aboutsummaryrefslogtreecommitdiffstats
path: root/searchlib
Commit message (Collapse)AuthorAgeFilesLines
* Add memory allocator to enum store.Tor Egge2022-02-113-2/+19
|
* Move memory allocator argument to be last argument.Tor Egge2022-02-112-3/+3
|
* Add memory allocator to array store.Tor Egge2022-02-113-4/+4
|
* Add memory allocator to unique store allocatorTor Egge2022-02-102-2/+2
|
* Merge pull request #21124 from ↵Lester Solbakken2022-02-101-0/+1
|\ | | | | | | | | vespa-engine/havardpe/run-onnx-model-to-detect-output-sizes run onnx model to detect unknown output sizes
| * run onnx model to detect unknown output sizesHåvard Pettersen2022-02-091-0/+1
| |
* | Keep using same memory allocator when resizing rcu vector.Tor Egge2022-02-091-1/+1
|/
* vector of string_id tends to become very large. Use mmap allocation ↵Henning Baldersheim2022-02-081-3/+3
| | | | automatically.
* Merge pull request #21082 from vespa-engine/balder/reduce-code-visibilityHenning Baldersheim2022-02-072-7/+11
|\ | | | | Reduce code visibility and include only what you need from config lib…
| * Reduce code visibility and include only what you need from config library.Henning Baldersheim2022-02-062-7/+11
| |
* | Check return value from Sync() and Close()Henning Baldersheim2022-02-064-9/+12
|/
* geo distance for streamingArne H Juul2022-02-034-22/+38
| | | | | | | * try to make the distance rank features work also in streaming mode * don't check the index environment too closely (it has wrong information) * expect matching locations from the query to have the field name (without _zcurve suffix) as their "attribute"
* avoid using uninitialized valuesArne H Juul2022-02-021-0/+2
|
* Merge pull request #20998 from vespa-engine/arnej/add-km-distance-outputArne H Juul2022-02-017-2/+323
|\ | | | | add extra output with distance in km
| * add unit tests for new featuresArne H Juul2022-01-312-1/+67
| |
| * add great_circle_distance featureArne H Juul2022-01-314-0/+249
| |
| * add extra output with distance in kmArne H Juul2022-01-311-1/+7
| |
* | Merge pull request #20999 from ↵Henning Baldersheim2022-01-3110-75/+112
|\ \ | |/ |/| | | | | vespa-engine/geirst/cpu-category-tag-document-store-tasks Tag all document store tasks executed on the shared executor with cpu…
| * Tag all document store tasks executed on the shared executor with cpu category.Geir Storli2022-01-3110-75/+112
| |
* | Add missing includeHenning Baldersheim2022-01-311-0/+1
| |
* | Merge pull request #20966 from ↵Tor Egge2022-01-3132-445/+272
|\ \ | | | | | | | | | | | | vespa-engine/balder/reduce-code-duplication-for-write Refactor to reduce code duplication for write path.
| * | Refactor to reduce code duplication for write path.Henning Baldersheim2022-01-2732-445/+272
| | |
* | | Merge pull request #20965 from vespa-engine/balder/gc-code-only-used-in-testsTor Egge2022-01-3127-258/+79
|\ \ \ | | | | | | | | GC unused Load/Save code for StringEnum
| * | | Hide Fast_BufferedFile better.Henning Baldersheim2022-01-2724-74/+75
| | | |
| * | | GC unused Load/Save code for StringEnumHenning Baldersheim2022-01-273-184/+4
| |/ /
* | / Since we no longer have a minimum 1G summary file size, we should no longer ↵Henning Baldersheim2022-01-311-1/+6
| |/ |/| | | | | | | | | | | have 1G initial buffer. Reducing first allocation to 64M. Use named literals.
* | Merge pull request #20980 from ↵Henning Baldersheim2022-01-285-18/+27
|\ \ | | | | | | | | | | | | vespa-engine/geirst/tag-tasks-and-threads-with-cpu-category Tag tasks and threads with cpu category
| * | Tag tls and operation response thread pools with cpu category WRITE.Geir Storli2022-01-281-9/+11
| | |
| * | Tag cpu category for tasks run in the proton shared executor.Geir Storli2022-01-284-9/+16
| |/
* | convert ITEM_GEO_LOCATION_TERM to something usefulArne H Juul2022-01-282-6/+9
| |
* | Merge pull request #20968 from ↵Henning Baldersheim2022-01-281-16/+18
|\ \ | | | | | | | | | | | | vespa-engine/balder/ignore-spread-in-the-active-file Ignore the current active file when computing spread.
| * | Locate similar methods together and improve comment.Henning Baldersheim2022-01-281-18/+18
| | |
| * | Ignore the current active file when computing spread.Henning Baldersheim2022-01-281-3/+5
| |/
* | Merge pull request #20973 from vespa-engine/arnej/just-fix-dist-featureHenning Baldersheim2022-01-281-2/+3
|\ \ | | | | | | pick the right attribute for streaming
| * | pick the right attribute for streamingArne H Juul2022-01-281-2/+3
| | | | | | | | | | | | | | | | | | | | | * This would detect the wrong field name because with SearchVisitor all fields are available as attributes; SearchVisitor::PositionInserter would not be triggered because the wrong field name was hinted for attribute access.
* | | Merge pull request #20956 from ↵Håvard Pettersen2022-01-287-26/+95
|\ \ \ | |_|/ |/| | | | | | | | vespa-engine/havardpe/auto-unboxing-of-scalar-results-from-interpreted-ranking-expressions auto-unbox scalar results from interpreted ranking expressions
| * | auto-unbox scalar results from interpreted ranking expressionsHåvard Pettersen2022-01-277-26/+95
| |/
* / Refactor to reduce code duplication.Henning Baldersheim2022-01-272-221/+135
|/
* Add [[nodiscard]] to Read() tooHenning Baldersheim2022-01-262-3/+5
|
* Add [[nodiscard]] to Write2 and CheckedWrite tooHenning Baldersheim2022-01-263-4/+6
|
* Add missing newlineHenning Baldersheim2022-01-261-1/+1
|
* Add sanity checks on Sync() and Close() callsHenning Baldersheim2022-01-261-5/+13
|
* Scope FastOS_File to avoid explicit Close().Henning Baldersheim2022-01-262-51/+49
|
* - Avoid explicit Close() when not necessary. Destructor does the necessary.Henning Baldersheim2022-01-2614-300/+36
| | | | | - Check Close() and Sync() results. - GC unused code.
* - assert result of Close() in destructor of FastOS_File and FastOS_BufferedFile.Henning Baldersheim2022-01-267-26/+4
| | | | | | | - Check result of Close() - Check result of Sync() - Scope FastOS_File to avoid explicit Close(). - NULL -> nullptr
* Improve diskindex write error detection.Tor Egge2022-01-264-35/+39
|
* Only require vespalib::Executor instead of vespalib::ThreadExecutorHenning Baldersheim2022-01-194-6/+6
|
* Avoid spinning up a separate thread just for waiting when you can just as ↵Henning Baldersheim2022-01-192-11/+17
| | | | easily do it yourself.
* Since fusion task are now smaller, there is no need to explicit limit number ↵Henning Baldersheim2022-01-194-16/+12
| | | | of concurrent tasks.
* Schedule small tasks for field merge.Tor Egge2022-01-178-26/+179
|