mizosoft / methanol

90%

master: 89%

LAST BUILD ON BRANCH bcs-touch-ups

branch: bcs-touch-ups

CHANGE BRANCH

x

Reset

• bcs-touch-ups
• 34-disable-decompression-for-HEADS
• 35-strip-content-encoding
• adapter-codec-touchups
• adapter-improvements
• allow-creating-websockets
• bugfix/build-on-aarch64
• cache-touch-ups
• cache-updates
• cache-updates-2
• cancellation
• crp-cwp-touchups
• dead-code
• decrease-scheduler-core-size
• dependency-updates
• doc-updates
• dont-releast-on-bg-revalidation
• fail-on-duplicate-directives
• fix-46
• fix-decoder-hanging
• fix-decompression-hang
• fix-method-copy
• fix-missed-response
• fix-snapshotting
• fix-timeout-memory-leak
• fix-windows-diskstore-http-caching
• headers-accumulator-spec
• http-caching
• invalid-expires
• investigate-ci-failure
• jackson-polyformat
• jaxb-jakarta
• jdk-api-updates
• kotlin-extension-updates
• kotlin-extensions
• local-properties
• master
• mockexecutor-tostring-race
• multipartbodypublisher-formpart-test
• name-fallbackexecutor-thread
• normalize-dates
• offer-to-add
• pass_emit_failures_downstream
• plain-native-tests
• proper-tostring
• properly-setup-coveralls
• quarkus-native-test
• redis
• redis-store-touch-ups
• refactor-workflows
• remove-join
• response-publisher
• responsebuilder-static-factory
• responsepayload-is
• retain-status-header
• retryer
• save-request-copy
• set-headers
• smarter-closure
• spi-touchups
• spring-boot-fix
• static-analysis
• structured-suffix
• test-subscriber-trylock
• test-timeout-and-interruption
• test-timeout-touch-ups
• unify-fallback-executor
• use-daemon-threads
• use-linkedhashmap
• v1.5.0
• v1.6.0
• v1.7.0
• v1.8.0
• v1.8.1
• v1.8.2
• v1.8.3
• wbp-memory-optimization

Build # #437

Build Type

Pull #100

github

Committed by mizosoft

Commit Message

ByteChannelBodySubscriber touch-ups

 - Amongst other things, a race issue that lead the channel to read nothing
   when interrupted by another thread instead of throwing is fixed. This was
   caused by the implementation relying on AbstractInterruptibleChannel::end
   to detect interruption even if detected by waiting on BlockingQueue::take.
   This was not reliable as when a thread is interrupted, it first unparks
   parked threads, then calls the channel's interruption hook. This makes
   it possible for the interruption to be detected by the queue, but not
   yet by the channel (AbstractInterruptibleChannel::end). Now the channel
   throws a ClosedByInterruptionException when an InterruptedException is
   thrown by BlockingQueue::take.

 - Another notable change is that an upstream error is logged and not recorded
   if it is found that there is no way for the reader to notice it due to channel
   closure.

Pull Request Pull Request #100: ByteChannelBodySubscriber touch-ups

Run Details

109 of 117 new or added lines in 3 files covered. (93.16%)

3 existing lines in 1 file now uncovered.

7338 of 8113 relevant lines covered (90.45%)

0.9 hits per line