Good place to host auto-generated code?

Question

We have a bunch of auto-generated classes, which are mostly Axis2 misfires, skeletons, etc. For some complex wsdls, Axis2 generates TON java beans, stubs, etc. And I'm sure there are other cases where auto-generation b.

Now we consider them as other members of the first class of our code base and store them in the same packages.

However, when performing refactoring, cleaning, etc. it becomes difficult to get rid of the warnings coming from these automatically generated classes. For example, if I try to clear the code to use the generic Java1.5 tools, there is no good way to find out how many of these intruder classes are ours and not auto-generated.

Should I split these auto-generated parts into another package? How do you guys store these artifacts in the repository?

EDIT: I see “generate during build process” in several answers below. Although I see the benefits of this, I don’t quite understand how I can get away from checking the repository.

My code has a compilation of time dependencies on some of these classes, and for me the build at design time is "ctrl-s" in eclipse. We use ant-scripts to generate compilation, run tests, and generate results.

Answer 1

, (- -src), . . , . , .

Answer 2

:

• • .
  • . ( , WSDL)
• • .
  • .jar, .
  • IDE ( maven)

Answer 3

. , , , , .. .

Answer 4

, . . WSDL, WSDL, .

, .jar , , , .

, , , , .

Answer 5

, , JAR ZIP , . , , IDE.

Answer 6

with maven and axistools-maven-plugin, the generated sources are places in another source folder in the target directory. this destination directory is where maven generates all the files and stuff, so it can be cleaned up.
This is very confusing because the generated files also appear in a different source folder in the IDE.