but the run fails on a missing step — why is undefined-step checking opt-in?
Clean lint, then the run dies immediately on a step nobody wrote. Turning the flag on finds it: If it can find it, why is it not on by default? Feels like the most valuable check…
asked by Ingrid Solberg241 ·Why one fixture set for ui, api and hybrid, and which of the fixtures are worker-scoped?
Reading steps/fixtures.ts in the project I inherited and trying to work out what I am extending. Two things I cannot find a straight answer to. Why is there one test rather than…
asked by Ingrid Solberg241 ·Getting the HTML report out of the Docker image we run the suite in
We run the suite inside ghcr.io/siri1410/sdods-server on a self-hosted agent. The run works and the exit code is right, then the container goes away and takes the report with it.…
asked by Ingrid Solberg241 ·sdods har record --interactive exits before the browser even opens
Trying to capture a flow by hand rather than through a scenario. Fine, added --name checkout. Codegen opened, I clicked through the flow, closed the window, and got this: Nothing…
asked by Ingrid Solberg241 ·Should each tester run sdods mcp locally, or should we host one server for the team?
Eight testers, one repo, one staging environment. Right now everybody has stdio in their editor and their own checkout. It works, but nobody can ask about a run somebody else…
asked by Ingrid Solberg241 ·Exactly which .env files get read, in which order, and how do I confirm it?
I have inherited a repo with .env, .env.local, .env.staging and .env.staging.local at the root and two more inside the project folder. Nobody remembers why. Before I delete any…
asked by Ingrid Solberg241 ·Quarantined scenarios still run in our release gate and still fail it
We quarantine flaky scenarios in the web UI while somebody fixes them. The dashboard shows them as quarantined, so I assumed that was the end of it. The release gate disagrees.…
asked by Ingrid Solberg241 ·
Answered (6)
How do you test a node-graph editor — blocks, edges and connection handles?
Half our product is a workflow builder: a canvas with about 190 block types, edges between them, and connection handles on each node. None of it is text or roles — the nodes…
asked by Nadia Belkacem8,611 ·heal_events, heal_locator_stats and sdods heal report — which do I use for what?
Building a small internal dashboard on top of the MCP server. There appear to be three ways to get at healing data and I cannot tell from the tool descriptions where the…
asked by Zeynep Arslan64 ·net::ERR_ABORTED on an authenticated route is our single biggest source of failures
Top error signature in a clean single-worker run, by a wide margin: 55 of the 192 failures I captured, plus a big pile of click timeouts that I am fairly sure are downstream of…
asked by Elsa Nyman333 ·335 of 390 scenarios failed at -w 4 with the user pool exhausted, and doctor is all green
Full regression at four workers. 55 passed, 335 failed, and nearly all of the failures are this: The passes are all scenarios that need no user, or need @user:admin, which…
asked by Chandra Pillai873 ·probeTimeoutMs of 1500 is too short on our staging box — every candidate scores 0
Locally everything heals. On staging the same locators fail with HEAL_FAILED and every candidate in the hint reads 0.00, including ones I know match, because I can open the page…
asked by Rasha Halabi572 ·One viewer account, twelve read-only scenarios, and the suite behaves as if -w 1
A partner gives us exactly one viewer account and will not create more — it is their system, their policy, and no amount of asking has moved it. Twelve scenarios carry…
asked by Rasha Halabi572 ·