cpp/celeborn/client/reader/CelebornInputStream.cpp (9 lines): - line 77: // TODO: in java this is UNSAFE and PLATFORM related. hand-crafted here, - line 86: // TODO: compression is not supported yet! - line 122: // TODO: support the failure retrying, including excluding the failed - line 134: // TODO: in java this is UNSAFE and PLATFORM related. hand-crafted here, - line 163: // TODO: support retrying when createReader failed. Maybe switch to peer - line 173: // TODO: support localPartitionReader... - line 184: // TODO: support DfsPartitionReader... - line 197: // TODO: support skipLocation functionality... - line 198: // TODO: the currLocationIndex_ management is a mess. might be worker/src/main/scala/org/apache/celeborn/service/deploy/worker/PushDataHandler.scala (4 lines): - line 310: // TODO Currently if the worker is in soft split status, given the guess that the client - line 387: // TODO Currently if the worker is in soft split status, given the guess that the client - line 1292: // TODO replica - line 1515: // TODO just info log for ended attempt worker/src/main/scala/org/apache/celeborn/service/deploy/worker/storage/PartitionMetaHandler.scala (3 lines): - line 152: // TODO: When region is finished, flush the data to be ready for the reading, in scenarios that - line 206: // TODO: force flush the index file channel in scenarios which the upstream task writes and - line 265: // TODO: force flush the index file channel in scenarios which the upstream task writes and common/src/main/java/org/apache/celeborn/common/network/sasl/registration/RegistrationRpcHandler.java (2 lines): - line 62: * TODO: This should be made configurable. For now, we only support ANONYMOUS for client-auth and - line 151: // TODO: not validating the auth init request. Should we? common/src/main/java/org/apache/celeborn/common/network/sasl/registration/RegistrationClientBootstrap.java (2 lines): - line 68: * TODO: This should be made configurable. For now, we only support ANONYMOUS for client-auth and - line 157: // TODO: client validates required/supported mechanism is present cpp/celeborn/network/Message.h (2 lines): - line 100: // TODO: add decode method when required - line 129: // TODO: add decode method when required client-flink/flink-2.0/src/main/java/org/apache/celeborn/plugin/flink/tiered/CelebornTierMasterAgent.java (2 lines): - line 209: // TODO we should impl this when we support JM failover. - line 215: // TODO we should impl this when we support JM failover. client-flink/flink-1.20/src/main/java/org/apache/celeborn/plugin/flink/tiered/CelebornTierMasterAgent.java (2 lines): - line 209: // TODO we should impl this when we support JM failover. - line 215: // TODO we should impl this when we support JM failover. client/src/main/scala/org/apache/celeborn/client/commit/CommitHandler.scala (2 lines): - line 216: // TODO: if isSegmentGranularityVisible is set to true, it is necessary to handle the pending - line 517: // TODO: if support upstream task write and downstream task read simultaneously, common/src/main/scala/org/apache/celeborn/common/protocol/message/ControlMessages.scala (2 lines): - line 571: // TODO change message type to GeneratedMessageV3 - line 1043: // TODO change return type to GeneratedMessageV3 cpp/celeborn/client/reader/WorkerPartitionReader.cpp (2 lines): - line 58: // TODO: it might not be safe to call blocking & might failing command - line 86: // TODO: add metric or time tracing cpp/celeborn/memory/ByteBuffer.h (2 lines): - line 112: // TODO: this interface is called rapidly. maybe need optimization. - line 141: // TODO: currently the appender is not allowed to grow. common/src/main/scala/org/apache/celeborn/common/CelebornConf.scala (1 line): - line 1357: // TODO related to `WORKER_DIRECT_MEMORY_RATIO_PAUSE_RECEIVE`, client-flink/common/src/main/java/org/apache/celeborn/plugin/flink/ShuffleResourceTracker.java (1 line): - line 102: // TODO if we support partition replica for map partition we need refactor this client-flink/flink-1.20/src/main/java/org/apache/celeborn/plugin/flink/tiered/CelebornTierConsumerAgent.java (1 line): - line 185: // TODO: if fail to open reader, may the downstream task start before than upstream task, service/src/main/scala/org/apache/celeborn/server/common/http/authentication/SpnegoAuthenticationHandler.scala (1 line): - line 66: // TODO: support to config kerberos.name.rules and kerberos.rule.mechanism client/src/main/java/org/apache/celeborn/client/read/LocalPartitionReader.java (1 line): - line 285: // TODO implement similar to {@link WorkerPartitionReader} master/src/main/scala/org/apache/celeborn/service/deploy/master/tags/TagsManager.scala (1 line): - line 46: // TODO: Make configStore.getTags return ConcurrentMap build/sbt-launch-lib.bash (1 line): - line 23: # TODO - Should we merge the main SBT script with this library? client-flink/flink-2.0/src/main/java/org/apache/celeborn/plugin/flink/tiered/CelebornTierConsumerAgent.java (1 line): - line 185: // TODO: if fail to open reader, may the downstream task start before than upstream task, cpp/celeborn/client/reader/WorkerPartitionReader.h (1 line): - line 90: // TODO: add other params, such as fetchChunkRetryCnt, fetchChunkMaxRetry worker/src/main/scala/org/apache/celeborn/service/deploy/worker/storage/FlushTask.scala (1 line): - line 57: // TODO: force flush file channel in scenarios where the upstream task writes and the downstream task reads simultaneously, such as flink hybrid shuffle. openapi/openapi-client/src/main/java/org/apache/celeborn/rest/v1/master/invoker/ApiClient.java (1 line): - line 393: // TODO: implement debugging mode master/src/main/java/org/apache/celeborn/service/deploy/master/clustermeta/ha/MetaHandler.java (1 line): - line 186: // TODO: Remove `WorkerRemove` in 0.7.x version to guarantee upgrade compatibility. cpp/celeborn/network/NettyRpcEndpointRef.h (1 line): - line 42: // TODO: refactor to template function when needed. client-flink/common/src/main/java/org/apache/celeborn/plugin/flink/client/FlinkShuffleClientImpl.java (1 line): - line 220: // TODO: in segment granularity visible senarios, when the downstream reduce tasks start early cpp/celeborn/utils/StackTrace.h (1 line): - line 31: // TODO: Deprecate in favor of folly::symbolizer. cpp/celeborn/network/TransportClient.cpp (1 line): - line 160: // TODO: auto-disconnect if the connection is idle for a long time? common/src/main/java/org/apache/celeborn/common/network/sasl/SaslClientBootstrap.java (1 line): - line 72: // TODO: Hardcoding the SASL mechanism to DIGEST-MD5 for Connection Authentication. This common/src/main/java/org/apache/celeborn/common/network/sasl/CelebornSaslClient.java (1 line): - line 46: *

TODO: Currently, we have hardcoded Sasl mechanism to be DIGEST-MD5 for connection auth. We cpp/celeborn/protocol/PartitionLocation.h (1 line): - line 77: // TODO: RoaringBitmap is not supported yet. client/src/main/scala/org/apache/celeborn/client/commit/MapPartitionCommitHandler.scala (1 line): - line 238: // TODO: if support the downstream map task start early before the upstream reduce task, it should cpp/celeborn/utils/Exceptions.h (1 line): - line 33: // TODO: maybe remove the file permanently... client/src/main/scala/org/apache/celeborn/client/ChangePartitionManager.scala (1 line): - line 406: // TODO: should record the new partition locations and acknowledge the new partitionLocations to downstream task, openapi/openapi-client/src/main/java/org/apache/celeborn/rest/v1/worker/invoker/ApiClient.java (1 line): - line 393: // TODO: implement debugging mode cpp/celeborn/utils/StackTrace.cpp (1 line): - line 164: // TODO: lineNumbers has been disabled since 2009. web/vite.config.ts (1 line): - line 28: // TODO: Get the backend service address and then change this uri.