updated refactor patch (Jira 388)

[jg...@simulexinc.com]

Before the discussion ended, the consensus seemed to be affirmative that Java 5+ would be used at least server-side.   

Without getting into the various forks of that discussion, is something like what I posted for 388 valuable?   Should I continue on to other packages (focused primarily on outrigger for this round)?

As a reminder, the bulk of the refactor was in correcting (not all, but some) compiler warnings and updating the classes to use Java 5 idioms (generics, for-each, etc.).

The updates to the patch that I made were as follows:
1) Added suppressed warnings in points where the "unused" method was used by Serialization or reflection.
2) Fixed error in initial patch (had nixed readNoData, now it's back; screwed up on context, because it was a Serialization-read, not a space-read). 
3) Added WeakHashSet class.   The majority of WeakHashMap usage in Outrigger used only the keyset.   The WeakHashSet class is a small wrapper of WeakHashMap that clarifies the intent a little (by implementing Set).   I isolated this change somewhat (submitting a patch with and without), in case there were any objections.

jamesG

Re: updated refactor patch (Jira 388)

[Peter Firmstone <ji...@zeus.net.au>]

Thanks James,

Yes I think this effort will be valuable, I'm currently putting time I 
have into the modular build effort (River-300) and source rename, I'll 
add your River-388 patches after.

Cheers,

Peter.

jgrahn@simulexinc.com wrote:
> Before the discussion ended, the consensus seemed to be affirmative that Java 5+ would be used at least server-side.   
>
> Without getting into the various forks of that discussion, is something like what I posted for 388 valuable?   Should I continue on to other packages (focused primarily on outrigger for this round)?
>
> As a reminder, the bulk of the refactor was in correcting (not all, but some) compiler warnings and updating the classes to use Java 5 idioms (generics, for-each, etc.).
>
> The updates to the patch that I made were as follows:
> 1) Added suppressed warnings in points where the "unused" method was used by Serialization or reflection.
> 2) Fixed error in initial patch (had nixed readNoData, now it's back; screwed up on context, because it was a Serialization-read, not a space-read). 
> 3) Added WeakHashSet class.   The majority of WeakHashMap usage in Outrigger used only the keyset.   The WeakHashSet class is a small wrapper of WeakHashMap that clarifies the intent a little (by implementing Set).   I isolated this change somewhat (submitting a patch with and without), in case there were any objections.
>
> jamesG
>
>
>