Login  |  Register
E-mail:
Password:
Forgot your password?
E-Mail:

Authorization

DBMS Linter Modifications List of 21.07.10

21.07.2010

Platforms:

  • RTOS-32 support added.
  • Problems associated with the operation of system calls on QNX eliminated.

Interfaces:

ADO.NET:

  • .NET 4 support added.

OleDB:

  • Problems with reading the values of character fields over 255 bytes in length and BLOB fields via ADO eliminated.

DBExpress:

  • DBExpress 2 support added.

ODBC:

  • Allocation of memory for converting long queries added.
  • Operations with long responses corrected.

JDBC:

  • Synchronization problems eliminated.

LinCPP:

  • Conversion functions corrected.

LinQT:

  • Several new examples added.

Stored procedure language:

  • An example with a pre-translated query added.

RELAPI library:

  • Combo-box functionality corrected.
  • Getting font metric corrected to avoid the situation when the text in the column is not fully displayed (when auto-setting the field width).

STIL library:

  • UTF-8 support enhanced (new function for hot keys added).

COMMON library:

  • Functions for working with multilingual messages added.

Utilities:

lindeskx:

  • Query browser updated and corrected.
  • Looping in local message processing cycles corrected.
  • One nesting level limitation for asynchronous queries introduced.
  • Query halt button added to the advanced query editor; instead of the cancel button, connection break error processing button added.
  • Problems with export/import of NULL values to BLOB fields eliminated.
  • Apostrophe processing problems eliminated.
  • CSV support added to export Excel tables to (export and import support, as well as correct processing of delimiters that are part of data rows).
  • A great number of minor errors corrected.

linadm:

  • Automated startup of the replication server during the automated startup of the database corrected.
  • Replication server management corrected.
  • Replication startup for default database corrected.

lhb:

  • Archiving and saving error processing corrected. In particular, kernel connection and open files are now closed correctly.
  • Information required for processing the log section beginning added to lhb archives.

inl:

  • Name and password buffer overfill problems corrected.
  • Behavior in case of discovering tab after command completion (semicolon) corrected.
  • Multilingual support corrected.

testdb:

  • Problem data page number output corrected.

gendb:

  • Incorrect work of the utility with Russian directory names in some compilers corrected.

Client network driver (dbc_tcp):

  • Problem with broken connection (Error 1069) after several days due to overfilling of watchdog corrected.

dbf2lin:

  • Message output mechanism redesigned, interface language selection support depending on the current locale added.

lin2dbf:

  • Message output mechanism redesigned, interface language selection support depending on the current locale added.

schema:

  • Message output mechanism redesigned, multilingual support added.

EWC (ERWin converter):

  • Message output mechanism redesigned, multilingual support added.

spc:

  • Message output mechanism redesigned, multilingual support added.

crmsg:

  • C message file generation problems corrected.

DBMS kernel and SQL translator:

  • According to the standard, REFERENCES privilege is now necessary for creating foreign keys (instead of ALTER that was formerly used). New BACKUP privilege enabling the user to perform hot backup copying of the database or its objects added.
  • Output of information on system queue size during kernel startup extended and corrected.
  • DBMS kernel information structures extended to use large-size buffers for inter-task exchange.
  • New strategy for working with aggregative indexes in case the number of rows in start sets is largely different (scanning the index of the smaller table and search in the index of the bigger table for each selected value). Calculated results of one-variable predicates (and results of two-variable predicates) are taken into account.
  • The number of modified pages of temporary tables/work files in the kernel page pool limited to optimize work with the pool.
  • Capabilities of working with the local time extended. Extended format of the "time zone" argument in TO_GMTIME and TO_LOCALTIME functions allowed. As a result, functions of converting time from Greenwich to local and vice versa (plus daylight saving time) are implemented. Creation of queries depending on daylight saving time simplified.
  • Algorithm for selecting the default phrase filter for BLOB columns changed. As a result, BLOB data in encodings other than CP866, CP1251 and KOI8-R can be indexed.
  • User message functionality for AUDIT added (one-time writing of a user message with predefined text and marking all messages from the current connection with the predefined text in the USERTEXT field).
  • Expressions, including non-constant, in the DEFAULT construction, allowed. «GENERATED BY DEFAULT (expression)» construction allowed. For GENERATED BY DEFAULT columns and columns with DEFAULT expressions, the same capability of deleting and modifying these expressions as for constant DEFAULT values implemented.
  • Pre-translated query integrity check before execution by the kernel added (in addition to the same check performed by the front end when binding the parameters.

Elimination of problems:

  • Logging problems and problems of the kernel during backup copying:
    • working with BLOB using parallel cursors in the READ COMMITTED mode;
    • processing AUTOINC values during warm restart;
    • missing system table element descriptor marks;
    • mismatch of the number of versions in the table descriptor and the data file in the backup database;
    • incorrect column number counter in an aggregative primary key;
    • accumulation of log files during hot backup copying after executing the operations in the system channel;
    • incorrect actions in case of mismatch during control sum check;
    • disparity of queues and database files in case of deleting temporary tables.
  • Query processing problems:
    • using sub-queries in CASE constructions;
    • calculating full text search predicates in hierarchic queries;
    • operation of the index merging strategy in case there are long rows;
    • calculating IS [NOT] NULL predicates by aggregative index;
    • delay in using the response to one of the start sets in one-index strategy that in some cases led to significant increase in intermediate results;
    • calculating complex queries with express JOIN and external links;
    • calculating MAX aggregative function by aggregative index.
  • Problems with incorrect work with rows longer than 4K and long queries:
    • executing queries with grouping and result row over 4K.
  • Extra uniqueness checks when creating primary and unique keys eliminated, which accelerated the creation of such keys several times.
  • Problems with encoding when getting column descriptions eliminated.
  • Work with device cache corrected.
  • Work with column queues when addressing large tables at random optimized.
  • Batch loading takes into account the specific of "cyclic" tables.
  • Comparison of real variables when working with indexes corrected: now this comparison is equally accurate as in all other cases.
  • Channel number correctness and relevance checks added in Linter Multiversion.
  • Execution of queries with NOWAIT modifier in Linter Multiversion corrected (transitions to standby mode eliminated).
  • Calculation of the maximum permitted row length in Linter Multiversion corrected in order to avoid problems with row unpacking.
  • Several problems with pre-translated query cache eliminated.
  • Several problems with DEFAULT and CHECK constructions corrected.
  • Channel context error corrected.
  • Full index locking in CORRECT INDEX removed (now only one element of the index is locked).
  • No more problems with system queue element damage after unsuccessful search that led to errors in system tables.
Main | Back to the list
ѕрокрутить вверх