Skip to content

ci: streamline Rocky dependencies and improve build failure diagnostics - #7

Open
3for wants to merge 55 commits into
developfrom
fix/ci-gradle-only
Open

3for wants to merge 55 commits into
developfrom
fix/ci-gradle-only

Conversation

@3for

@3for 3for commented Sep 10, 2026

Copy link
Copy Markdown
Owner

What does this PR do?

  • Reduce Rocky Linux build dependencies, use C.utf8, and configure JAVA_HOME explicitly.
  • Increase DNF download concurrency and disable weak dependency installation.
  • Capture Gradle console output while preserving build failure exit codes.
  • Upload **/logs/tron-test.log and ci-logs/*.log only on failure, with 7-day retention.
  • Disable automatic test retries while retaining the configuration as comments.
  • Remove the deleted multinode workflow from PR cancellation handling.

Why are these changes required?

Unnecessary packages increase CI setup time. Missing console logs make test failures harder to diagnose, while automatic retries can hide intermittent failures. Workflow cancellation should also reflect the workflows that still exist.

This PR has been tested by:

  • Unit Tests
  • Manual Testing

xxo1shine and others added 5 commits September 11, 2026 14:15
…#6930)

Node and network API responses populated two fields from the wrong source values because of copy-and-paste mapping errors.

Map needSyncFromPeer from the corresponding peer state and assign UDP inbound traffic to the udpInTraffic protobuf field.
…me (tronprotocol#6950)

* chore(deps): upgrade grpc-java from 1.83.0 to 1.83.1

1. bump grpcVersion to 1.83.1 to pick up the upstream fix for
   grpc/grpc-java#12930 (PR grpc/grpc-java#12942), which enforces
   connection.remote().maxActiveStreams(maxStreams) at handler startup
2. drop GrpcNettyMaxConcurrentStreamsLimiter, the local protocol-negotiator
   shim that applied the same limit while 1.83.0 left the remote endpoint
   unbounded until the client acknowledged SETTINGS

* chore(deps): upgrade jackson from 2.18.6 to 2.18.10

bump jackson-databind from 2.18.6 to 2.18.10 to pick up cumulative fixes from the 2.18.x line

* chore(deps): upgrade logback to 1.3.16 and slf4j to 2.0.17

1. bump logback-classic from 1.2.13 to 1.3.16 and slf4j-api,
   jcl-over-slf4j, jul-to-slf4j from 1.7.36 to 2.0.17; logback 1.3
   requires the slf4j 2.0 provider model, and 1.3.16 is the last 1.3.x
   release and the ceiling for the x86_64 JDK 8 build, since 1.5.x
   requires JDK 11
2. rename DelayingShutdownHook to DefaultShutdownHook in the toolkit
   logback.xml; logback 1.3 removed the old class and only auto-maps
   the legacy name with a startup warning
3. drop the CONSOLE appender from the toolkit logback.xml; no logger
   ever referenced it, so it never emitted output on 1.2 either, and
   logback 1.3 now flags it with an unreferenced-appender warning
4. accept one known 1.3.x behavior change: SizeAndTimeBasedRollingPolicy
   now throttles its maxFileSize comparison to once per 60s
   (SimpleInvocationGate) instead of the adaptive ~100-800ms gate of
   1.2.13, so under sustained heavy logging a file can overshoot the
   500MB cap by up to 60s of writes before the %i rollover fires;
   time-based rollover and totalSizeCap/maxHistory cleanup are ungated
   and unaffected
5. note for operators running a custom --log-config file: well-formed
   1.2-era configs using standard elements keep working unchanged
   (jmxConfigurator degrades to an ignored-property warning, the legacy
   shutdown hook name is auto-mapped), and malformed XML still fails
   fast via TronError(LOG_LOAD) exactly as on 1.2; however, a config
   that references an uninstantiable class (e.g. a custom appender
   missing from the classpath) now aborts the whole appender-ref phase
   instead of losing just that one appender, so the node starts with no
   log output while the ERROR statuses are printed to stdout by
   LogService

* chore(deps): upgrade commons-lang3/collections4 and drop commons-math

1. bump commons-lang3 from 3.4 to 3.20.0; the runtime classpath already
   resolved 3.18.0 through libp2p 2.2.9's transitive requirement, so
   align the declaration with what actually ships and move past the
   CVE-2025-48924 range that the nominal 3.4 still sits in
2. bump commons-collections4 from 4.1 to 4.6.0
3. remove commons-math 2.2; no source file imports
   org.apache.commons.math and nothing else in the dependency graph
   requests it

* chore(deps): remove joda-time and use JDK time APIs

1. drop the joda-time 2.3 dependency.
2. replace the six new DateTime(millis) log-formatting call sites in
   DynamicPropertiesStore, DposTask and DposService with a new
   Time.getIsoTimeString helper backed by java.time; its formatter
   (yyyy-MM-dd'T'HH:mm:ss.SSSXXX in the system zone) reproduces joda's
   DateTime.toString() output byte for byte where the JDK and joda 2.3
   time-zone databases agree (UTC nodes are unaffected); zones whose
   rules changed after joda's 2013-era tzdb, e.g. Europe/Moscow, now
   render the corrected offset for the same instant.
3. replace DateTime.now() day arithmetic in four test classes with the
   java.time equivalent, ZonedDateTime.now().minusDays(n)/plusDays(n)
   .toInstant().toEpochMilli(), keeping joda's calendar semantics
   one-to-one, and map plain DateTime.now().getMillis() to
   System.currentTimeMillis()
* feat(api): sanitize HTTP API error responses

Standard HTTP error paths used to expose internal details to clients:
Util.processError prefixed every message with the Java exception class
name, several servlets printed raw Throwable.getMessage() directly, and
the two solidity query endpoints returned bare-text error bodies.

Centralize the client-facing text decision in Util.processError:

* keep the raw non-blank message only for the exact runtime types
  JsonFormat.ParseException, ContractValidateException and
  MaintenanceUnavailableException; a null, empty or whitespace-only
  message falls back to "internal server error"
* preserve the events-deprecation message only for the exact
  IllegalArgumentException type carrying EVENTS_DEPRECATED_MSG
* write the fixed rate-limit and INVALID address messages, along with
  existing GetBlock validation messages, through the package-private
  writeAuditedError helper; these audited callers bypass exception
  classification, and printErrorMsg is private to the shared writer
* return {"Error":"internal server error"} for every other exception,
  with no exception class name

Client-visible changes:

* all processError-based error bodies lose the "class <FQCN> : "
  prefix; unclassified raw messages become "internal server error"
* the rate-limit rejection body becomes
  {"Error":"lack of computing resources"} on every endpoint extending
  RateLimiterServlet, including full-node, solidity and PBFT /jsonrpc
* gettransactionbyid / gettransactioninfobyid on solidity return
  standard {"Error":...} JSON instead of bare text
* validateaddress, getBrokerage and getReward replace leaked library
  messages in their failure branches with existing fixed texts; the
  "INVALID address" body is now written via writeAuditedError and loses
  the space after the colon
* getblock keeps its exact error bodies (refactor only)

Cover Solidity transaction and transaction-info GET/POST input errors,
backend failures, successful lookups and missing records directly with
mocked Wallet calls and in-memory requests and responses. Replace the
transaction servlet tests that accidentally exercised POST in both cases,
changed global stdout and used a shared temporary response file.

Verify both endpoint and global rate-limit rejections across the three
JSON-RPC servlet variants, including status, response body and the absence
of business dispatch on rejection.

HTTP status codes, success responses, request validation rules and
gRPC behavior are unchanged. JSON-RPC behavior is unchanged except for
the shared HTTP rate-limit response described above.

Closes tronprotocol#6936

* fix(api): keep server-side failure logging at error level

The previous commit routed four catch-all blocks through the shared
processError entry point, which logs at debug. Those four catches cover
server-side work only: getburntrx, getnodeinfo and getpendingsize read
no request parameters, and in getreward malformed addresses are already
handled by the preceding DecoderException | IllegalArgumentException
catch. Their failures therefore left no trace under the default log
configuration, where the API topic is INFO.

Add a dedicated processServerError entry point that logs at error and
then applies the same sanitization, and use it at those four call sites.
Logging the exception once inside the helper keeps a single record at
any log level, instead of pairing an error log in the servlet with the
debug log in the shared path.

The shared Exception entry point keeps debug on purpose: its callers
also cover request parsing, so an unauthenticated client can fail it
cheaply and repeatedly, and an unconditional stack trace per request
would amplify that into log pressure. Distinguishing client from server
faults on that path is the parameter/internal split tracked as follow-up
in tronprotocol#6936.

Client-facing responses are unchanged.
@3for
3for force-pushed the fix/ci-gradle-only branch from 04905f4 to 1a1ea8c Compare September 18, 2026 09:02
@3for
3for force-pushed the fix/ci-gradle-only branch from 43b86de to 92f2588 Compare September 23, 2026 03:51
halibobo1205 and others added 18 commits September 24, 2026 16:48
)

* feat(vm): add optimized storage layout and alias write checks

* feat(api): separate outbound JSON parsing

* perf(vm): cache alias check at Storage construction

---------

Co-authored-by: halibobo1205 <halibobo1205@gmail.com>
tronprotocol#6982)

* refactor(api): replace response copy wrappers with Jetty content count

* fix(api): preserve legacy Unicode replacement in JsonFormat

Use JDK UTF-8 conversion to replace malformed surrogates in parser error descriptions and isolated low surrogates in string output. Match the legacy writer's question-mark replacement while preserving valid Unicode and parsing behavior.

Cover identifier, ABI escape, numeric errors, and low-surrogate output with regression tests, including strict UTF-8 checks through native Jetty responses.
- Remove Development Tools and unnecessary packages
- Install only JDK 8, git-core and zstd with weak dependencies disabled
- Use C.utf8 to avoid installing glibc-langpack-en
- Resolve and validate JAVA_HOME so Gradle no longer requires which
- Collect tron-test.log, rotated logs and JUnit XML across five PR build jobs
- Run artifact uploads even when preceding steps fail
- Use distinct artifact names per job and matrix configuration
- Retain artifacts for 7 days and warn when no files are found
Remove the stale integration-test-multinode.yml reference to avoid
unnecessary API requests when cancelling workflows for closed PRs.
- Save build, RocksDB test and coverage output with tee and plain console mode
- Use Bash pipefail to preserve failures when capturing stdout and stderr
- Include console logs and HTML test reports in diagnostic artifacts
- Preserve existing test retry and base coverage failure policies
Remove the test-retry plugin and retry configuration shared by test and
testWithRocksDb so test failures fail the task without retrying.
- Limit diagnostic artifacts to **/logs/tron-test.log and ci-logs/*.log
- Upload logs only when a preceding step fails
- Include base test failures tolerated by continue-on-error
- Keep the existing 7-day retention period
- Set DNF max_parallel_downloads to 10 for Rocky dependency installation
- Restore the test-retry plugin and configuration as comments, keeping retries disabled
Set DNF minrate=256k and timeout=30 to abort persistently slow
connections and allow fallback to another mirror.
Remove --no-build-cache from Debian and baseline coverage jobs
to preserve the cache behavior introduced in tronprotocol#6938.
- Upload diagnostics on failure or cancellation with a 4-minute timeout.
- Add step timeouts to reserve time for diagnostic uploads.
- Keep Gradle daemon shutdown failures non-fatal.
- Retain job-level timeouts for base tests to avoid accepting a partial coverage baseline after a tolerated step timeout.
- Preserve outcome checks for tolerated base test failures.
@3for
3for force-pushed the fix/ci-gradle-only branch from 92f2588 to 430e431 Compare October 10, 2026 12:19
Federico2014 and others added 28 commits October 11, 2026 21:08
* fix(jsonrpc): handle null parameters consistently

Validate nullable JSON-RPC arguments before they are dereferenced.
Return stable protocol errors instead of jsonrpc4j's unhandled -32001
fallback, which exposes Java exception types.

Validate eth_call's required transaction argument before its block
selector. The double-null case now returns -32602 instead of -32600.

Align eth_uninstallFilter lookup misses with geth: null, unknown and
already removed IDs return false. Empty and non-hexadecimal strings
also return false when no filter matches, without new format checks.
Non-null lookup misses now return false instead of -32000.
Use the value returned by remove to report whether a filter was removed
and avoid the race between checking and removing the same entry.

Remove the obsolete uninstall ItemNotFoundException declaration and
error mapping. Java binaries remain compatible, but source callers
catching that checked exception, or implementations still declaring it,
may need adjustment when recompiled.
Keep getFilterChanges and getFilterLogs on their existing lookup-miss
errors, and preserve request-source precedence and filter isolation.

Treat null block-detail flags as false and retain DTO defaults for
explicitly null optional fields. Preserve filter wildcard semantics.

Add binding, service, wire and VM execution regressions for null
arguments, error contracts, filter removal, source isolation and
concurrent uninstall results.

* fix(jsonrpc): reject null required positional parameters

Return invalid params for an explicit null filter ID in
eth_uninstallFilter, eth_getFilterChanges and eth_getFilterLogs, and for
a null fullTransactionObjects flag in eth_getBlockByHash and
eth_getBlockByNumber. The block hash or selector is still validated
first, and the flag is checked before any block lookup, so the result
does not depend on whether the block exists. Unknown non-null filter
IDs keep their existing results.

This replaces the earlier null-as-false handling of the flag and the
filter ID. go-ethereum v1.17.6 rejects null for these required
arguments; earlier releases decoded it as a zero value.

* refactor(jsonrpc): share the null check for required parameters

Route the null checks for the ten required JSON-RPC parameter
positions through one requireParam helper instead of repeating the
same if/throw block at each site. The "invalid params" and "invalid
filter request" messages are unchanged and are now both private
constants in TronJsonRpcImpl.

Remove the null check from the LogFilter constructor. In production
code, LogFilter is only built through LogFilterWrapper from
eth_getLogs and eth_newFilter, and both reject a null filter request
before reaching it, so the constructor check and its unit test are
dropped. Null filter requests remain covered at the service and wire
level.

Extract getBlockByHash(byte[]) and use it from ethGetBlockByHash and
getBlockByJsonHash, so ethGetBlockByHash no longer inlines the block
lookup. The validation order is unchanged.

No behavior change.

* test(jsonrpc): use the three-argument TronJsonRpcImpl constructor

Adapt the tests added for null parameter handling to the constructor
that now takes the Manager, and drop the removed setManager call.
Tests built only from mocks pass a mocked Manager, and tests backed by
the shared database pass dbManager, matching the existing tests.
…col#6985)

Replace jsonrpc4j's unhandled -32001 fallback with a fixed -32603
response so unmapped exception types and raw messages no longer reach
clients. Keep the documented -32001 contract for net_version and
eth_chainId through explicit mappings with fixed message and data.

Propagate VirtualMachineError, ThreadDeath, LinkageError, and TronError
through the JSON-RPC boundary. Scan complete cause chains without
allocating a visited set, detecting cycles without missing fatal causes.
Classify Throwables escaping dispatch before logging or mapping them.

Before rethrowing an Error, make a best-effort attempt to commit an
empty HTTP 500 so the container does not render its default error page.
Preserve the original error if cleanup fails, and leave an already
committed response untouched.

Route single requests through handleRequest because the servlet handle
API catches Throwable. Catch RuntimeException, IOException, and Error at
both dispatch boundaries, then propagate a classified fatal cause or
return -32603 when a response is appropriate. This also covers non-fatal
Errors from interceptor hooks. Discard partial dispatch output on
failure.

Recover each failed batch element without discarding earlier results or
skipping later requests. Share internal-error construction across
serialization, dispatch, and response-parsing failures. Count only the
replacement when parsing fails, retaining existing overflow rules and
the current single/batch handling of requests without an id.

Log unmapped exceptions, chain-identity failures, and asynchronous
log-query failures and interruptions at DEBUG with the Throwable,
classifying fatal causes before logging the failures. Log batch
request-serialization and response-parsing recoveries at DEBUG with the
element index and the Throwable. Log the first escaped dispatch failure
in each batch at ERROR with the Throwable; repeats use DEBUG with only
the index and exception type. Restore interrupted status for
asynchronous log queries.

Reject Boolean, object, and array request IDs before dispatch with
-32600 and id:null. Reject non-null scalar params with -32600, preserve
valid IDs, and isolate invalid batch elements so their siblings run.
Missing, null, array, and object params keep their existing semantics.

Remove the obsolete request replay wrapper and HTTP status provider.
Add resolver, servlet, embedded-Jetty, chain-identity, asynchronous
failure, dispatch-contract, and request-envelope regression coverage.
Exercise the real HTTP interceptor chain and observe commitment without
performing it on the servlet's behalf.

Test interceptor Errors with a real server and preserve the existing
notification response-suppression rules with characterization tests.

Add a package-private server injection seam for servlet tests without
changing production initialization.
The node.rpc.maxRstStream and secondsPerWindow options defaulted to
0, which left the HTTP/2 RST_STREAM frame limit unset, and the limit
was configured only when both values were positive. Give them
sensible defaults, validate them at load time, and apply the limit
unconditionally so the configuration behaves consistently.

- NodeConfig: default maxRstStream/secondsPerWindow to 1000/5.
  In postProcess, reject negative values for either option and reject
  Integer.MAX_VALUE specifically for maxRstStream, which grpc-java
  treats as disabling the limit. Independently replace each zero with
  its default and log a warning.
- RpcService: always configure maxRstFramesPerWindow from the config
  values; validation stays solely in NodeConfig, consistent with the
  neighboring maxConcurrentCallsPerConnection option.
- reference.conf, docs/configuration.md and config/README.md:
  document the new defaults and the 0 fallback for compatibility.
- Tests: pin the Java and reference defaults, preserve positive
  overrides, and verify PING at the RST limit followed by GOAWAY
  after one additional reset using a prebuilt HTTP/2 frame burst.

Note: an explicit 0 now falls back to the default instead of leaving
the limit unset, and maxRstStream = 2147483647 is rejected at startup.
Remove verification entries for the disabled test-retry plugin
and its plugin marker artifact.
- Wrap PR build, test, coverage and Toolkit stages with a shared runner
- Collect JVM crash logs, daemon logs, memory samples and OOM counters
- Record test classes and methods associated with each test executor
- Show failure evidence in Actions annotations and job summaries
- Upload stage-specific diagnostics and available JUnit XML on failure
- Preserve command exit codes when diagnostic logging fails
- Use PR diagnostic scripts for base coverage builds and retain logs
  for tolerated failures

(cherry picked from commit 2f61ae1)
This reverts commit 430e431.

Explicitly exclude cancelled runs from all diagnostic upload conditions, while retaining uploads for ordinary failures and tolerated base test failures. Preserve the later Gradle and JVM diagnostic runner integration.
Merge upstream a901f85. Preserve disabled test-retry and cancellation upload guards while accepting the upstream Sonar plugin removal.
1. bump jsonrpc4j to 1.7, the latest release
2. add jakarta.servlet-api 5.0.0 as a test dependency; jsonrpc4j
   declares it optional, and Mockito needs it on the classpath to mock
   JsonRpcServer, whose new handle overload takes jakarta servlet
   types; the servlet dispatches through handleRequest and never
   calls that overload
3. drop javax.jws-api; jsonrpc4j 1.7 logs a missing
   javax.jws.WebParam at debug level instead of error, and no code
   uses @WebParam
4. accept one wire-level behavior change: when a positional parameter
   cannot be converted to the declared Java type (for example an
   object or array where a string is expected), the response changes
   from -32700 "JSON parse error" with id "null" to -32602 "Failed
   to read method parameter at index N" with the request id preserved
   and no data field; this also applies to batch entries, and a
   request without an id still gets this error with id null; the
   error is produced inside jsonrpc4j and does not pass through the
   error resolver
5. transitive changes on the runtime classpath: commons-codec 1.15 ->
   1.16.1, httpcore 4.4.13 -> 4.4.16, httpcore-nio 4.4.5 -> 4.4.16,
   jakarta.jws-api 3.0.0 added, net.iharder:base64 removed; slf4j-api
   and jackson keep resolving to 2.0.17 and 2.18.10
6. update the dispatch contract test, extend the error integration
   test with conversion errors in single, batch and id-less requests,
   and update the dependency verification metadata accordingly
Preserve the coverage-base diagnostic script checkout alongside upstream Gradle wrapper validation. Retain failure diagnostics, cancellation upload exclusions and disabled test retries.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.