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

Chaguar / orderer / 41386798000
85%

Build:
DEFAULT BRANCH: main
Ran 02 Nov 2026 04:22AM UTC
Jobs 1
Files 217
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

01 Nov 2026 09:18PM UTC coverage: 84.773% (-0.06%) from 84.836%
41386798000

push

github

web-flow
[blocksprovider] Fix BFT stop races and flaky BFT deliverer tests (#1379)

* blocksprovider: fix BFT stop races and the flaky deliverer tests (#1372)

Port the fixes from fabric-x-common PR #228, whose copy of this package
is unused and is being removed.

- BFTCensorshipMonitor no longer launches header receivers after Stop().
  Monitor() checks the stop flag at the top of every pass, and
  launchHeaderReceivers() checks it again under the mutex that Stop()
  takes, so a receiver can no longer outlive the monitor. The timer
  branch also skips censorship detection once stopped.
- BFTDeliverer.FetchBlocks re-checks stopFlag before setting
  blockReceiver. If Stop() ran after the DoneC check, it cancels the new
  stream instead of keeping a receiver that nothing will stop.
- The BFT deliverer test fake now ends each stream only when its own
  context is cancelled, as with gRPC. Before, every stream shared one
  fake whose CloseSend pushed an error onto the shared recvStepC, so an
  old stream that shut down late could fail the stream that replaced it.
  This caused the off-by-one failures in
  TestBFTDeliverer_CensorshipMonitorEvents.
- BFT deliverer tests now wait for the factory and dialer stubs to
  register a new monitor or connection, not only for the fake's call
  count, which is bumped before the stub runs. Otherwise a test could
  send on the previous monitor's error channel while holding the setup
  mutex, which blocks the stub and hangs the deliverer. It could also
  assert on monitorSet before the newest monitor was registered.
- Add TestBFTCensorshipMonitor_StopBeforeMonitor, and remove the racy
  Stop() from TestBFTCensorshipMonitor_NoSources.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Yoav Tock <TOCK@il.ibm.com>

* blocksprovider: address review of the BFT stop-race fixes

- Give each DeliverBlocks round its own fetch error channel and pass it
  to FetchBlocks, instead of a shared field that the next round
  re... (continued)

13 of 21 new or added lines in 2 files covered. (61.9%)

28 existing lines in 8 files now uncovered.

18545 of 21876 relevant lines covered (84.77%)

51734.99 hits per line

Uncovered Changes

Lines Coverage ∆ File
6
92.02
-2.54% common/deliverclient/blocksprovider/bft_deliverer.go
2
93.42
-1.62% common/deliverclient/blocksprovider/bft_censorship_monitor.go

Coverage Regressions

Lines Coverage ∆ File
10
83.06
-2.32% common/ledger/blkstorage/blockfile_mgr.go
4
88.24
-5.88% common/ledger/blkstorage/blocks_itr.go
4
84.39
-1.95% node/assembler/batch_fetcher.go
3
90.57
0.0% node/comm/deliver.go
2
93.42
-1.62% common/deliverclient/blocksprovider/bft_censorship_monitor.go
2
82.69
-3.85% common/ledger/blockledger/util.go
2
93.22
1.22% node/delivery/consensus_decision_deliver_client.go
1
92.02
-2.54% common/deliverclient/blocksprovider/bft_deliverer.go
Jobs
ID Job ID Ran Files Coverage
1 41386798000.1 02 Nov 2026 04:22AM UTC 217
84.77
GitHub Action Run
Source Files on build 41386798000
  • Tree
  • List 217
  • Changed 14
  • Source Changed 0
  • Coverage Changed 14
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • fb99f8e5 on github
  • Prev Build on main (#41271906187)
  • Next Build on main (#41576353900)
  • 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