1. Bring trace DSQL, DYN and BLR trace logging up-to-date
2. Remove redundant CSB pointer dereferences in BLR parser and request compiler
3. Convert RIGHT JOIN to LEFT JOIN early during BLR parsing
4. Check BLR syntax for unititalized contexts usage
5. Some type-safety and const-correctness fixes
2. Minor cleanup
3. Fixed BREAK/LEAVE code to support FOR EXECUTE STATEMENT loops
4. Added support for loop labels and LEAVE <label> statement
5. Fixed SUBSTRING to support value expressions as its arguments
Exception is that we allow mixing datatypes with (var)char (except blob), because all datatypes are convertable to character-type and because this functionallity was available from 1st day COALESCE is introduced inside Firebird.
2) Fix bug with expressions without fields inside aggregate-functions.
Example :
SELECT
(SELECT Count(1) FROM RDB$DATABASE)
FROM
RDB$RELATIONS
1) RELEASE SAVEPOINT statement is added
2) SAVEPOINT statement now re-establishes existing savepoint with the same name
3) all savepoint stuff is changed to use single BLR verb
2) Enabled aggregate-functions from different parent_context (with lower scope_level) to be used inside another aggregate-function
Example:
SELECT
MAX((SELECT COUNT(*) FROM RDB$RELATIONS))
FROM
RDB$RELATIONS
2. ROWS_AFFECTED is renamed to ROW_COUNT
(see SQL2002 specification)
3. CONNECTION_ID/TRANSACTION_ID are renamed to CURRENT_CONNECTION/CURRENT_TRANSACTION
(see SQL2002 recommendations for context functions)
4. Solved most of the shift/reduce conflicts and all reduce/reduce conflicts in the grammar
5. Fixed ambiguity in ORDER BY clause, when both the value and the clause itself could contain COLLATE word
(now explicit collation sequence of the value is used in the ordering logic)
6. Some of newly introduced tokens are made non-reserved
7. Added code to enable ORDER/GROUP BY ordinal when asterisk is used in the select list
(not sure which version it should go, so it's commented out so far)
