review request for 7005608: diamond conversion of JCA and crypto providers

[Brad Wetmore]

>> Minor nit, can you add a space in line 221 between
>> "BigInteger,BlindingParameters"
>
> [and Max replied]
>> I thought the formal style is no-space. At least in Map.java it's
>> public interface Map<K,V>.
>
> I'm not familiar with any formal style. The original Java code
> conventions doc [1] hasn't been updated since 1999 and so doesn't cover
> generics. I didn't see a more recent version.

I'm not aware of one either.

> The code base isn't consistent. I found about 3,000 uses of Map<...,...>
> in the JDK, and about 1,800 of those had a space after the comma and
> about 1,200 did not.To my eye it seems that single-letter type
> parameters such as Map<K,V> tend not to have the space whereas type
> parameters that name actual types, e.g. Map<String, Element> tend to
> have the space in there. However, the code base isn't consistent on that
> either.
>
> I'm happy to change the code to <BigInteger, BlindingParameters> though.
>
> [1] http://www.oracle.com/technetwork/java/codeconv-138413.html

Apparently we had the same idea to ask Alex Buckley, which you've 
summarized pretty well above.  I'm ok with either, probably a slight 
preference for the space given the discussion above.

>> Not having a lot of experience yet with <>, the only ones I wasn't
>> sure about
>> were the ones in X509Factor.parseX509orPKCS7Cert. (line 415 & 425) I
>> assume it
>> just picks up the outer return type?
>
> [and Max replied]
>> I read the byte codes generated, and the parameters on lines 415 and
>> 425 are not used. Maybe they are good for code readability.
 >
> Recall that generics are implemented by erasure, so no type parameters
> will end up in the bytecodes in any case.

Yes of course, but this being the first instance of seeing this 
construct, I had to stop and think about it a bit.

I'll have to parse the rest of this discussion when we're back.  I still 
have some support issues to deal with before I can go home tonight.

Brad