The most recent Stable (GA) release of MariaDB Connector/ODBC is:MariaDB Connector/ODBC 3.2.5
DownloadRelease NotesChangelogAbout MariaDB Connector/ODBC
Release date: 2 Aug 2018
For the highlights of this release, see the release notes.
The revision number links will take you to the revision's page on GitHub. On GitHub you can view more details of the revision and view diffs of the code modified in that revision.
Revision #baf2287 2018-07-31 23:23:29 +0200
Fix of memory leaks in the driver. Query structure delete routine was called in the wrong place for SQL_DROP, and thus wasn't invoked for single statement queries. Other leak would only occur in case of multistatement query preparing, and batches were not allowed for the connection. One more leak would occur if a batch of statements was optimized to be executed via text protocol. Also the fix of the memory leak in tests. Not really a problem, but created noise in valgrind output.
2018-07-31 13:39:02 +0200
Update of libmariabd module to v3.0.6
2018-07-27 17:35:33 +0200
The fix and the testcase. In case if the query is a multistatement, connector splitted it into individual statements, prepare one by one, and then execute one by one. In case if one of statements depends on execution of one of previous statements, application can get error(think of tmp table creation with select from it) or incorrect results. We can't do anything in case of SQLPrepare + SQLExecute. The patch fixes that for SQLExecDirect. Now connector prepares and immediately executes each statement in the batch.
2018-07-24 18:22:34 +0200
Wrong display size for decimal fields.
2018-07-23 15:17:00 +0200
Not including libmariadb into source package
2018-07-19 12:49:41 +0200
FileDSN creation - fix and testcase
2018-07-17 14:31:11 +0200
The fix and the testcase. Connector does some query parsing, in particular to know the type of the query. Mostly it is interested if query returns result. It didn't evaluate WITH correctly
2018-07-13 01:02:18 +0200
The fix and the testcase. Those aggregate functions, that caused error, returned LONGLONG value, and Access was getting it as SQL_C_LONG. The problem was that connector returned length of the data as 8(field's size), and not 4(requested C-type size), as specs prescribe.
2018-07-11 17:05:28 +0200
The fix and the testcase. Connector did not return length of string returned for SQL_IDENTIFIER_QUOTE_CHAR info type.
2018-07-10 12:50:27 +0200
Changed version to 3.0 in README files.
2018-07-09 12:47:04 +0200
The fix and the testcase. Display and column length did notinclude fractional part for (date) time types. Also octet length for those types was length of they string representation, while specs say it should be size of corresponding ODBC data structs.
2018-07-04 18:06:26 +0200
Fix of build with dynamic C/C linking
2018-07-03 01:05:41 +0200
Scale(DecimalDigits) IRD field wasn't set for (date) time/timestamp fields with second fractional part. Thus SQLDescribeCol retulrned 0 in DegitalDigitsPtr
2018-07-02 01:19:59 +0200
The fix and the testcase. Date(time) types had numerous issues with bulk operations. We provided array of buffers, and C/C expects array of pointers to buffers.
2018-06-27 00:51:56 +0200
The fix and the testcase. For fixed length types SQL_DESC_OCTET_LENGTH initialized by type length, rather than by BufferLength parameter value.
2018-06-25 14:11:34 +0200
The fix and extended old testcase. DESC statement caused error with the connector.
2018-06-18 14:36:35 +0200
the testcase and some amendments to the fix. Used OctetLengthPtr, since IndicatorPtr may be different from length ptr Version bump -> 3.0.6
2018-06-14 21:33:01 -0400
The field length is not set for DATE, TIME, or DATETIME values assigned during SQLFetch calls. Without the length set, Crystal Reports interprets all DATE, TIME, and DATETIME values as NULL. Setting *ArdRecord->IndicatorPtr to the size of the appropriate data structure in MADB_CopyMadbTimestamp.
This page is: Copyright © 2025 MariaDB. All rights reserved.