finish deprecation of spi.PersistenceUnitTransactionType

[lukasj]

@gavinking we've missed part in PersistenceUnitInfo; the name of the new method is long, but do you have better idea?

fixes #477

[gavinking]

@lukasj Well, actually I didn't miss it, I deliberately left it out with the plan in mind that we would simply change the return type of the older method in JPA 4.0. That seemed like a better solution than introducing an ugly-named method that would hang around after the original was removed. WDYT?

[lukasj]

I'm thinking about possible transition path and I see two ways it "can" be offered.

Let's assume that everything marked for removal will get removed and v4 contains many incompatible changes

• one change for which one cannot prepare beforehand more or less doesn't matter, this one is only about changing import (ie sed can take care of it) vs.
• opt-in to move now and possibly don't care in the future, just delete ~3 lines when it comes

I'm inclining to think 2nd approach is better for most, even though I personally don't use wildcard imports and so I can use the sed way.

In any case, the intent should be clearly communicated - be it this or that. Do you prefer me to update the javadoc instead?

[gavinking]

the intent should be clearly communicated

Yes, that's fair.

Do you prefer me to update the javadoc instead?

That's my (weak) preference, but only because I didn't want to be left with the longer, uglier method name after the migration.

But honestly I don't care much. This is an SPI anyway, it's not in the face of the end user.

[lukasj]

who could we check this with.. @scottmarlow do you have any preference here (from the AS vendor POV)?

[lukasj]

add a comment instead