Any news on the next build?

snorkel · Post by **snorkel** » Thu 12 Feb 2009 17:06

Just wondering if there was any news on the next build?
fixes, features etc?

Plash · Post by **Plash** » Fri 13 Feb 2009 16:45

Now we have the following list of fixes for the next build:

Added UpnpreparedExecute option of TCustomPgDataSet that allows simple execution of statements
Added support for Schemas metadata kind in TPgMetaData component
Bug with not working CachedUpdates in TVirtualTable fixed
Bug with suppressing errors in TVirtualTable.LoadFromFile fixed
Problem with restoring current row after refresh fixed
Fixed bug when statements, executed on TClientDataSet.ApplyUpdates, were not shown in DBMonitor
Fixed access violation on TClientDataSet.ApplyUpdates call
Fixed bug with not executing prepared statement second time
Fixed bug with saving TPgLargeObjectField.BlobType property in DFM
Fixed bug with TPgAlerter started at design-time when Active=False
Fixed memory leak in UniDirectional mode
Fixed bug with detecting read-only fields when ExtendedFieldsInfo=False

snorkel · Post by **snorkel** » Thu 19 Feb 2009 19:45

Plash wrote:Now we have the following list of fixes for the next build:

Added UpnpreparedExecute option of TCustomPgDataSet that allows simple execution of statements

Will the unpreparedexecute be a published property? Also if this is set will I be able two execute two queries like this:
sql.add('select get_refcursor();');
sql.add('fetch all from return_cursor;');
execute;

Also will the enhanced error messages be available in the pgerror exception type? Currently if I use protocol 2 I lose all that.

Plash · Post by **Plash** » Fri 20 Feb 2009 08:58

The UnpreparedExecute option will be published. It allows to add several statements to the SQL property. Error information will be full if you use this option with the 3.0 protocol.

snorkel · Post by **snorkel** » Fri 20 Feb 2009 17:54

Plash wrote:The UnpreparedExecute option will be published. It allows to add several statements to the SQL property. Error information will be full if you use this option with the 3.0 protocol.

Sounds perfect, will the performance be similar to the 2.0 protocol?

Plash · Post by **Plash** » Mon 23 Feb 2009 09:04

No, in our tests with SSH tunnel the performance is worse than with the 2.0 protocol. The speed with the 3.0 protocol and simple execution is between the speed of the prepared execution and the speed of the 2.0 protocol.