RFR: 8151542: URL resources for multi-release jar files have a #runtime fragment appended to them

[Chris Hegarty]

> On 1 May 2016, at 08:45, Alan Bateman <Alan.Bateman at oracle.com> wrote:
> 
> On 30/04/2016 01:10, Steve Drach wrote:
>> Hopefully the last one ;-)  This webrev removes the lowercase of protocol, and incorporates better (in my mind) seperation of choices for choosing the loader, similar to what Paul suggested.  Everything else remains the same.  Only URLClassPath changed from previous webrev.
>> 
>> http://cr.openjdk.java.net/~sdrach/8151542/webrev.04/index.html <http://cr.openjdk.java.net/%7Esdrach/8151542/webrev.04/index.html>
>> 
> This version looks okay but a comment to explain why the protocol is checked without regard to case would be useful. Also would it be possible to fix the overly long lines (L377-378) as that is very inconsistent with the existing code and annoying when looking at side-by-side diffs.


+1

Also, it may be worth considering an overloaded JarLoader constructor
that accepts a “jar:” URL, rather than stripping and wrapping.

Also, some jar URLs will now go through JarLoader that previously didn’t.
I’m not saying that this wrong, just trying to understand the behavioural 
change. For example, some URLs that previously didn’t have their jar index
checked now will.

I’ve been looking at this code, for another reason, sorry you have had to
deal with it, it’s a bit of a mess.  We may need to revisit some of this when
8155770 is resolved.

-Chris.