summaryrefslogtreecommitdiffstats
path: root/vespalib/src/tests/datastore
Commit message (Collapse)AuthorAgeFilesLines
* Always use array of AtomicEntryRef in unique store remapper.Tor Egge2022-03-171-10/+13
|
* Use AtomicEntryRef in MultiValueMappingBase.Tor Egge2022-03-091-20/+10
|
* Use AtomicEntryRef as key for unique store btree dictionary.Tor Egge2022-03-032-3/+3
|
* Move memory allocator argument to be last argument.Tor Egge2022-02-111-5/+4
|
* Add memory allocator to array store.Tor Egge2022-02-111-4/+16
|
* Add memory allocator to unique store allocatorTor Egge2022-02-104-4/+32
|
* ignore local core files produced by gtest during death testingHåvard Pettersen2022-01-271-0/+1
|
* Rename max_buffers_ratio to active_buffers_ratio.Tor Egge2021-12-141-6/+6
| | | | Add comment before calculation of max buffers to compact at once.
* Limit buffers to compact based on number of active and free buffers.Tor Egge2021-12-141-9/+31
|
* Add CompactBufferCandidates, used to select buffers to compact.Tor Egge2021-12-132-0/+100
|
* Add CompactionSpec.Tor Egge2021-12-083-3/+13
| | | | | Pass CompactionStrategy to DataStoreBase::startCompactWorstBuffers in preparation for compacting more buffers in the same compaction round.
* Add EntryRefFilter class.Tor Egge2021-12-061-22/+29
|
* Filter early on buffer id and pass vector of entries in normalize_valuesTor Egge2021-12-031-29/+148
| | | | and foreach_value ShardedHashMap member functions to limit number of callbacks.
* Don't try to move dictionary keys that won't move.Tor Egge2021-12-021-1/+25
|
* Avoid livelock when running sharded hash map unit test with valgrind.Tor Egge2021-11-281-0/+15
|
* Detect direct free of element to held buffer.Tor Egge2021-11-191-0/+52
|
* Update 2019 Oath copyrights.gjoranv2021-10-274-4/+4
|
* Update Verizon Media copyright notices.gjoranv2021-10-074-4/+4
|
* Update 2018 copyright notices.gjoranv2021-10-072-2/+2
|
* Update 2017 copyright notices.gjoranv2021-10-078-8/+8
|
* Extend test of array store compaction context.Tor Egge2021-08-181-10/+34
|
* Consider reusing active buffer.Tor Egge2021-08-163-15/+45
|
* Revert "Consider reusing active buffer."Henning Baldersheim2021-08-153-45/+15
|
* Consider reusing active buffer.Tor Egge2021-08-133-15/+45
|
* Use stats from all active buffers of same type when resizing a buffer.Tor Egge2021-08-122-16/+16
|
* Use grow factor when considering switching to a new buffer.Tor Egge2021-08-111-10/+10
| | | | This enables earlier ramp-up of buffer sizes.
* Use a list instead of a set to make building faster.Henning Baldersheim2021-06-141-0/+33
| | | | Then sort and uniq before applying the list.
* Add explicit unit test for compact_worst_shard().Tor Egge2021-04-141-0/+13
|
* Compact enum store dictionary when needed.Tor Egge2021-04-131-11/+60
|
* Sort unique values during load if needed.Tor Egge2021-04-081-3/+3
|
* Add noexcept specifier to lambda.Tor Egge2021-04-071-1/+1
|
* Extend unit test for ShardedHashMap.Tor Egge2021-04-071-0/+61
|
* Rename unit test type parameters.Tor Egge2021-04-061-16/+16
|
* Follow codestyle: getComp -> get_compHenning Baldersheim2021-04-061-4/+4
|
* Add ShardedHashComparator so that a single divison will be used for both ↵Henning Baldersheim2021-03-311-4/+4
| | | | | | dividend and remainder. The compiler will also be smarter about it as it is a known constsnt compile time both.
* Handle UniqueStoreDictionary without B-tree.Tor Egge2021-03-302-8/+53
|
* Factor out ReadSnapshot from UniqueStoreDictionary.Tor Egge2021-03-291-1/+1
|
* Rename SimpleHashMap to ShardedHashMap.Tor Egge2021-03-265-31/+31
|
* Report memory usage from unique store dictionary hash.Tor Egge2021-03-253-21/+88
|
* Track name change from ordered / unordered to btree / hash.Tor Egge2021-03-241-19/+19
|
* Extend UniqueStore unit test.Tor Egge2021-03-181-41/+117
|
* Update unordered dictionary.Tor Egge2021-03-182-3/+3
|
* Wire in ordering enum in EnumStore.Tor Egge2021-03-171-1/+1
|
* Extend FixedSizeHashMap unit test.Tor Egge2021-03-121-4/+86
|
* Remove duplicate tests.Tor Egge2021-03-112-171/+26
|
* Update prev_node_idx in loop when removing entry.Tor Egge2021-03-112-0/+321
| | | | | Use next_node_idx instead of next. Use first_used instead of used_gen or usedGen.
* Add lock free hash map.Tor Egge2021-03-103-0/+244
|
* Add simple equal to Comparator interface.Henning Baldersheim2021-03-091-0/+3
|
* Convert comparator from being a simple comparator to using an explicit less ↵Henning Baldersheim2021-03-091-1/+1
| | | | method.
* Take dead elements into account (and subtract them) when calculating how ↵Geir Storli2021-03-011-3/+30
| | | | | | | | | | | | many arrays to allocate in a datastore buffer. This avoids a problem were allocated memory can accumulate over time in components using an ArrayStore. If all documents in an array attribute vector changes from one value class to another, all elements in the buffers of the previous value class are marked dead. Those buffers will eventually be compacted. Without this fix the wanted size of the resulting compacted buffer is calculated too high, and we allocate memory we are not going to use. If we move to yet another value class later, the same problem occurs again and more memory is allocated.