Issue 652 - Parametrized include

[kmoco2am]

This code changes implements static placeholders to be able to parametrize SQL fragments during parsing phase.

[harawata]

Thank you @kmoco2am !

Looking at the test case, I think this is achievable using mybatis-velocity.
Please see my comment on another pull request.
The sample directive that I posted on Gist is a simple column list enhancement, but you should be able to modify it to suit your needs.

What do you think?

Regards,
Iwao

[lrfsat]

From my point of view: parametrized include is more understandable, flexible and predictable than columnalias directive, or something based on columnalias.

[harawata]

columnalias is just an example.
My point is that velocity custom directives is very flexible, powerful and less intrusive.

I wrote a new custom directive substitute and post it on Gist.
As an usage example, I rewrote the test case in this pull request using the directive.

p.s.
MyBatis already has ${var} expression, so I used @var@ instead.

[kmoco2am]

Thank you for pointing me to the velocity integration project - I did not know about that before.

I think that Velocity can solve almost any issue because it is full templating engine. I just don't need all this functionality. Parametrized include is something missing in the whole mechanism of MyBatis statement construction by XML. Generally, I would be able to replace all , and others by Velocity. However, I don't want to add another syntax to our project just to be able to have this functionality (combining XML include and Velocity syntax). Some of the examples using Velocity for parametrized include are more complex (from syntax perspective), may be too much.

[harawata]

We have discussed this addition on the dev list and decided it's a good addition to MyBatis :-)

Could you please add the following changes to this PR before merging?

1. Documentation. The source of the doc is src/site/xdoc/sqlmap-xml.xml in xdoc format.
2. Independent test case. Current test case in the PR does not fail even if the name/value of the placeholder is modified, so please create a new test case which actually validates the behavior.

Thank you both for your contribution!

[kmoco2am]

That's great it has been accepted ;-)

Let me add more tests and documentation.

[kmoco2am]

I have updated test cases and also added documentation for this feature.

[harawata]

Thank you very much!
I will look into it as soon as I have time.

[harawata]

1. Remove unused imports and 2) import each class explicitly instead of a package.
   Please see some other classes for reference.

[kmoco2am]

Agree, I have changed IDE configuration to fix that,

[harawata]

Hi, @kmoco2am

It's merged. Thanks again for your contribution!
Note that it might take a while until the DTD file to be updated.

I also added some tests that verify the expected behavior of this feature.
Please let me know if I am misunderstanding something.

Regards,
Iwao