Saturday, May 30, 2015

New features in Attracs Bold

It is very easy to believe that Bold for Delphi is dead and the best action is to leave Delphi and just port your existing applications to some other platforms like C#.

I would like to inform that Attracs version of Bold is very actively developed.
Beside all small bug fixes and changes this is some bigger improvements.
Thanks to our skilled consultants that made it possible.

Autosubscription. 
You do not need to call DefaultSubscribe in derived code. Bold handle that automatically resulting in cleaner code.

No hash lookup for datatypes
Internally Bold now call hash lookups for datatypes. This is fast but when done frequently the time will add up. Attracs version use hardcoded datatypes instead as this changes very seldom or never. Together with some other optimization Bold can now handle about twice as many objects per second.

Developer Express grid and editors is boldaware
Not much to say about that. Much nicer components than the original in Bold

Caching of OCL result
Boldaware components evaluate OCL expressions when size of component is changed. For grids this is done for all cells. Using a cache is more effective.

Spanfetch avoid lazyfetch
In Bold it is very easy to trigger a lazy fetch of data. This means that one SQL statement is generated each time data is needed. This can be merged to one final SQL per table that is more effective. So this means a parsing of OCL and calculate the most effective SQL from that. Now there is even support for OCL variables but still a bit unstable. A very powerful and complicated addition to Bold.

ViewModels
When you have a many Bold-handles on a form it is easy to make a mess. Viewmodel is a new component that collect the handles in a nice treeview. It have also some other features so very little code is required on forms. In some cases no code at all. Less code also means less chances for bugs.

LogServer
This is the latest feature. Usually we log to a plain text-file. But that means a lot of information is lost. Because then we need to parse raw text to get information like performance, frequency of events, exceptions etc. Log to a database is better because it is more structured. Data is saved on the proper place.
This also means we can log more as it is easy to find data with SQL and diskspace is cheap. We log all user actions to a database by intercept events also from timers etc. We can also log all modifications of objects. The log can contain time, user, value etc. Hopefully this means much better chance to reproduce those random bugs in live environment.