* Remove BytesRef usage from XContentParser and its subclasses
This removes all the BytesRef usage from XContentParser in favor of directly
returning a CharBuffer (this was originally what was returned, it was just
immediately wraped in a BytesRef).
Relates to #28504
* Rename method after Ryan's feedback
This commit fixes the Javadocs for the class o.e.x.r.j.RollupIndexer as
these Javadocs were referring to instance methods on the class
incorrectly (using a this prefix).
Original commit: elastic/x-pack-elasticsearch@fdcc7338f9
The original example resulted in a 400 error due to the example being `-` separated instead of the default `.` separation.
```
failed to parse date field [2001-01-01] with format [YYYY.MM.dd]
```
Adds a usage example of the JLH score used in significant terms aggregation.
All other methods to calculate significance score have such an example
Closes#28513
This wraps the stream (`.streamInput()`) that is passed to many of the
`createParser` instances in the enclosing (or a new) try-with-resources block.
This ensures the `BytesReference.streamInput()` is closed.
Relates to elastic/x-pack-elasticsearch#28504
Original commit: elastic/x-pack-elasticsearch@7546e3b4d4
* Wrap stream passed to createParser in try-with-resources
This wraps the stream (`.streamInput()`) that is passed to many of the
`createParser` instances in the enclosing (or a new) try-with-resources block.
This ensures the `BytesReference.streamInput()` is closed.
Relates to #28504
* Use try-with-resources instead of closing in a finally block
This change ensures that we ignore terms removed from the analysis rather than returning a match_no_docs query for the part
that contain the stop word. For instance a query like "the AND fox" should ignore "the" if it is considered as a stop word instead of
adding a match_no_docs query.
This change also fixes the analysis of prefix terms that start with a stop word (e.g. `the*`). In such case if `analyze_wildcard` is true and `the`
is considered as a stop word this part of the query is rewritten into a match_no_docs query. Since it's a prefix query this change forces the prefix query
on `the` even if it is removed from the analysis.
Fixes#28855Fixes#28856
Pruning tombstones is quite expensive since we have to walk though all
deletes in the live version map and acquire a lock on every value even though
it's impossible to prune it. This change does a pre-check if a delete is old enough
and if not it skips acquireing the lock.
This moves the `payload = null;` statement to above the asynchronous
HTTP call. This helps to avoid a race condition relative to `doClose`
asserting that it is `null`.
This is only a realistic situation during a shutdown situation because
the thread responds immediately before it can be nullified and freeable,
which is not a realistic scenario in any other situation.
Original commit: elastic/x-pack-elasticsearch@eb3a6ff118
Remove functions without a backing matrix agg
MatrixAgg works across multiple fields and exposing it directly in SQL
does not work. Instead isolated functions are exposed which get folded
and optimized into one matrix agg per field. Thus not all matrix
functions can be exposed in SQL, at least at this time.
Instead of depending on the plugin directly, depend on the plugin client
jar (matrix-agg-client)
Remove outdated test
Original commit: elastic/x-pack-elasticsearch@ec9b31bf59
The type of BUFFER_LENGTH needs to be an integer (not NULL) and unsigned
indicate the opposite of signed.
Change isSigned from Object to primitive
Since all the consumer of isSigned expect a primitive, an Object is just causing troubles by being null.
Update description table
Original commit: elastic/x-pack-elasticsearch@8e1960bdb5
Use AggregatorTestCase's `newIndexSearcher()` instead. Lucene's
version can randomly wrap with IndexReader with things we can't handle
like ParallelCompositeReader
Original commit: elastic/x-pack-elasticsearch@b4c0e9a601
The test job helper randomizes the index name with 1-10 characters,
which can lead to randomized index names to overlap and show fewer
caps than the test expects.
The solution is to just use index names "0"-"24" to ensure none
of the names overlap, and thus the caps don't overlap.
Original commit: elastic/x-pack-elasticsearch@74a6d13213
The arrangement of the final latch meant the latch could countdown,
then the test ends before the await() triggers which caused the
thread to be interrupted and fail. The whole arrangement was incorrect
anyhow.
We need to await the latch before sending the search response as before,
but move the final atomicBoolean to the second time the persistent
task status is updated which is a signal that we are done
and can end the test
If these tests continues to be flaky, we should probably just remove them.
The headers are tested elsewhere and not required to be tested in this
context.
Original commit: elastic/x-pack-elasticsearch@0cf5603972
Increase the default limit of `index.highlight.max_analyzed_offset` to 1M instead of previous 10K.
Enhance an error message when offset increased to include field name, index name and doc_id.
Relates to https://github.com/elastic/kibana/issues/16764
* Clarifies how the query_string splits textual part to build a query
Whitespaces are not considered as operators anymore in 6x but the documentation is not clear about it.
This commit changes the example in the documentation and adds a note regarding whitespaces and operators.
Closes#28719
It is only a comment, but can confuse those reading the code
Used 6.0 as an arbitrary elasticsearch.version value since it is version that required Java 8
Values for the network.host setting can often contain a colon which is a
character that is considered special by YAML (these arise in IPv6
addresses and some of the special tags like ":ipv4"). As such, these
values need to be quoted or a YAML parser will be unhappy with
them. This commit adds a note to the docs regarding this.
Analysis limits contain settings that affect the resources
used by ML jobs. Those limits always take place. However,
explictly setting them is not required as they have reasonable
defaults. For a long time those defaults lived on the c++ side.
The job could just not have any explicit limits and that meant
defaults would be used at the c++ side. This has the disadvantage
that it is not obvious to the users what these settings are set to.
Additionally, users might not be aware of the settings existence.
On top of that, since 6.1, the default model_memory_limit was lowered
from 4GB to 1GB. For BWC, this meant that jobs where model_memory_limit
is null, the default of 4GB applies. Jobs that were created from 6.1
onwards, contain an explicit setting for model_memory_limit, which is
1GB unless the user sets it differently. This adds additional confusion.
This commit makes analysis limits an always explicit setting on the job.
Regardless of whether the user sets custom limits or not, the job object
(and response) will contain the full analysis limits values.
The possibilities for interpretation of missing values are:
- the entire analysis_limits is null: this may only happen for jobs
created prior to 6.1. Thus we set the model_memory_limit to 4GB.
- analysis_limits are non-null but model_memory_limit is: this also
may only happen for jobs prior to 6.1. Again, we set memory limit to
4GB.
- model_memory_limit is non-null: this either means the user set an
explicit value or the job was created from 6.1 onwards and it has
the explicit default of 1GB. We simply keep the given value.
For categorization_examples_limit the default has always been 4, so
we fill that in when it's missing.
Finally, note that we still need to handle potential null values
for the situation of a mixed cluster.
Original commit: elastic/x-pack-elasticsearch@5b6994ef75
I'm really really sad to be removing the cli-fixture but I've had
trouble with it leaking recently it is pretty slow. Beyond that, we'd
prefer that our test fixture only fixture things that are external
depndencies.
So, yeah, I'm removing it. So we get faster tests and no chance of
leaking processes. We lose some "realness" in the tests. Instead of
interacting with the CLI like a real user we embed it in the test
process. That means we don't test the forking, we don't test the
executable jar, and we don't test the jLine console detection stuff. On
the other hand we were kind of forcing the jLine console detection stuff
in a funky way with the fixture anyway. And we test the executable jar
in the packaging tests. And that'll have to do.
I haven't renamed `RemoteCli` because it'd bloat this commit with
mechanical changes that'd make it hard to review. I'll rename it in a
followup commit.
This also updates jLine so we can disable blinking to matching
parentheses during testing. I have no clue why, but this wasn't
happening when we used the fixture. The trouble with the blinking is
that it is based on *time* so it slows things down. Worse, it works
inconsistently! Sometimes it spits out sensible ascii codes and
sometimes it, well, spits out weird garbage. When you use it in person
it works fine though. So we keep it on when not testing.
Cleans up some redundancy in when testing CLI errors. Less copy and
paste good.
I was tempted to disable the xterm emulation entirely while working on
this because upgrading jLine changed a few things and it was a real pain
to update. But If we turned that off then we'd have *nothing* testing
the colors and such. That'd be a shame because we use color in the
output to commicate stuff. I like it so I don't want to break it.
While I was there, I replaces the cli connector's `PrintWriter` with a
`BufferedWriter`. The `PrintWriter` was kind of a trap because `println`
would fail to work properly on windows because we force the terminal
into xterm mode and it doesn't know what to do with windows line
endings. Windows.....
Additionally I fixed a race condition between disabling echo when
reading passwords and fast writers. We were disabling the echo shortly
after sending the prompt. A fast enough writer could send us text before
the echo disable kicked in. Now I delegate to `LineReader#readLine`
with a special echo mask that disables echo. This is both easier to test
and doesn't seem to have the race condition. This race condition was
failing the tests because they are so much faster now. Yay!
Original commit: elastic/x-pack-elasticsearch@d0ec027396
When virtual lock is not possible because JNA is unavailable, we log a
warning message. Yet, this log message refers to mlockall rather than
virtual lock, presumably because of a copy/paste error. This commit
fixes this issue.
This commit specifies that the working directory of the destroy task for
destroying test VMs is the root of the build. This is necessary in case
the build was run from a sub-directory, the Vagrant command would then
not be able to locate the Vagrantfile for the VMs in question.
Today we do not destroy Vagrant boxes before tests. This is because
constantly reprovisioning these boxes is time-consuming. Yet, not
destroying these boxes can lead to state being left around that impacts
subsequent test runs. To address this, we now always destroy these boxes
before tests and provide a flag to set if this is not desired while
iterating locally.
Add basic support for catalog parameters in SYS COLUMN
Pass an empty string instead of a null inside the prepared statement
Don't use pattern for catalog in getColumns
Original commit: elastic/x-pack-elasticsearch@17e9e851a0
Lee Hinman
Igor Motov
Jason Tedor
Lisa Cawley
Jonathan Pool
Aaron Mildenstein
Menno Oudshoorn
Sergey Galkin
Jim Ferenczi
Alexander Reelsen
Tim Roes
Simon Willnauer
Chris Earle
Costin Leau
polyfractal
Mayya Sharipova
Andrew Odendaal
Ivan Brusic
olcbean
Luca Cavanna
Bojan Matić
Nik Everett
Yu
FUJI Goro
Dimitris Athanasiou
lzh3636
David Roberts
David Kyle
Ryan Ernst