September - December 2016

September 2016

  • Found / reported bugs: CORE-5353
  • Created / modified QA tests for ~20 tracker tickets
  • Added test for new feature in 4.0: EXCEPTION and ERROR_MESSAGE context variables
  • Prepared standard Oracle SCOTT schema for using in FB tests
  • Collected stacktraces for crashes in FB 2.5 which occur on checking user rights for trigger while it is modified in another connection
  • Benchmark for encrypted / non-encrypted database: encryption effects performance by ~4..8%
  • OLTP-EMUL: misc changes
October 2016
  • Found / reported bugs: CORE-5364, CORE-5385
  • Created / modified QA tests for ~25 tracker tickets
  • OLTP-EMUL: added checking for FB crash or connection trouble inside block which does initial data populaton
  • Found crash after changing DB-level trigger state form inactive to active
  • Continued collecting stacktraces for FB 2.5.7 crashes
  • Started creation of prototype for integral test for SQL SECURITY clause
  • Found performance regressions in FB 4.0
  • Continued collect samples for integral test of CORE-3362 (cursor stability)
  • OLTP-EMUL: attempt to run test on special FB build provided by Slavomir Skopalik. Failed with dozen of crashes.
November 2016
  • Found / reported bugs: CORE-5408
  • Created / modified QA tests for ~30 tracker tickets 
  • OLTP-EMUL: changed scripts to be able to use this test with enabled replication
  • Found difference in execution plans while making test for CORE-3450
  • Started implementing test as attempt to reproduce crash when trigger source is changed while other attachments make DML on table(s)
  • OLTP-EMUL: more benchmarks of FB 2.5.7 vs 3.0.0 vs 3.0.1
December 2016
  • Found / reported bugs: CORE-5413, CORE-5418, CORE-5423, CORE-5424, CORE-5426
  • Created / modified QA tests for ~40 tracker tickets 
  • Found poor performance of SIMILAR TO comparing with LIKE on trivial search of line containing at least one character from [a-z]
  • Upgraded fbtest to its last version on the Windows testing VM
  • Found that FB service often cannot be correctly stopped after bugcheck
  • OLTP-EMUL: misc changes

Pavel Zotov

Moscow, Russian Federation