Generate concise Java code for tests

[mximp]

When +junit exists in eo code separate mechanism of Java code generation is enabled. It's driven by junit.xsl and rename-junit-inners.xsl. It generates inner Java classes for each inner EO object which leads to long class names during Java compilation.
The only option is to re-design how we generate Java code for tests.
Probably we need to generate normal (not inner) Java class for each object just like it's done for normal objects.

[mximp]

@yegor256 please assisgn

[yegor256]

@mximp I think we can stay with inner objects, but do something with their names. Why not?

[yegor256]

@Graur please, help with this one

[mximp]

@mximp I think we can stay with inner objects, but do something with their names. Why not?

@yegor256 The issues is that it's Java compiler who is generating the names for inner classes: it prepends inner class name with the outer classes. No matter what names we pick for them with certain level of nesting the compiler will produce long enough .class name to fail.
For example the following is the result of 40 nesting classes:

EOorg.EOeolang.EOexamples.EOtestTest.EOω1blah0.EOω2blah1.EOω3blah2.EOω4blah3.EOω5blah4.EOω6blah5.EOω7blah6.EOω8blah7.EOω9blah8.EOω10blah9.EOω11blah10.EOω12blah11.EOω13blah12.EOω14blah13.EOω15blah14.EOω16blah15.EOω17blah16.EOω18blah17.EOω19blah18.EOω20blah19.EOω21blah20.EOω22blah21.EOω23blah22.EOω24blah23.EOω25blah24.EOω26blah25.EOω27blah26.EOω28blah27.EOω29blah28.EOω30blah29.EOω31blah30.EOω32blah31.EOω33blah32.EOω34blah33.EOω35blah34.EOω36blah35.EOω37blah36.EOω38blah37.EOω39blah38.EOω40blah39

So the only solution I see here is to avoid nesting... Or at least nest classes into the same outer not into each other so we have names like: EOorg.EOeolang.EOexamples.EOtestTest.EOω1blah0, EOorg.EOeolang.EOexamples.EOtestTest.EOω2blah1, etc.

[yegor256]

@mximp got it. I like the second option: keep all classes in the same file, but reposition them to avoid nesting.

[yegor256]

@mximp let's do this?

[mximp]

@Graur I believe it was with you.

[levBagryansky]

@maxonfjvipon please assign me

[0pdd]

@mximp 2 puzzles #3266, #3267 are still not solved.