Summary: 512 instances, 462 unique
Text Count
// TODO: shouldn't have to trim, sha files should not have trailing newline 1
// TODO: Fix matrix stats to treat neg inf as any other value 1
// TODO: reconsider reverting to simplified approach once https://github.com/Azure/azure-sdk-for-java/issues/26064 3
* TODO once checksum verification on snapshotting is implemented this test needs to be fixed or split into several 1
// TODO: Evaluate if we still need to catch this exception 1
// TODO we might use the new NIO2 API to get real notification? 1
// TODO: use FieldHighlighter::maybeAddPassage 1
// TODO: Should all existing templates be deleted first? 1
// TODO: remove this property in 8.0 1
// TODO: temporary work-around for #5531 2
// TODO this method helps because of asMultiBucketAggregator. Once we remove it we can move this logic into the aggregators. 1
// TODO: (maybe) Create a new bucket for all distant docs and merge down to shardSize buckets at end 1
//TODO: please convert to unit tests! 1
// TODO: would be nice to have more random data here 2
// TODO these ain't gonna work with runtime fields 1
// TODO once composite is plugged in to the values source registry or at least understands Date values source types use it 1
// TODO: Implement this. 1
// TODO Maybe just clear out non-ongoing snapshot recoveries is the node is master eligible, so that we don't 1
// TODO: We should move these datastructures to IndexNameExpressionResolver, this will give the following benefits: 1
// TODO: pluggability? 1
// TODO: remove this leniency, these settings should exist together and be validated 2
// TODO: check for non public classes that seem like tests 1
// TODO - need to check with mapping that this is indeed a text field.... 1
// TODO: We haven't had timestamp for Index operations in Lucene yet, we need to loosen this check without timestamp. 1
case ParsedScrollId.QUERY_AND_FETCH_TYPE: // TODO can we get rid of this? 1
// TODO: WTF? 1
// TODO: Standardize exception types 1
return ScoreMode.COMPLETE; // TODO: how can we know if the script relies on scores? 1
// TODO: we need another name than "http transport"....so confusing with transportClient... 1
// TODO how do we identify dynamically that its a binary value? 1
// TODO: ensure empty/no PARAMETERS if parameterTypes.length == 0? 1
// TODO: specialize based on compiled.type: for ALL and prefixes (sinkState >= 0 ) we can avoid i/o and just set bits. 1
// TODO how might we use JtsSpatialContextFactory to configure the context (esp. for non-geo)? 1
// TODO this one seems useless 1
// TODO: how can we compute this? 1
// TODO: not great we don't test deletes GC here: 1
// TODO: Should it expose a count (current approach) or return null when there are no more values? 1
// TODO: Should we rather pass roundUp=true? 1
// TODO resolve 1
// TODO: JNA doesn't have netbsd support, but this mechanism should work there too. 1
// TODO: I think startTime() should become part of ActionRequest and that should be used both for index name 1
// TODO: change order when delete index api also updates the data stream the index to be removed is member of 1
// TODO: Generalize this class and add it as a utility 1
// TODO: Enable restore for open indices 1
// TODO: Find a way to remove this query and let doToQuery(...) just return the query from JoinUtil.createJoinQuery(...) 1
// TODO: can we handle missing bucket when using index sort optimization ? 1
//TODO: cleanup 1
* //TODO COMPAT will be deprecated in jdk14 https://bugs.openjdk.java.net/browse/JDK-8232906 1
// TODO this is probably super slow, but just used for testing atm 1
// TODO: remove this leniency 1
// TODO we should allow this, but at the moment levels is used to build bookkeeping variables 1
* TODO: Add shard level cleanups 1
// TODO: be and g and both based on the bose-einstein model. 1
// TODO: There must be a better option than this. 1
// TODO: to calculate the highest upper bound for the two types and return that. 1
// TODO as we move analyzers out of the core we need to move some of these into HighlighterWithAnalyzersTests 1
// TODO: can we move this to the .env package and make it package-private? 1
// TODO support Z?? 1
// TODO: In the rare case we still haven't reached a correct promotion we need 1
// Temporary shim for aliases. TODO deprecate after they are moved 1
// TODO remove when everything is parametrized 1
// TODO: report on child aggs as well 1
// TODO: improve serialization of BigInteger 1
// TODO: if we really need we can break out name (e.g. "hack" or "scriptEngineService" or whatever). 1
// TODO: It is worth revisiting the design choice of creating a placeholder entry in snapshots-in-progress here once we have a cache 1
// TODO: it appears that this can never be true? 1
// TODO: this is wrong, i.e. we are comparing apples and oranges at this point. It would be better if we always rescored all 1
} else if (id == 1 || id == 3) { // TODO this bwc layer can be removed once this is back-ported to 5.3 QUERY_AND_FETCH is removed 1
// TODO: when types are complete removed just use String instead for the id: 1
// TODO: remove once refresh doesn't fail immediately if there a master block: 1
// TODO: This is nonsense. We allow the value to be null (via constructor), but don't allow it to be set to null. This means 1
// TODO: remove this when guice is removed, and exceptions are cleaned up 1
// try the other suffix...TODO: get rid of this, just support ending in .txt 1
// TODO: Remove these once https://github.com/elastic/elasticsearch/issues/46091 is fixed 1
// TODO why does wildcard resolver throw exceptions regarding non wildcarded expressions? This should not be done here. 1
// This is a random starting point alpha. TODO: revisit this with actual testing and/or make it configurable 1
// TODO: this class can be removed in master once 7.x is bumped to 7.4.0 1
* TODO: The number of bytes written during the merge should be accessible in OneMerge. 1
// TODO: Remove this class when https://github.com/elastic/elasticsearch/issues/32981 is addressed. 1
// TODO: we should fail the entire request because the topdocs 1
// TODO: allow publishing multiple addresses 1
// TODO: remove this once ctx isn't added to update script params in 7.0 1
// TODO lookup the min and max rounding when appropriate 1
// TODO: check that object did not change when stream is reopened (e.g. based on etag) 1
* and any delete/write operations. TODO: This will not work if we also want to clean up at the shard level as those will involve writes 1
// TODO we usually have a heap here so we could use that to build the results sorted. 1
* This class contains all the settings which are required and owned by {TODO link ShardIndexingPressure}. These will be 1
// TODO do we need to also fail this if we can't delete the pending commit file? 1
// TODO: is MANAGEMENT TP the right TP? Right now this is an admin api (see action name). 1
// TODO: remove this when the transition from Joda to Java datetimes is completed 3
// TODO: should verify this against the Vagrantfile, but would need to do so in afterEvaluate once vagrantfile is unmodifiable 1
// TODO this is fairly expensive to compute. Can we avoid needing it? 1
// TODO these are used on tests only - move into utils class 2
// TODO: this is a hack!! We should instead keep track of incoming (relocated) shards since we know 1
// TODO - clean this up when OSS flavor is removed in all of the code base 2
// TODO we should be checking mapped field types, not lucene field types 1
// TODO: remove this when the transition from Joda to Java datetimes is completed 2
// TODO fix for backport https://github.com/elastic/elasticsearch/pull/52547 1
// TODO: replace this with #closeContexts 1
// TODO: I think we can remove these arrays. it isn't worth the effort, for operations on all indices. 1
// TODO: check that the testing tasks are included in the right task based on the name ( from the rule ) 1
// TODO it'd be faster to iterate many ords at once rather than one at a time 1
// TODO: https://github.com/elastic/elasticsearch/issues/51836 1
// TODO: do we want to call updateTask.onFailure here? 1
// TODO: improve serialization of BigInteger 1
// TODO merge the contents of this file into BucketOrder. The way it is now is relic. 1
// TODO: AggregatedDfs. Currently the idf can be different then when executing a normal search with explain. 1
// TODO: report on sub collectors 1
// TODO: maybe track these in bitsets too? this is trickier... 1
// TODO: only here temporarily for BWC development, remove once complete 2
// TODO: move this out if we need it for other reasons 1
// TODO: feels hacky, a block disables state persistence, and then we clean the allocated shards, maybe another flag in blocks? 1
// TODO this is used in date range's collect so we should optimize it too 3
// TODO: Is there a reason not to get the prepared rounding in the supplier itself? 1
// TODO: move PathUtils to be package-private here instead of 1
// TODO move the request to the correct node. maybe here or somehow do it as part of startup for reindex in general.... 1
return Float.MAX_VALUE; // TODO: what would be a good upper bound? 2
* TODO: We should be able to create passages incrementally, starting from the offset of the first match and expanding or not 1
// TODO: also support the structured formats of geo points 1
// TODO: override toXContent 1
// TODO we should refactor to remove this, since buckets should be immutable after they are generated. 1
// TODO if perf is a problem, we can specialize a subclass to avoid conditionals on each iteration 1
// TODO: ADD SEQ NO! 1
// TODO after backporting we expect this never to happen in 8.x, so adjust this to throw an exception instead. 1
// TODO: no jars should be an error 1
// TODO: figure out why GCE is so bad like this 1
// TODO: can we use one instance of the script for all buckets? it should be stateless? 1
// TODO missing node names should only be allowed in tests 1
// TODO: Support for exclusive ranges, pending resolution of #40601 1
// TODO: currently we load up the suggester for reporting its size 1
// TODO: consider moving this to StoreStats 1
// TODO rename metadataToXContent and have only SearchPhaseExecutionException use it, which prints out complex objects 1
// TODO: use diamond once JI-9019884 is fixed 2
// TODO: sometimes use _bulk API 1
ExecutionMode execution = ExecutionMode.MAP; // TODO global ords not implemented yet, only supports "map" 1
errors.merge(e.getErrors()); // TODO: source 2
* TODO: We should maybe rewrite this with the flexible query parser which matches the same syntax with more freedom. 1
// TODO: maybe we should just put a scheduled job in threadPool? 1
// TODO this needs heavy cleanups before we can rewrite it 1
// TODO: We shouldn't use the generic thread pool here as we already execute this from the generic pool. 1
// TODO maybe we need a getter on Namedanalyzer that tells if this uses broken Analysis 1
// TODO stop ignoring parameters such as charset... 1
// TODO Gradle should provide property based configuration here 1
// TODO: We shouldn't be blanket catching and suppressing all exceptions here and instead handle them safely upstream. 2
// TODO Discuss moving compileOptions.getCompilerArgs() to use provider api with Gradle team. 1
// TODO we allow null here for the _all field, but this is forbidden in the parser. Re-check 1
// TODO keep track of unsuccessful requests in different entries 1
// TODO we can maybe factor out MapperService here entirely by introducing an interface for the lookup? 1
// TODO: why do we support suffix of LICENSE *and* LICENSE.txt?? 1
// TODO: shouldn't this skip, not parse? 1
// TODO we should makes this a hard validation and throw an exception but we need a good way to add backwards layer 1
// TODO: bind the overridden keys using multibinder 1
/* TODO: this silently ignores a timezone argument, whereas NumberFieldType#docValueFormat throws if given a time zone. 1
// TODO: make this a proper setting validation logic, requiring multi-settings validation 1
} // TODO if this logging statement is correct then we are missing an else here 1
// TODO: Index settings and mapping 1
public static final String SEARCH_SOURCE = "search"; // TODO: Make source of search enum? 1
// TODO execute CommandLoggingConfigurator.configureLoggingWithoutConfig() in the constructor of commands, not in beforeMain 1
// TODO we should check the checksum in lucene if we hit an exception 1
// TODO: move all aggs to the same package so that this stuff could be pkg-private 1
// TODO: can we enable this check for all origins except primary on the leader? 1
//TODO public for tests 5
// TODO Stop expecting nulls here 1
// TODO we should somehow inform the user that we failed 1
// TODO: forbidden apis does not yet support java 15, rethink using runtime version 1
// TODO: upgrade the HDRHistogram library 1
// TODO: fail shard? This will otherwise have the local / global checkpoint info lagging, or possibly have replicas 1
// TODO: get these fixed in aws sdk 3
// TODO: Upgrade handling (reindex script?) 1
// TODO: how to do it in the future? 1
// TODO: get this fixed 2
// TODO only millisecond granularity here! 1
// TODO revisit 1
// TODO: replace with bwc formatter 2
// TODO we can almost certainly do better here by building something fit for purpose rather than trying to lego together stuff 1
// TODO minJava >= 9 : use ClassLoader.getDefinedPackage and remove @SuppressForbidden 1
// TODO: Don't enable this leniency unless users explicitly opt-in 1
// TODO: this probably shouldn't apply to groovy at all? 1
// TODO: remove this hack when insecure defaults are removed from java 2
// TODO: Is this the correct exception type here? 3
// TODO: remove ZonedDateTime exception when JodaCompatibleDateTime is removed 1
// TODO it would be nice to assert this in the future 1
// TODO: Maybe it could also make sense to assume independent distributions of documents and eg. have: 1
* TODO: make this protected once Mapper and FieldMapper are merged together */ 1
* TODO: we execute this in parallel but it's a synced method. Yet, we might 1
if (masterNode != null) { // TODO: can we remove this? Is resending shard failures the responsibility of shardStateAction? 1
// TODO this would be better done bottom-up instead of top-down to avoid 1
// TODO: these simply trigger a noisy warning if its unable to clear the properties 1
// TODO: Should index creation not be throttled for primary shards? 1
* TODO: Make this more efficient 1
// TODO should we remove ones that are not part of full list? 1
// TODO: add more fine grained mem stats values to per segment info here 1
// TODO: TextBytesOptimization we can use a buffer here to convert it? maybe add a 1
// TODO: Make this phase optional #60156 2
// TODO: optimize to skip any bundles not connected to the candidate plugin? 1
* TODO: We could be smarter here and group the shards by index and then 1
index = 0; // TODO Should this be pos instead of 0? This assigns exact matches to the southernmost component. 2
// TODO: get this fixed 1
// TODO: follow the directions in the Javadoc for this method 1
* //TODO: instead of throwing error, find a way to sort long and unsigned_long together 1
// TODO: clean this up when types are gone 1
// TODO: Extend DirectSpellChecker in 4.1, to get the raw suggested words as BytesRef 1
// TODO revisit when moving to shorter paths on CI by using Teamcity 1
* TODO: remove this by moving defaults into mappers themselves 2
// TODO: since the plugins are unique by their directory name, this should only be a name check, version should not matter? 1
// TODO: Until we implement the retry mechanism for point in times (i.e., replace an unavailable shard with an equivalent copy), 1
* TODO: Remove in 8 1
// TODO: can we avoid having an exception here? 2
// TODO: put these in doPrivileged and move these to test-framework.policy 1
// TODO: we should add an easier to do this. It's too much of a song and dance.. 1
* TODO - Need to add package description 1
// TODO move to KeywordFieldMapper.Builder 1
// TODO (crazybob): if we can't proxy this object, can we proxy the other object? 1
// TODO: Count requests double in the circuit breaker if they need copying? 1
// TODO: create snapshot UUID in CloneSnapshotRequest and make this operation idempotent to cleanly deal with transport layer 1
// TODO: remove this property 1
// TODO: Method interceptors could actually enable us to implement 1
// TODO: Implement IMDSv2 behavior here. For now this just returns a 403 which makes the SDK fall back to IMDSv1 1
// TODO: should be validating if index options go "up" (but "down" is ok) 1
// TODO: remove this when the transition from Joda to Java datetimes is completed 4
// TODO: we should really try to get this fixed in ASM! 1
// TODO: The only implementations of this class use SAME which means the execution will not be 1
// TODO: stop using nulls here 1
// Make sure it's not the same type. TODO: Can we check for deeper loops? 1
// TODO: fix POI XWPF to not do this: https://bz.apache.org/bugzilla/show_bug.cgi?id=58597 1
// TODO: deprecate and remove, this is a noop token filter; it's here for backwards compat before we had "smartcn_tokenizer" 1
// TODO: more detailed ram usage? 1
// TODO: This should eventually be removed once ExceptionsHelper is moved to a core library jar 1
// TODO: Should we restore them at the end? 1
// TODO: Restart snapshot on another node? 1
// TODO: getThreadpool() 1
// TODO remove this and from plugin-security.policy when JDK 11 is the only one we support 1
// TODO perhaps pad out seasonal to a power of 2 and use a mask instead of modulo? 1
// TODO: sort shouldn't fail when sort field is mapped dynamically 1
* TODO we could be a bit smarter here, we don't need to fully sort necessarily 1
// TODO: remove THIS when hadoop is fixed 1
// TODO: Fallback for BWC with older predecessor (ES) versions. 1
// TODO: Remove this escape hatch in 8.x 1
// TODO We could use an FST internally to make things even faster and more compact 1
// TODO make this un-updateable 1
IndexRequestBuilder[] highCardBuilders = new IndexRequestBuilder[100]; // TODO randomize the size? 1
// TODO: do we need to report zeros? 1
// TODO: capture the connection instead so that we can cancel child tasks on the remote connections. 1
// TODO it would be nice to assert this in the future 1
// TODO: do this in 6.0: 1
// TODO: we could check IndexWriter.getMergingSegments and avoid adding merges that IW will just reject? 1
// TODO: passing null to an object seems bogus? 1
public static final String SYNC_COMMIT_ID = "sync_id"; // TODO: remove sync_id in 3.0 1
// TODO: remove this leniency, but tests bogusly rely on it 1
// TODO we might force a flush in the future since we have the write lock already even though recoveries 1
// TODO: Create a proper object to encapsulate the recovery context 1
// TODO: we should not ignore the exception and return a prefix query with the original term ? 1
* TODO: If the sort is based on score we should propagate the mininum competitive score when orderedGroups is full. 1
// TODO: Getting here means repository corruption. We should find a way of dealing with this instead of just 1
.filter(this::seemsLikeATest) // TODO when base classes are set, check for classes that extend them 1
// TODO: sometimes use bulk: 1
// TODO move subclasses that print out simple metadata to using addMetadata directly and support also numbers and booleans. 1
// TODO: Stop using nulls here 4
if (threadName.contains("[node_s_0]") // TODO: this can't possibly be right! single node and integ test are unrelated! 1
// TODO: make settings accessible only to code that registered the setting 1
// TODO revisit missing and unknown constants once Zen2 BWC is ready 1
* TODO: get rid of this and resolve it to concrete nodes in the rest layer 1
// TODO: this is public so allocation benchmark can access the default deciders...can we do that in another way? 1
// TODO: remove setting logging level via system property 1
// TODO: we should, instead, hold a "bytes reserved" of how large we anticipate this shard will be, e.g. for a shard 1
// TODO: Read translog forward in 9.0+ 1
// TODO this should be replaced by Setting.Property.HIDDEN or something like this. 1
// TODO deprecate and remove in API 2
// TODO: for the above example, remove the binding for BarImpl if the binding for FooImpl fails 1
// TODO remove in 6.0!! 1
// TODO: can we improve this? 1
// TODO: [Zach] we might want refactor and remove ExecutionMode#create(), moving that logic outside the enum 1
// TODO: can we get rid of this hack? 1
// TODO: fix PDFBox not to actually install bouncy castle like this 1
// TODO implement semi-sorting 1
// TODO is this really a good default number for max_merge_segment, what happens for large indices, 1
// TODO: Make Condition non-pluggable 1
// TODO: this catch is bogus, it means the old repo is already closed, 1
// TODO: if I create a proxy which implements all the interfaces of 1
// TODO: multi fields are really just copy fields, we just need to expose "sub fields" or something that can be part 1
// TODO: what is the difference between "number" and "numeric"? 1
// TODO: do not be lenient here 1
// TODO: Remove null usage here, by using a different aggregator for create 1
// TODO: Try to parse other named objects (e.g. stored scripts, ingest pipelines) that are part of core es as well? 1
// TODO need to figure out what subset of system indices OpenSearch Dashboards should have access to via these APIs 1
// TODO this is actually used date range's collect so we should optimize it 1
* TODO it'd be faster if we could apply the merging on the fly as we 1
// TODO this should be caught by lucene - EOF is almost certainly an index corruption 1
// TODO better way to know if the scripted metric received documents? 1
// TODO: shouldn't this be up to the ScoreMode? I.e., we should just invoke ScoreMode.combine, passing 0.0f for the 2
// TODO: deprecate this so we don't need to carry around a useless null in the wire format 1
// TODO: fix that 2
// TODO: The source might be parsed and available in the sourceLookup but that one uses unordered maps so different. 1
// TODO: perhaps introduce a ComponentTemplateMissingException? 1
// TODO: stop depending on nulls here 1
// TODO: Stop expecting nulls here 1
// TODO: we could fail the request because there is a high probability 1
// TODO: if some uses it in a scoring expression, they will get a nasty failure when evaluating...need a 2
// TODO: return a prettier name on non-Linux OS 1
// TODO: in 8.0+, only allow updating index templates, not adding new ones 1
// TODO why do we let patterns like "none,all" or "open,none,closed" get used. The location of 'none' in the array changes the 1
// TODO we can possibly support also relocating cases, where we cancel relocation and move... 1
final long phase1ThrottlingWaitTime = 0L; // TODO: return the actual throttle time 1
// TODO it'd be faster replace hasMetric and metric with something that returned a function from long to double. 1
// TODO find a better way to limit concurrent updates (and potential associated reroutes) while allowing tests to ensure that 1
// TODO: use FieldHighlighter::highlightOffsetsEnums and modify BoundedBreakIteratorScanner to work with it 1
// TODO: remove this with gradle 7.0, which no longer allows insecure urls 1
// TODO: make relocated async 1
// TODO perhaps pad out seasonal to a power of 2 and use a mask instead of modulo? 1
// TODO: This rate limiter has some concurrency issues between the two maybePause operations 1
// TODO: Change this to true when we can assume that scripts are pure functions 1
// TODO: we could specialize the single value case 1
// TODO this exception message is not really accurate for the string case. It's really disallowing regex + formatter 1
// TODO: this comparison is bogus! it would cause us to upgrade even with the same format 1
// TODO: randomize the wait for active shards value on index creation and ensure the appropriate 1
// TODO: fail if pattern contains a newline, it won't work (currently) 1
// should be able to move these tests to run against a proper cluster instead. TODO do this. 1
// TODO: set this with gradle or some other way that repros with seed? 1
// TODO should we throw an IAE if we have two nodes with the same host? 1
private static final int NUM_DOCS = 5; // TODO randomize the size? 1
// TODO consider changing this implementation to emit tokens as-we-go 1
// TODO we can potentially also consume the actual per shard results from the initial phase here in the aggregateDfs 1
// TODO: Implement custom Diff for tasks 1
// TODO a "bottom up" insert would save a couple of comparisons. Worth it? 1
// TODO apparently we don't use the MetadataSnapshot#.recoveryDiff(...) here but we should 1
* TODO - uncomment this logic from OpenSearch version 3 onwards 1
* TODO: Currently we need the concrete aggregators to build empty responses. This means that we cannot 1
// TODO: remove this in 6.0 1
// TODO: change caller logic in RoutingNodes so that this check can go away 1
// TODO: we could skip this step for snapshots with state SUCCESS 1
// TODO: Is this sufficient? 1
// TODO it may make sense to integrate the remote clusters responses as a shard response in the initial search phase and ignore hits coming 1
// TODO: can we pass trackTotalHits=subSearchContext.trackTotalHits(){ 1
// TODO - do we need something more extensible? Yet, this does the job for now... 1
// TODO this should only be allowed in tests 1
// TODO: remove this once cname is prepended to transport.publish_address by default in 8.0 1
// TODO we can probably fold this entire request ID dance into connection.sendReqeust but it will be a bigger refactoring 1
* Gradle provides "expansion" functionality using groovy's SimpleTemplatingEngine (TODO: check name). 1
// TODO: this logic can be removed when the current major version is 8 1
// TODO implement the single point optimization above 1
private long nextRebucketAt = 1000; // TODO this could almost certainly start higher when asMultiBucketAggregator is gone 1
// TODO: Maybe Fix for BWC 1
// TODO: resolve no longer generates invalid configs. Once VSConfig is immutable, we can drop this check 1
// TODO: Upgrade to newer non-blocking Azure SDK 11 and execute delete requests in parallel that way. 1
// TODO this could be incorrect... e.g. +1 + -1 1
// TODO: verify the classname exists in one of the jars! 1
// TODO: This can be moved to IndexNameExpressionResolver too, but this means that we will support wildcards and other expressions 1
// TODO cancel when a slice fails? 1
// TODO: Should we allow parallelism across repositories here maybe? 1
// TODO move to text-specific builder base class 1
// TODO: save a .clone() call 1
* TODO: should we update the in-sync allocation IDs once the data is deleted by the node? 1
// TODO: document and/or error if params contains _score? 1
// TODO: test non-aborting exceptions (Rob suggested field where positions overflow) 1
// TODO: Fallback for BWC with older OpenSearch versions. 1
* TODO: there are schemes to avoid powers of two, might want to investigate those 1
// TODO: in a subsequent change we hope to extend ThreadPool.Info to be more specific for the thread pool type 1
// TODO: useful to break down RAM usage here? 1
// TODO: Change this to true when we can assume that scripts are pure functions 1
// TODO: this interaction with sort comparators is really ugly... 1
// TODO: Add 3D support 1
// TODO: This request and its associated transport action can be folded into UpdateRequest which is its only concrete production code 1
// TODO remove context entirely when it isn't needed 1
// TODO: change to String[] 1
// TODO maybe we can make this concrete later - for now we just implement this in the base class for all initial phases 1
// TODO: Fix repo cleanup logic to handle these leaked snap-file and only exclude test-repo (the mock repo) here. 1
// TODO It'd be nice not to refresh on every slice. Instead we should refresh after the sub requests finish. 1
// TODO: This behavior breaks rest api returning http status 201 1
private static final int NUM_DOCS = 5; // TODO: randomize the size? 1
// TODO it'd be nice to combine this with BaseRestHandler's implementation. 1
// TODO Can this actually happen? Needs a test to exercise it, or else needs to be removed. 2
// TODO: Think of a fairer way to do this. Should prev.max = cur.min? 1
// TODO this one's two argument get is still broken 1
// TODO: remove ZonedDateTime exception when JodaCompatibleDateTime is removed 2
//TODO: This is a *temporary* workaround to ensure a timeout does not mask other problems 1
// TODO: if this is still too long, truncate and use ellipses 1
// TODO remove funky contexts 1
// TODO: Asserts the starting commit is a safe commit once peer-recovery sets global checkpoint. 1
// TODO: stop using LanguageIdentifier... 1
// TODO this lines up with FieldSortBuilder. Share? 1
// TODO: we should be able to default this to eg compile deps, but we need to move the licenses 1
// TODO: deprecate and remove, this is an alias to "smartcn_tokenizer"; it's here for backwards compat 1
// TODO: break down ram usage? 1
// TODO: please convert to unit tests! 1
// TODO both unwrap and subAggregator are only used here! 1
// TODO: do this without a serialization round-trip 1
// TODO: Callers should ask for the range that source should be retained. Thus we should always 1
* TODO: Should be renamed to appliedState / appliedClusterState 1
// TODO: rename all these "file" methods to "dir" 1
// TODO: clean all these property writes up, and don't allow any more in. these are all bogus! 1
// TODO: look into these and check how to get rid of them 1
// TODO: needs to be InetAddress[] 1
// TODO: add a dedicate recovery stats for the reset translog 1
// TODO: check for abstract classes that implement the naming conventions 1
* TODO currently the diversity feature of this agg offers only 'script' and 1
// TODO: move PathUtils to be package-private here instead of 1
// TODO this class can be foled into either IndicesService and partially into IndicesClusterStateService 1
// even though it's now possible to achieve better balance, so we have to do an explicit reroute. TODO fix this? 1
* //TODO: Remove this when system modules are available 1
// TODO: remove the fieldType parameter which is only used for bw compat with pre-2.0 1
// TODO review: what size cap should we put on this? 1
// TODO: deprecate and remove this field in favor of initialConnectionTimeout field that is of type TimeValue. 1
// TODO: check primary == null || primary.active() after all tests properly add ReplicaAfterPrimaryActiveAllocationDecider 1
// TODO: remove me! 1
// TODO: Only visible for testing. Perhaps move StubbableTransport from 1
// TODO: See if we can get precise result reporting. 1
// TODO: Remove this once fixed in the AWS SDK 1
// TODO: Include the current generation here once we remove keeping index-(N-1) around from #writeIndexGen 1
// TODO: move this to a unit test somewhere... 1
// TODO: expose docs stats for replicated closed indices 1
// by this version of OpenSearch. TODO TBD should we avoid indexing when possible? 1
// TODO: Once Java has a factory for those in java.lang.invoke.MethodHandles, use it: 1
from('src/main/packaging') // TODO: move all config/bin/_size/etc into packaging 1
* TODO: we should get rid of this hack, and make the rest tests use some other identifier 1
* {TODO link ShardIndexingPressureMemoryManager} and {TODO link ShardIndexingPressureStore} 1
// Only report "critical" exceptions - TODO: Reach out to the master node to get the latest shard state then report. 1
// TODO maybe we can respect throttling here too? 1
// TODO (crazybob): Filter out overridden members. 1
// TODO: can we remove this leniency yet?? 1
// TODO: create snapshot UUID in CreateSnapshotRequest and make this operation idempotent to cleanly deal with transport layer 1
// TODO: Use IndexUpgradeMergePolicy instead. We should be comparing codecs, 1
// TODO can we get around this? the cluster is RED, so what do we wait for? 1
* TODO: Refactor this as transport client is removed and this used to prepare settings for the transport client by 1
// TODO when using delete/none, we can still return the source as bytes by generating it (using the sourceContentType) 1
// TODO: Check if we could save the set null step 2
// TODO: add OpenBSD to Lucene Constants 1
// TODO: clean this up / only give it to engines that really must have it 1
// TODO: pass the latest timestamp from engine. 1
// TODO: Remove this transition in OpenSearch 3.0 1
// TODO: stop expecting nulls here 3
// TODO remove this filter once legacy ES versions are no longer supported 1
// TODO: We don't have to remove a cancelled shard from in-sync set once we have a strict resync implementation. 1
// TODO stop expecting nulls here 1
// TODO: use diamond operator once JI-9019884 is fixed 1
* TODO: Remove this section once BwC logic it references is removed 1
* TODO: Add unreferenced index metadata cleanup 1
// TODO: A better exception goes here 1
return offsetInstant.toEpochMilli(); // TODO or throw something? 1
// TODO remove once jvm.options are added to test system properties 1
// TODO maybe aggs should force millis because lots so of other places want nanos? 1
// (TODO: clean this up?) 1
// TODO: Is this the correct exception type here? 1
* TODO: optimize this to execute in a single CS update together with finalizing the latest snapshot 1
// TODO: this seems wrong, SearchPhaseResult should have a writeTo? 1
// TODO: The source might parsed and available in the sourceLookup but that one uses unordered maps so different. Do we care? 1
// TODO fix for backport of https://github.com/elastic/elasticsearch/pull/52547 1
// TODO improve this measure for when nested docs are in use 1
// Make sure it's not the same type. TODO: Can we check for deeper cycles? 1
// TODO it would be nice to share code between PreLoadMMapDirectory and HybridDirectory but due to the nesting aspect of 1
// TODO: this array is always of length 1, and testChildrenAggs fails if this is changed 1
// TODO it'd be faster to count the number of buckets in a list of these ords rather than one at a time 1
// TODO we only index single valued docs but the ordinals ends up with multi valued sometimes 1
// TODO - clean up this code when we remove all v6 bwc tests. 2
* TODO: move this into IngestService and make index/bulk actions call that 1
// TODO: add the differentiation on the metadata level in the next major release 1
// TODO: introduce a way to ignore internal statements so this assert is not triggered 1
// TODO: do we really need to store all this stuff? some if it might not speed up anything. 1
// TODO: However, for now we just return objectType that may require an extra cast. 1
// TODO: not ideal, make a better api for this (e.g. with jar metadata, and so on) 1
// TODO: Stop using null here 3
// TODO: do we really need to make the exception name in underscore casing? 1
// TODO: replace with constants class if needed (cf. org.apache.lucene.util.Constants) 1