[ https://issues.apache.org/jira/browse/HBASE-18957?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16197179#comment-16197179 ]
Sean Busbey commented on HBASE-18957: ------------------------------------- {quote}For stuff like filters, no harm a comment or method javadoc saying what is expected... Can add on commit.{quote} FilterList with an OR operator seems to be self explanatory already. The current regression is a bug AFAICT. {quote} Then the change in FilterList, is that part of the revert you and Peter are talking of doing? Thats odd, that there are two mentions of a/hbase-client/src/main/java/org/apache/hadoop/hbase/filter/FilterList.java in the patch. Am I reading it wrong? (First there is removal of some breaks, then there is more substantial change) {quote} Any changes to FilterList are a part of reverts. Peter's patch only adds a test (and at that a test that can be used without changing starting in branch-1.1). There are multipleinstances of FilterList because the patch is up to 6 commits depending on the branch (patch w/test plus a commit per thing reverted) so that they're straight-forward to track. e.g. here's what master changes look like on my local repo (that's the branch with the most reverts): {code} * 6abd43fba5 - (HEAD -> HBASE-18957) Revert "HBASE-17678 FilterList with MUST_PASS_ONE may lead to redundant cells returned" (45 minutes ago) * 00816870d5 - Revert "HBASE-17678 FilterList with MUST_PASS_ONE lead to redundancy cells returned - addendum" (45 minutes ago) * 33d40a82be - Revert "HBASE-15410 Utilize the max seek value when all Filters in MUST_PASS_ALL FilterList return SEEK_NEXT_USING_HINT" (45 minutes ago) * 49bbe00877 - Revert "HBASE-18160 Fix incorrect logic in FilterList.filterKeyValue" (45 minutes ago) * e900149a38 - Revert "HBASE-18904 Missing break in NEXT_ROW case of FilterList#mergeReturnCodeForOrOperator()" (45 minutes ago) * 0587c50298 - HBASE-18957 establish filter test that shows current behavior of OR operator on FamilyFilters. (45 minutes ago) * e2cef8aa80 - (origin/master, origin/HEAD) HBASE-18752 Recalculate the TimeRange in flushing snapshot to store file (6 hours ago) * fcdf96a0e8 - HBASE-18958 Removed the IS annotation from SpaceLimitingException (16 hours ago) {code} > add test that establishes branch-1 behavior for filterlist w/OR and revert as > needed to make it pass. > ----------------------------------------------------------------------------------------------------- > > Key: HBASE-18957 > URL: https://issues.apache.org/jira/browse/HBASE-18957 > Project: HBase > Issue Type: Sub-task > Components: Filters > Reporter: Sean Busbey > Assignee: Peter Somogyi > Priority: Critical > Fix For: 3.0.0, 1.4.0, 1.3.2, 1.5.0, 1.2.7, 1.1.13, 2.0.0-alpha-4 > > Attachments: HBASE-18957-branch-1.2.v0.patch, > HBASE-18957-branch-1.2.v1.patch, HBASE-18957-branch-1.4.v1.patch, > HBASE-18957-branch-1.v1.patch, HBASE-18957-master.v1.patch > > > we need a test that shows the expected behavior for filter lists that rely on > OR prior to our filterlist improvements so we have a baseline to show > compatibility (and/or document incompatibilities that end up being > introduced). > Specifically (paraphrased from HBASE-18368 description): Using 2 > FamilyFilters in a FilterList using MUST_PASS_ONE operator should return > results that match either of the FamilyFilters. -- This message was sent by Atlassian JIRA (v6.4.14#64029)