Friday, October 15, 2004

It was a quiet day with TJ away. I noticed AN and KA dealing with some minor crisis that needed TJ's involvement on his day off, but fortunately I was far removed from that.

There were quite a few minor changes to the RemoteResolver interface which needed attention before anything would compile properly, and I spent a lot of the day dealing with these. Many of the changes were straightforward, though time consuming (e.g. some data which was provided in local node ID form is now being provided in global URI form, which means that all references to it have to change). Changes like this take up a lot of the day, but don't leave much to report on.

The only thing worth mentioning is that the RemoteResolverFactory.newResolver method now takes an extra parameter called systemResolver. While it might be handy to have this object in some circumstances, I can't think of how it would be needed for the remote resolver. So after looking at the related code for some time I decided that I wouldn't need the system resolver, and have ignored it. On the other hand, I'm now concerned that I'm missing something important, so I'll need to have a word with SR about it on Monday.

The other thing about this method is that the signature changed from:

  public Resolver newResolver(ResolverSession resolverSession, boolean canWrite);
To the following similar form:
  public Resolver newResolver(boolean canWrite, ResolverSession resolverSession, Resolver systemResolver);
I've already mentioned the systemResolver parameter, but I'm a little lost as to why canWrite was moved, though it hardly matters. However, one thing that does matter is that the javadoc for this method in the interface was not updated... something I should really chastise AM about.

No comments: