supabase / supabase-swift / 35244197659
89%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 135
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 89.361% (+0.1%) from 89.227%
35244197659

push

github

web-flow 
test(realtime): cover _handleMessage over a real connection (#1362)

_handleMessage scored CRAP 30 — CC 5 at 0% coverage. It is private and
only reachable from the _task.receive() loop, and while WebSocketTests
already stands up a real loopback server, every test there drives
traffic client→server only. Nothing ever arrived for _handleMessage to
handle.

Teaching LoopbackWebSocketServer to push a frame is enough to reach it,
so this needs no production change at all. The extraction the issue
proposed turned out unnecessary: coverage, not complexity, was carrying
the whole score.

Two frames go out, not one. The second can only arrive if handling the
first re-armed the receive loop, so the test fails if _scheduleReceive()
stops being called on the success path — verified by mutation, where
dropping that call fails the test and leaves the rest of the suite
green.

The second test pins that nothing surfaces after close. It is
deliberately not claimed as coverage of the isClosed guard: two
mechanisms enforce that (the guard, and events having already
finished), and mutation shows removing either alone leaves the test
green. It pins the property, not the line.

_handleMessage goes to 81.2% coverage and CRAP 5.2; _scheduleReceive
and _trigger go from untested to 100%. The remaining 18.8% is the
@unknown default branch, which cannot be reached because an unknown
URLSessionWebSocketTask.Message case cannot be synthesized.

Fixes SDK-1872

Co-authored-by: Claude <noreply@anthropic.com>

10692 of 11965 relevant lines covered (89.36%)

897562.06 hits per line

Jobs
ID Job ID Ran Files Coverage
1 35244197659.1 135
89.36
 GitHub Action Run
Source Files on build 35244197659