maria-developers team mailing list archive

[Sergei Golubchik <serg@xxxxxxxxxxx>]

Hi, Blessing!

On Jun 08, Blessing Atie wrote:
> Hi Serg, Colin!
> 
> The following is a rundown on my progress with the above task.  I
> started this project by getting the source of 10.1 from Github,
> setting my working environment by building and running the tests on
> ubuntu 14 then proceeded with an analysis by studying plugins and how
> they interact with the server.(Especially storage engines plugins).
> 
> With  analysis and sources investigations being done, I realized with
> help from assigned mentor(Serg)  that our task is about implementing
> MYSQL_UDF_PLUGIN with the functionalities of the CONNECT engine which
> is similar in functionality to table functions in a way.

Not with the functionality of the CONNECT engine. I only mentioned
CONNECT as an example of what a table function can do. But those table
functions are too complex for us to do in this GSoC.

If you want a "functionality of engine", look at the Sequence engine.
A table function with the functionality of the Sequence engine - that's
fine for an example.

> This past week, I worked on   the interface for our plugin by declaring a
> structure with version and some useful init/deinit function pointers to
> work with our  UDFs(table UDF only  for now). I am currently studying the
> Connect engine to see how I can implement its functionalities  as a UDF
> plugin.

No, please don't. Study Sequence engine, if you must.

> For this week, I intend to look at other databases for example table
> functions that will be good candidates that we can implement with our
> plugin(This was not mentioned in my proposal but I intend to implement a
> couple of them using our plugin even if we are short of time after the
> final evaluation).

Good, you cannot know if your API is useful, unless you actually try to
use it (by writing few table function examples).

Regards,
Sergei