testing – Page 19 – Türker Yıldırım

Asserting an event was dispatched, not that a mail was sent

NoteMay 31, 20221 min read

Testing the effect couples the test to the listener, and testing the event tests the thing the code under test is actually responsible for. The listener gets its…

A load test with an assertion is a test; without one it is a benchmark

NoteMay 26, 20222 min read

A load test that reports numbers requires somebody to interpret them; one that asserts a threshold fails a build. The ramp shape matters as much as the target:…

A feature flag needs an expiry date at creation

NoteMay 19, 20222 min read

Creating a flag is a deliberate act and removing one is not, so without a mechanism the count only goes up. A failing test is the only mechanism…

Style variations, and the theme that ships four designs

Deep diveMay 18, 202210 min read

6.0 in May, and styles/ as a directory of theme.json fragments. Four client sites that differ only in colour and type.

An SLO that survived contact with the business

Deep diveMay 17, 202214 min read

99.9% was written in a document in 2020 and had never been measured. Three teams with three definitions of “up”.

Sharded pub/sub, and the cluster that could not broadcast

Deep diveApril 25, 20229 min read

A broadcast that goes to every node on every publish. Classic pub/sub does not scale in a cluster, and 7.0 gives channels that respect slots.

Playwright traces uploaded only when a test fails

NoteApril 20, 20221 min read

A trace is large and is only interesting when something went wrong, so recording on retry and uploading on failure is the arrangement that stays affordable. if: failure()…

Model::shouldBeStrict in development, and what it catches

NoteApril 13, 20222 min read

One call turns on three separate checks that each convert a silent misbehaviour into an exception. The third is the one with the largest immediate yield, because a…

Redis functions replace the Lua scripts nobody could version

Deep diveApril 12, 202210 min read

7.0 in April, and a script registry that survives a restart. EVAL with a SHA nobody can map back to source.

A test that asserts on a duration is a flaky test

NoteApril 6, 20221 min read

A CI runner is shared, throttled and occasionally slow, so any assertion about wall-clock time fails eventually for reasons unrelated to the code. Query counts and memory are…