A new variant of inter-task memory exchange implemented for communication between the kernel, clients and network utilities. Support of several new compilers added, including the new versions of Visual Studio (C, C++, C#) and PHP. New selection options added to the distribution (new network utilities and many other selection parameters).
UNIXх64 Systems
Library references can now be created in the /usr/lib64 directory.
QNX
Timer deinitialization in case of client deinitialization organized.
INTROS-VM
The situation with insufficient size of 64k for an object during lindesk build eliminated.
Itanium
DBMS Linter is being ported to the Itanium platform.
DBMS Kernel and SQL Translator:
A numeric sequence can now be generated by an SQL query (syntax: SELECT LEVEL FROM (SELECT 1) CONNECT BY LEVEL < value; ).
Hierarchic query processing optimized for many parent elements having only one child element each.
The REBUILD command can now rebuild only the indexes without rebuilding the converter. Syntax: REBUILD TABLE table_name INDEXES ONLY.
A command added to locally correct an index page in which problems were detected. Syntax: CORRECT INDEX index_name ON table_name FOR [FILE file_number] PAGE page_number.
Using NCHAR and NCHAR VARYING arguments for FINDBLOB and COUNTBLOB functions allowed.
Transaction work in the Optimistic mode optimized.
Codepage description correctness checks (on codepage creation) added.
Query parameter values can now be entered into LINTER.LOG.
A NaN value can be returned as a result of any scalar operation with floating-point numbers if at least one of the arguments is a NaN value.
The work of the GREATEST and LEAST functions for string-type arguments corrected (varying-length type arguments allowed; CHAR/VARCHAR arguments cannot be mixed in order to avoid incorrect processing).
Utilities:
.NET
Database conversion utility added.
lindeskx
DEFAULT value output during schema uploading for the BOOLEAN type improved.
Cyclic channel opening when pressing F8 eliminated.
testdb
Incorrect warnings about not supporting long files (for tables containing such files) eliminated.
It is now possible to detect the existence of several rows in a simple index for the same rowid at the -i 2 level (in the previous version, only level -i 3 was possible).
Complete index check (-i 3) for Version 6.1 optimized.
inl
The time of parameter values entry by the user excluded from the query processing time calculation.
The output of result column names starting with spaces enhanced.
migration
Single-file saving support added.
Qt3Support is no longer used for Qt4 in all utilities based on the Qt interface.
Interfaces:
Call
Thread stack size can now be set by default or from a user program (by calling the inter_control function with the ICR_SET_STACK_SIZE operation code).
BOOLEAN parameter values can now be entered in the lower case (t, f, true, false).
LinAPI
A new field (OPEN_DESC structure) has been added to the t_Connection structure to get the database description in the asynchronous mode in the AstConnect function.
LINCPPAPI
If a connection is broken, it cannot be re-established for asynchronous commands, and the saved queries buffer cannot be run. Any asynchronous calls of LINCPPAPI functions are now fully supported by the asynchronous LINAPI calls. When the saved queries buffer is run, all the queries are executed synchronously. Three tests for the LINCPPAPI library asynchronous mode have been added to the examples.
Error code generation in input-output operations added.
ADO.NET
Entity Framework support added.
DDEX compatibility corrected.
Batch work added.
ODBC
ODBC 3.8 support added.
Working with non-Unicode drivers from MS Access improved.
Argument correctness checks added for some functions.
PHP
Example syntax corrected, i.e. all PHP code prefixes “
MT version build for UNIX platforms corrected.
Problems Solved:
with possible omission of phrase index generation after restoring the database from the lhb archive;
with possible data access conflicts between two channels in the same transaction;
with omission of the action after a channel adding a row waits;
with slow deletion of large indexes due to displacement of pages from the pool;
with the insufficient size of the buffer for PDF file indexing;
with TEST TABLE command looping if an error is generated before table check;
with writing BLOB values for a certain size of source data;
with processing NCHAR arguments in the XML function;
with processing EXISTS and NOT EXISTS predicates in the CASE construction;
with the LOGIN ERROR TIMEOUT functionality;
with the execution of PARTITION BY for the LEAD and LAG functions;
with executing UNION/INTERSECT/EXCEPT and CASE operations with string-type operands one of which is in the UTF-8 encoding and the other is not;
with the TABLESAMPLE construction for Versions 5.9 and 6.0;
with the LIMIT N, -1 selection limitation use;
with insertion to a table containing the GENERATED AS IDENTITY column;
with incorrect calculation of a DEFAULT value address;
with incorrect encoding for the USER pseudocolumn value;
with incorrect results when joining a column with itself;
with the operation of the MULTIME function if the long tick value overfills.
Documentation:
Updated versions of all DBMS Linter documents published on the Web site.
