• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

floatdrop / moq-go / 31673100289
89%

Build:
DEFAULT BRANCH: draft-19
Ran 13 Aug 2026 06:19AM UTC
Jobs 1
Files 100
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

13 Aug 2026 06:14AM UTC coverage: 87.187% (+0.1%) from 87.05%
31673100289

push

github

web-flow
test(session): control-stream violations, verifier sentinels, fetch cancel (#71)

Three gaps chosen for what they protect rather than what they add to the
number, which moves only 86.7% -> 87.1%. The session package's remaining
uncovered mass is diffuse, so picking by size would have bought noise.

dispatchControl was the one worth doing. Table 5 in §10 leaves GOAWAY as the
only message valid on the control stream after SETUP and §3.5 gives
PROTOCOL_VIOLATION for anything else, but only the GOAWAY branch had a test.
That matters more than a normal gap because the dispatcher recently lost a
mechanism for per-rule close codes that nothing ever constructed, and
removing it rested on the claim that every violation here is a
PROTOCOL_VIOLATION — a claim with nothing asserting it. Both the code and the
reason are now checked.

TokenDeniedError.Unwrap is the documented survival path for a third-party
verifier's own error across our normalisation, and nothing called it. A
policy that cannot tell "expired" from "forbidden" once the error crosses
this boundary is a real loss for anyone implementing TokenVerifier, and it
breaks silently.

IncomingFetchStream.Cancel was at 0% across every package. Its comment
records what it does NOT guard: resetting the wrong half does not compile,
since ReceiveStream has no CancelWrite. It guards that Cancel reaches the
peer at all, so the assertion is that the writer sees the stream fail.

All three verified by mutation. The Cancel test needed a second pass — its
first form turned a stubbed-out Cancel into a 60s package timeout rather than
an assertion, because the write blocks instead of erroring; it now fails in
5s.

Not tested, deliberately: WriteMessageAfterSetup and WithMaxFilterRanges,
which are exercised through pkg/relay and read as 0% here only through
per-package attribution; the Error/Unwrap string methods; and the
isDataStream markers Go never calls.

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

8914 of 10224 relevant lines covered (87.19%)

1.99 hits per line

Jobs
ID Job ID Ran Files Coverage
1 31673100289.1 13 Aug 2026 06:19AM UTC 100
87.19
GitHub Action Run
Source Files on build 31673100289
  • Tree
  • List 100
  • Changed 6
  • Source Changed 0
  • Coverage Changed 6
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • Github Actions Build #31673100289
  • d45a92d3 on github
  • Prev Build on draft-19 (#31671683862)
  • Next Build on draft-19 (#31674203330)
  • Delete
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE TRIAL · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc