Re: [interchange] Correct logic of DisplayErrors

On Fri, 8 Sep 2017, David Christensen wrote:

> I personally fail to see the usefulness of this division as documented,

That's straightforward.

Interchange's global/catalog distinction was grounded in the idea that the 
whole daemon would have certain important configuration made by an expert, 
and individual catalogs would be done by developers who were perhaps less 
expert in matters of security and stability.

Thus the ability to reload a catalog without restarting the whole 
Interchange daemon.

If DisplayErrors is off globally, that means the admin (the global 
overlord) doesn't want any catalogs to have it on.

The reason you didn't get it, and I don't mind if it changes, is that I 
think almost nobody runs Interchange this way anymore.

Most installations I'm aware of (and I'm not just talking about End 
Point's, but also those we've worked on that were set up by others) have 
one Interchange daemon per company/business and one catalog for each site.

The global and catalog configurations and code are almost always managed 
by the same people.

> plus the code itself had been inconsistent when it came to how it 
> applied the DisplayErrors criteria (some call sites checked either/or 
> and some were both/and).

Sure, there were bugs one way or the other.

> Whatever the decision ends up being should be made consistent, plus 
> there are probably additional things that can/should be made to respect 
> this directive as well.

Right-o.

I'd be interested to hear what other users think.

Jon