CheckIndex already validates SortedDocValues properly: reads every
document's ordinal and validates derefing all the ordinals back to bytes
from the terms dictionary.
It should not do an additional (very slow) pass where it treats the
field as if it were binary (doc -> ord -> byte[]), this is slow and
doesn't validate any additional index data.
Now that the term dictionary of SortedDocValues may be compressed, it is
especially slow to misuse the docvalues field in this way.
skipBytes() is a "relative" version of seek(), but DataInput previously
implemented it via read() calls, because DataInput's API does not
include absolute positioning methods (seek, getFilePointer).
This resulted in inefficiencies: calls to skipBytes() would cause
buffers to be allocated, bytes copied, etc.
Instead, make the subclass implement skipBytes() explicitly. The old
DataInput implementation is marked deprecated and renamed to skipBytesSlowly().
Some subclasses still implement skipBytes() via skipBytesSlowly(), to be
fixed in future improvements.
TestDirectIODirectory will currently fail if run on an unsupported
filesystem (e.g. tmpfs). Add an "assume" that probes if the filesystem
supports Direct I/O.
Also tweak javadocs to indicate correct @throws clauses for the
IndexInput and IndexOutput. You'll get an IOException (translated from
EINVAL) if the filesystem doesn't support it, not a UOE.
* Updated SOLR-8138 files for Solr 9.
This code was mostly written by Michael Suzuki, i just tweaked it to load, and updated the version of ui-grid to the 4.10 version.
* unused file, we use the .min version.
* add an entry for the ui-grid project to license file.
Co-authored-by: epugh@opensourceconnections.com <>
The 'testBackupAndRestore' method in this class was asserting that the
collection created by restore had the expected number of cores-per-node,
but the logic to compute that expected cores-per-node value failed to
account for a rarely-triggered branch that adds a 'createNodeSet' param
to the restore.
This commit updates the test logic to compute the expected
cores-per-node value when createNodeSet is passed.
Also used by FunctionScoreQuery.boostByQuery.
Could throw an exception when the query implements TwoPhaseIterator
and when the score is requested repeatedly.
Co-authored-by: Chris Hostetter <hossman@apache.org>
The recent addition of support for a "readonly" mode for collections
opens the door to restoring to already-existing collections.
This commit adds a codepath to allow this. Any compatible existing
collection may be used for restoration, including the collection that
was the original source of the backup.
Upgrade from icu 62.2 to 68.2, with Unicode 13 support.
Modify GenerateUTR30DataFiles to take the release tag as a program
argument. Gradle populates this automatically, removing a manual step
from regeneration process.
* relocate xslt related classes into scripting contrib
* relocating files to scripting and seperating out unit tests
* relocate files under test-files/scripting/solr, similar to how we do it in other contribs. deals with some issues in finding files
* Reformatting using the Google Java Format...
* use actual param name, not the variable to properly test api!
* Clean up references to paths, and deal with the mish mash of Xslt and XSLT in class names.
* Move XSLT processing out of XMLLoader
* Move TransformerProvider.Dedupe getTransformer logic.
Co-authored-by: epugh@opensourceconnections.com <>
Co-authored-by: David Smiley <dsmiley@apache.org>
Gus Heck
Kevin Watters
Robert Muir
Dawid Weiss
Eric Pugh
Jason Gerlowski
Peter Gromov
Ilan Ginzburg
epugh@opensourceconnections.com
Houston Putman
noblepaul
Tobias Kaessmann
Ignacio Vera
David Smiley
Jason Gerlowski
Chris Hostetter
Timothy Potter
Alessandro Benedetti