(CORE-4713)     A “BLOB not found” error would occur on rollback after inserting into a table with an expression index.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-4700)     There was a mistake in testing for shadows when executing GDS_DROP_DATABASE. Fix has been backported from Firebird 3.0.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-4690)     DISTINCT vs non-DISTINCT could affect the result of a subquery involving the IN() criterium.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4678)     A “BLOB not found” error could be thrown (wrongly) in uncommitted transactions on Classic or SuperClassic.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4673)     Computed indexes based on a computed column were storing NULL for all keys.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4670)     A constraint violation error could be swallowed in some cases.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4644)     Open database errors could occur under heavy concurrent load.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-4631)     When the shared memory region backing the lock table could not be remapped, the status vector would return the error message “Lock manager out of room” with no low-level details (e.g. OS level error), obscuring the actual cause of the problem.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4627)     A deadlock could occur between the TIP page lock and the monitoring lock when an ON CONNECT trigger accessed the monitoring tables.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-4618)     Rollback would fail to undo changes when a MERGE statement updated the same target rows multiple times and PLAN MERGE was used.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4615)     The Classic server could hang with (near) 100% CPU load, caused by looping of an AST process under some conditions.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-4599)     The REPLACE() function was working incorrectly with multi-byte character sets.

fixed by A. dos Santos Fernandes

    ~ ~ ~     

(CORE-4530)     DB_KEY based join of two tables could suffer in performance in a way that was traced to the order of the expressions in the statement.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4475)     After fb_shutdown(), end-of-process cleanup should not do anything. A situation was reported where it seemed possible to create a file during cleanup. This was corrected.

fixed by A. Person

    ~ ~ ~     

(CORE-4384)     Problems would occur when a table grew beyond 65535 pointer pages.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4383)     Garbage collection on indexes and BLOBs was not working for update_in_place().

fixed by D. Yemanov, D. Sibiryakov

    ~ ~ ~     

(CORE-4382)     User savepoints were not being released on commit.

fixed by D. Sibiryakov

    ~ ~ ~     

(CORE-4616)     The server could crash while the same context variable was being accessed concurrently by separate threads.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4588)     Superserver would die when client was disconnected abnormally during an index navigational scan.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4500)     Firebird was crashing after an unsuccessful remapping of the lock table's shared memory.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4075)     Server bugchecks or crashes would occur when an exception was raised in a calculated index.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-4604)     During EXECUTE STATEMENT, the char_length() size of a UTF8 varchar would increase unexpectedly.

fixed by A. dos Santos Fernandes

    ~ ~ ~     

(CORE-4566)     The output parameter or argument would have an incorrect size when an executable block, a procedure or a function used a system field in the metadata character set.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-4624)     Firebird was mishandling colon characters in POSIX mount table entries.

fixed by A. Peshkov

    ~ ~ ~