summaryrefslogtreecommitdiffstats
path: root/searchlib/src/tests/common
Commit message (Collapse)AuthorAgeFilesLines
* unit test all constructors and full API for GeoLocationArne Juul2020-07-171-9/+267
|
* [lo,hi] -> [low,high]Arne Juul2020-07-151-30/+30
|
* remove now-duplicate testArne Juul2020-07-152-102/+0
|
* style fixesArne Juul2020-07-151-12/+9
|
* fix bugsArne Juul2020-07-151-2/+0
|
* finish rewrite to use GeoLocationArne Juul2020-07-152-105/+110
|
* split out geo location parser into its own distinct classArne Juul2020-07-152-38/+12
|
* refactor geo location parsingArne Juul2020-07-151-2/+2
|
* add common geo location parsingArne Juul2020-07-152-0/+144
|
* Consolidate search for GTest.Tor Egge2020-07-072-2/+0
|
* Use find_package to find gtest library.Tor Egge2020-06-292-2/+4
|
* Prepare to support matched-elements-only for arrays and weighted sets of ↵Geir Storli2020-05-204-61/+64
| | | | | | | primitive types. This renames StructFieldMapper to MatchingElementsFields with added support for basic fields. Also some terminology is aligned.
* Add protection to avoid going out of bounds when handling an empty bitvector.Henning Baldersheim2020-05-091-0/+16
|
* Follow up on code PR comments.Henning Baldersheim2020-05-081-12/+12
|
* In order to handle and/or/andnot where the left hand side is longer than the ↵Henning Baldersheim2020-05-071-0/+41
| | | | | | | | | right hand size, we handle it as if it had been false padded. This is to handle the case where you end up below a SourceBlender, where the disk indexes have different ages and docId limits. These padded bits will never be accessed, as they will never be chosen by the source blender. But having well defined behavior is always good.
* Reduce code duplication. Rand48 class is in vespalib.Tor Egge2020-04-251-3/+3
|
* Move sequenced task executors to staging vespalibHenning Baldersheim2020-04-048-727/+0
|
* adaptive sequenced executor (experimental)Håvard Pettersen2020-03-263-18/+317
|
* Wire in control of executor type.Henning Baldersheim2020-02-281-2/+2
|
* Add single threaded thoughput optimized executor with high and low watermark ↵Henning Baldersheim2020-02-271-1/+7
| | | | at 25% / 75%.
* Add getStats and setTaskLimit to interface to make it easy to swap ↵Henning Baldersheim2020-02-272-37/+37
| | | | | | implementation. Also make do with ISequenceHandlerInterface.
* GC unused include and update copyright.Henning Baldersheim2020-02-252-1/+2
|
* Add benchmark for sequenced task executor.Henning Baldersheim2020-02-242-0/+31
|
* Add testing of growing of bitvector when count is invalidHenning Baldersheim2020-01-291-7/+15
|
* Maintain the cached bitCount to avoid cost query time.Henning Baldersheim2020-01-221-11/+11
|
* Include algorithm header file when needed.Tor Egge2020-01-081-0/+1
|
* Add test for distribution of tasksHenning Baldersheim2020-01-041-0/+11
|
* Improve the distribution by using a indirect mapping that should improveHenning Baldersheim2020-01-031-7/+5
| | | | chance for better loaddistribution.
* Update tests after changing how executors are selected.Henning Baldersheim2020-01-031-6/+4
|
* GC some unused code related to fs4 protocol that has now gone missing.Henning Baldersheim2019-09-273-675/+0
|
* added function identifying which elements matchedHåvard Pettersen2019-09-194-0/+115
| | | | | only a skeleton for now; outlines the interface between the summary generator and the matcher.
* support binary values in feature setsHåvard Pettersen2019-06-131-44/+49
| | | | this will be used to store serialized tensors
* Let capacity be set based on amount of memory allocated.Henning Baldersheim2019-06-051-8/+9
|
* Move RcuVector and relevant support classes to vespalibTor Brede Vekterli2019-05-233-313/+0
| | | | | | | | | | | | | Having RCU support available across all our C++ modules open up new opportunities for optimizations. This changes the following: - `RcuVector` moved from `searchlib` to `vespalib` - `MemoryUsage` moved from `searchlib` to `vespalib` - Introduce a simplified, more generic `GrowStrategy` in `vespalib` used by the moved `RcuVector` which does not have any notion of documents. Existing searchlib `GrowStrategy` gets a simple function to convert to this generic strategy.
* Remove files used by old unit test runner.Tor Egge2019-02-1913-13/+0
|
* Follow api changeHenning Baldersheim2018-12-191-6/+2
|
* Return information if a buffer is held.Henning Baldersheim2018-12-181-6/+6
| | | | | Use std::make_unique. Minor unifying cleanup.
* Avoid depending of ResultSet. Use more simpler types. Also remove some ↵Henning Baldersheim2018-08-221-71/+2
| | | | unused code and hide some implementation details.
* - Remove the execute(string, ...) and force the use of ExecutorId.Henning Baldersheim2018-06-171-6/+5
| | | | | - Remove some double bookkeeping in AttributeWriter. - Ensure that we always use attribute.getNamePrefix() to compute executor id.
* Use a explicit type to avoid confusion on what an intger number mean. ↵Henning Baldersheim2018-06-111-2/+2
| | | | ComponentId or ExecutorId.
* Add test for getNumExecutorsHenning Baldersheim2018-06-111-0/+5
|
* Merge pull request #4989 from vespa-engine/balder/filter-qflagsHenning Baldersheim2018-02-091-2/+2
|\ | | | | Ensure that we filter and QFLAGS and do not automatically say we have…
| * Ensure that we filter and QFLAGS and do not automatically say we have the ↵Henning Baldersheim2018-02-091-2/+2
| | | | | | | | same capabilities as our master.
* | Compute nodes queried/replied more correctly during merge to get it more ↵Henning Baldersheim2018-02-091-2/+2
|/ | | | correct in MLD cases.
* Also keep track of nodes queried/replied.Henning Baldersheim2018-02-061-13/+11
|
* Replace tab with spaces.Geir Storli2017-09-131-2/+2
|
* Moved databuffer and compresssion to vespalibHenning Baldersheim2017-08-281-1/+3
|
* Include only what you need.Henning Baldersheim2017-08-182-16/+12
|
* Add some more tests for clearIntervalHenning Baldersheim2017-08-021-0/+31
|
* Update copyright headersJon Bratseth2017-06-1418-18/+18
|