Firebird Documentation IndexFirebird 3.0.2 Release NotesBugs Fixed → Firebird 3.0.1 Release: Bug Fixes
Firebird Home Firebird Home Prev: Bugs FixedFirebird Documentation IndexUp: Bugs FixedNext: Firebird 3.0.0 Release: Bug Fixes

Firebird 3.0.1 Release: Bug Fixes

Core Engine
Server Crashes
API/Remote Interface
Utilities
Builds
POSIX-Specific

The following bug fixes were reported prior to the v.3.0.1 official release:

Core Engine

(CORE-5355)     XpbBuilder was failing to create a new TPB.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5351)     LEFT JOIN would incorrectly push a UDF into the inner stream, causing wrong results.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5341)     A user-defined collation failed to work with BLOBs.

fixed by A. dos Santos Fernandes

    ~ ~ ~     

(CORE-5339)     An assertion could occur in createDatabase() when doing the overwrite check.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5337)     A subquery in the insert list expression was ignoring changes made earlier in the same executable block.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-5329)     A database could suffer partial corruption in the “use all space” (no-reserve) mode.

fixed by V. Khorsun & D. Yemanov

    ~ ~ ~     

(CORE-5327)     BUGCHECK 165 (cannot find tip page)

fixed by V. Khorsun

    ~ ~ ~     

(CORE-5322)     Cascade deletion in a self-referencing table could raise the error “no current record for fetch operation”.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-5313)     A “Data type unknown” error was occurring with the LIST operation.

fixed by A. dos Santos Fernandes

    ~ ~ ~     

(CORE-5307)     Random “invalid transaction handle” errors were being returned from the engine.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5305)     ON UPDATE CASCADE was failing for a self-referencing Foreign Key.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-5302)     REGRESSION: Performance was degraded when bulk inserting into a table with indices.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-5292)     Trying to encrypt a database in the absence of the the appropriate key could corrupt it.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5289)     A small memory leak could occur when an empty result set was returned.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-5279)     Granting of privileges to views was broken.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5277)     Parameters with multi-byte character sets were being allowed to exceed the length limit for VARCHAR fields.

fixed by A. dos Santos Fernandes

    ~ ~ ~     

(CORE-5275)     Expression index could become inconsistent if CREATE INDEX was interrupted after b-tree creation but before committing.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5268)     Nested OR conditions could lead to incorrest results.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5248)     Consistency between roles and privileges in GRANT syntax needed improving in accordance with the SQL standard.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5244)     An attempt to execute a SELECT statement using isc_dsql_execute_immediate() would produce the wrong error message.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5241)     Affected rows were not being counted for some update operations with views.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5236)     IN/ANY/ALL predicates could cause sub-optimal execution (late filtering) of joins.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5233)     Unaligned memory access in MET_format() would cause segfault on big-endian machines.

fixed by James Clarke

    ~ ~ ~     

(CORE-5232)     32-bit locks on big-endian systems would lead to deadlock.

fixed by James Clarke

    ~ ~ ~     

(CORE-5231)     EXECUTE STATEMENT would throw a BLR error if the number of output parameters exceeded 256.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5226)     An incorrect result set (missing records) could be returned by an ORDER plan query navigating on a descending index.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5225)     Authentication was ending with first plug-in that had the user but would fail there, instead of continuing with the next plug-in.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5222)     SELECT WITH LOCK could raise unexpected update conflict errors under concurrent load.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5213)     A database could get decrypted after changing a couple of bytes in the database header without 'agreement' from the crypt plug-in.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5210)     POST_EVENT was not working with the Fb 3.0 server and client.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-5193)     The operator IS exhibited a precedence problem.

fixed by A. dos Santos Fernandes

    ~ ~ ~     

(CORE-5183)     REGRESSION: line/column numbering in error messages were being garbled if alias.name syntax was used in the offending statement.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5142)     The error “no current record to fetch” was thrown if some record was to be deleted both by the statement itself and by some trigger fired during statement execution.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-1746)     An expression index could be created while inserts into the table were in progress.

fixed by V. Khorsun

    ~ ~ ~     

Server Crashes

(CORE-5234)     Access violation from certain UDFs would cause the server to crash.

fixed by V. Khorsun

    ~ ~ ~     

API/Remote Interface

(CORE-5347)     Race conditions could arise when initializing wire compression.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5296)     A network protocol error was exhibited when performing a callback to a client for the database crypt key.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5278)     Some SPB parameters were failing in FB 3.0.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5224)     Transaction id tags for the Services API did not support new 48-bit transaction ids.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5154)     The Services API would not work with non-ASCII database names.

fixed by D. Sibiryakov

    ~ ~ ~     

Utilities

fbsvcmgr

(CORE-5270)     fbsvcmgr was not producing any error when it attempted to shut down a database without a specified timeout (prp_force_shutdown N).

fixed by A. Peshkov

    ~ ~ ~     

gbak

(CORE-5228)     Restore could hang if the database contained more than 4 billion records.

fixed by D. Yemanov

    ~ ~ ~     

gfix

(CORE-5295)     Validation could read after the end-of-file when operating on a multi-file database.

fixed by V. Khorsun

    ~ ~ ~     

isql

(CORE-5308)     The BLOBVIEW command was causing isql to hang.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5294)     isql would leak memory when SHOW GRANTS was used on a new, empty database.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5220)     isql -[e]X[tract_metadata] output was omitting double quotes for the COLLATE <collation> clause of a CREATE DOMAIN statement when <collation> was for any non-ascii character set.

fixed by A. dos Santos Fernandes

    ~ ~ ~     

(CORE-5218)     isql -[e]X[tract_metadata] was not exporting explicitly defined names for NOT NULL constraints for Dialect 3 databases.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5217)     isql -[e]X[tract_metadata] could crash while exporting an exception with message text length longer than 127 bytes.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5207)     isql -[e]X[tract_metadata] could generate invalid GRANT USAGE statements for domains.

fixed by D. Yemanov

    ~ ~ ~     

Monitoring and Trace

(CORE-5246)     A string truncation error would occur while selecting from the MON$ tables if some user-defined context variable was longer than 255 bytes.

fixed by D. Yemanov

    ~ ~ ~     

(CORE-5203)     A trace session could fully utilize a single CPU core.

fixed by A. Peshkov

    ~ ~ ~     

nBackup

(CORE-5264)     A database located on a raw device could not be unlocked from an nBackup-locked state.

fixed by A. Peshkov

    ~ ~ ~     

Builds

(CORE-5348)     Databases created during the process of the Windows builds were owned by a Windows administrator, not SYSDBA.

fixed by V. Khorsun

    ~ ~ ~     

(CORE-5284)     Firebird failed to build with USE_VALGRIND

fixed by A. Peshkov

    ~ ~ ~     

POSIX-Specific

(CORE-5335)     A segfault could occur when op_que_events was used on a port that was not prepared for events processing.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5285)     A segfault could occur if an attachment was closed while requests were still open.

fixed by A. Peshkov

    ~ ~ ~     

(CORE-5249)     Incorrect use of the system call readlink().

fixed by A. Peshkov

    ~ ~ ~     

Prev: Bugs FixedFirebird Documentation IndexUp: Bugs FixedNext: Firebird 3.0.0 Release: Bug Fixes
Firebird Documentation IndexFirebird 3.0.2 Release NotesBugs Fixed → Firebird 3.0.1 Release: Bug Fixes