aboutsummaryrefslogtreecommitdiffstats
path: root/searchlib
Commit message (Collapse)AuthorAgeFilesLines
* Simplify ISearchContext.Geir Storli2017-03-162-42/+21
|
* Change IRequestContext to return attributes as IAttributeVector.Geir Storli2017-03-153-25/+27
|
* Move asDocumentWeightAttribute() to IAttributeVector.Geir Storli2017-03-153-1/+6
|
* Extend IAttributeVector API with function for creating a ISearchContext.Geir Storli2017-03-158-36/+85
|
* Merge pull request #2002 from ↵Tor Egge2017-03-1442-209/+78
|\ | | | | | | | | yahoo/geirst/prepare-for-extending-iattributevector-api Geirst/prepare for extending iattributevector api
| * Move AttributeVector::SearchContext::Params to separate class in searchcommon.Geir Storli2017-03-1439-106/+75
| |
| * Move search::attribute::ISearchContext to searchcommon in preparation of ↵Geir Storli2017-03-144-103/+3
| | | | | | | | extending IAttributeVector API.
* | Follow up comments on readability by consitent naming and compare.Henning Baldersheim2017-03-144-15/+21
| |
* | Increase number of allowed files to 64k.Henning Baldersheim2017-03-133-14/+68
|/ | | | | Sacrifice accuracy of size to 64 bytes. This only affect memory representation.
* Merge pull request #1992 from ↵Henning Baldersheim2017-03-1315-33/+64
|\ | | | | | | | | yahoo/vekterli/clean-compile-with-winline-and-fno-omit-frame-pointer Clean compile with -Winline and -fno-omit-frame-pointer
| * Let sub parser be template arg instead of C function pointerTor Brede Vekterli2017-03-132-32/+32
| |
| * Create explicit, non implicitly inlined function definitionsTor Brede Vekterli2017-03-1313-1/+32
| |
* | Merge pull request #1984 from yahoo/balder/move-to-helper-classHåvard Pettersen2017-03-137-121/+147
|\ \ | | | | | | Move to helper class.
| * | Move to helper class.Henning Baldersheim2017-03-107-121/+147
| |/
* / Easy readability of merge loop.Henning Baldersheim2017-03-101-2/+3
|/
* Enable inlinining of the most frequent and optimized code path.Henning Baldersheim2017-03-092-12/+17
|
* override and final.Henning Baldersheim2017-03-095-38/+23
|
* Merge pull request #1975 from yahoo/geirst/improve-imported-attribute-unit-testGeir Storli2017-03-091-7/+9
|\ | | | | Improve testing of acquireEnumGuard() by checking that guard is taken…
| * Use std::async instead of explicit std::thread.Geir Storli2017-03-091-5/+2
| |
| * Improve testing of acquireEnumGuard() by checking that guard is taken on ↵Geir Storli2017-03-091-7/+12
| | | | | | | | underlying target attribute.
* | Merge pull request #1973 from yahoo/balder/deinline-destructors-2Henning Baldersheim2017-03-09103-681/+986
|\ \ | |/ |/| Balder/deinline destructors 2
| * Deinline destructorsi and use -Winline gcc option.Henning Baldersheim2017-03-0924-322/+364
| |
| * deiniline destructorsHenning Baldersheim2017-03-096-2/+10
| |
| * deiniline destructorsHenning Baldersheim2017-03-098-58/+103
| |
| * Hmm, this was needed to avoid getting unresolved symbol in prod_feature_test ↵Henning Baldersheim2017-03-082-0/+3
| | | | | | | | | | | | on ~VectorBase<string....>. But only in test.... That is very odd.
| * gc includesHenning Baldersheim2017-03-081-7/+6
| |
| * deinline large destructorsHenning Baldersheim2017-03-082-0/+3
| |
| * Implement a default destructor to avoid the automatic inlining of large ↵Henning Baldersheim2017-03-0860-291/+490
| | | | | | | | destructors.
| * Specify destructors explicit to control inlining.Henning Baldersheim2017-03-082-1/+7
| | | | | | | | | | Conflicts: build_settings.cmake
* | Merge pull request #1962 from yahoo/toregge/reduce-enumstore-tree-fanoutHenning Baldersheim2017-03-091-1/+1
|\ \ | |/ |/| Reduce enumstore btree fanout.
| * Reduce enumstore btree fanout.Tor Egge2017-03-081-1/+1
| |
* | Merge pull request #1927 from ↵Henning Baldersheim2017-03-0810-634/+306
|\ \ | | | | | | | | | | | | yahoo/balder/put-typeid-and-bufferr-close-to-reduce-number-of-memory-accesses Balder/put typeid and bufferr close to reduce number of memory accesses
| * | Use a public interface for testing.Henning Baldersheim2017-03-043-24/+18
| | |
| * | Inline very frequently used get method.Henning Baldersheim2017-03-042-38/+22
| | |
| * | - override, code unification.Henning Baldersheim2017-03-044-358/+160
| | | | | | | | | | | | - colocate the buffer and the typeId since they are frequently used together.
| * | Align code and GC some unused code.Henning Baldersheim2017-03-045-49/+31
| | |
| * | Put the 2 small words together at the front. They will then be located ↵Henning Baldersheim2017-03-045-168/+78
| | | | | | | | | | | | | | | | | | closer to the members accessed in the _store too. Remove some (void) signarures.
* | | Merge pull request #1928 from yahoo/balder/optimise-andnotHenning Baldersheim2017-03-0825-146/+379
|\ \ \ | | | | | | | | Balder/optimise andnot
| * | | Test notSelfHenning Baldersheim2017-03-081-5/+27
| | | |
| * | | Use templates and iterator range.Henning Baldersheim2017-03-086-58/+71
| | | |
| * | | Get rid of the dirty use of faked unique_ptr.Henning Baldersheim2017-03-086-24/+40
| | | |
| * | | Add non-dirty orChildren.Henning Baldersheim2017-03-083-8/+29
| | | |
| * | | 2nd try at repairing the magic rebase....Henning Baldersheim2017-03-046-6/+12
| | | |
| * | | andnot_hits_into was more code than worthwile.Henning Baldersheim2017-03-048-98/+1
| | | |
| * | | Implement andnot_hits_intoHenning Baldersheim2017-03-042-0/+8
| | | |
| * | | Use dual not and or instead of special andnot..Henning Baldersheim2017-03-043-3/+12
| | | |
| * | | Optimise includes.Henning Baldersheim2017-03-041-6/+4
| | | |
| * | | Major typoHenning Baldersheim2017-03-042-2/+2
| | | |
| * | | Add andnot test.Henning Baldersheim2017-03-042-0/+38
| | | |
| * | | Must ensure space has been reserved.Henning Baldersheim2017-03-041-0/+4
| | | |