Lets a caller pin `app.common.logging`'s level for the duration of a
test run via `--log-level <trace|debug|info|warn|error>`. The flag is
off by default, so when absent the runner doesn't touch logger state
and stdout looks exactly as before.
When passed, the runner calls `(l/setup! {:app level})` right before
dispatching to the test block, so production code exercised by tests
emits only at the requested level or higher.
pnpm run test:quiet -- --focus frontend-tests.logic.groups-test \
--log-level warn
Composes with `--focus`; the two flags are independent.
Caveats worth knowing: top-level log calls fired at namespace load
time run before the runner parses CLI options and therefore slip past
this flag; direct `println` / `js/console.log` calls bypass the
logging system entirely and are unaffected.
Introduces `pnpm run test:quiet` for non-interactive runs (CI, scripted
invocations, agent loops). It runs the same pipeline as `pnpm run test`
— `build:wasm`, then `build:test`, then `node target/tests/test.js` —
but buffers each build step's stdout and stderr and only replays them
when that step exits non-zero. Test-runner output streams through
unchanged, so failures and the summary are never hidden. Short progress
hints (`Building wasm...`, `Building test bundle...`, `Running tests...`)
are written to stderr, leaving stdout to carry only the test results
for clean capture and parsing.
Forwards arguments verbatim, so `pnpm run test:quiet -- --focus ...`
composes with the existing `--focus` flag. The default `pnpm run test`
script and its output are unchanged.
Also documents the new command in the developer guide and updates the
frontend testing memory to recommend it for agent runs.
Previously `pnpm run test` always ran the full frontend-tests suite,
which made tight iteration on a single namespace or var painful. The
runner now accepts `--focus <ns>` or `--focus <ns>/<var>` and executes
only the matching tests, preserving each namespace's `:once` and `:each`
fixtures so behavior matches a full-suite run.
pnpm run test -- --focus frontend-tests.logic.groups-test
pnpm run test -- --focus frontend-tests.logic.groups-test/some-test
Also updates the developer guide and the testing memory so the flag is
discoverable from both docs and agent context.
Memories use a system of progressive disclosure:
Starting from a root memory, memories reference other memories using explicit
references.
The new system of hierarchical memories replaces AGENTS.md files.
GitHub #9215
Co-authored-by: Michael Panchenko <michael.panchenko@oraios-ai.de>
Co-authored-by: Codex <codex@openai.com>
Add SERENA_UPDATE_VERSION env var (in devenv docker-compose.yml)
to dynamically update Serena on agentic devenv without requiring
an image rebuild.
Apply for update to v1.5.0 (also changing initial installation
in Dockerfile to this version).
* ✨ Add search bar to prototype interaction destination dropdown
On pages with many boards the destination dropdown becomes hard to
navigate. Add an optional `searchable?` flag to the shared select
component that renders a case-insensitive filter input at the top of
the dropdown, and opt it in for the interaction destination select.
- Filtering reuses the already-computed option list (no extra queries).
- Arrow-key navigation tracks the filtered list.
- Search clears automatically when the dropdown closes, so reopening
starts with the full list.
- New `labels.no-matches` i18n key renders when nothing matches.
Closes#8618
Signed-off-by: moorsecopers99 <patellscott18@gmail.com>
* ✨ Use trigger input as search field in destination dropdown
Per review on #9006, the searchable select now uses the visible trigger
input as the filter field itself rather than an extra sticky input
inside the dropdown. The trigger behaves like a filterable select:
typing filters the options without permitting free-text values.
Signed-off-by: moorsecopers99 <vadanamihai409@gmail.com>
* ♻️ Update css on old select component
---------
Signed-off-by: moorsecopers99 <patellscott18@gmail.com>
Signed-off-by: moorsecopers99 <vadanamihai409@gmail.com>
Co-authored-by: moorsecopers99 <patellscott18@gmail.com>
Co-authored-by: moorsecopers99 <vadanamihai409@gmail.com>
Co-authored-by: Alejandro Alonso <alejandro.alonso@kaleidos.net>
* ✨ Bound the size of plugin task responses
When using the integrated remote MCP server, bound response size.
All responses are passed to LLMs, which themselves impose bounds.
This is a measure to bound memory usage in the centrally provided
MCP server.
GitHub #9493
* ✨ Bound parallelism in ExportShapeTool
Use an integer semaphore to bound parallel requests to this
memory-intensive tool, thus bounding memory usage.
GitHub #9493
* ✨ Add (manual) integration test script for ExportShapeTool parallelism
Add dependency tsx to facilitate executions.
GitHub #9493
* ✨ Make number of parallel export requests configurable in ExportShapeTool
Use env var PENPOT_MCP_EXPORT_SHAPE_MAX_PARALLEL_REQUESTS to configure
the maximum number of requests in multi-user mode (default 0, no limit).
Add :uri and :scheme/:host keys to exceptions raised by
`validate-uri` for better error diagnostics. Also fix a bug
where (str url) was used instead of (str uri) in the
host-missing exception path.
Update the existing blocked-target test to verify the new :uri
key, and add three new tests covering scheme rejection, missing
host, and DNS failure error paths. All 27 tests pass with 60
assertions and 0 failures.
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
