Re: Adoption of RPG OA (Summary)

Yep, that's clear - and that's what I realized after reading your post - in another part of my reply I think I said that a handler could be a trojan horse - it can do anything a called program can.

So there's a risk, just the same as you describe here. Hey, would either of use let Leif Svalgaard put a handler OR a program on our machines? :-) No offense to Leif - we think he's trustworthy.

I'm glad you raised this issue - it hasn't really been mentioned much, if at all, and it needs to be considered.

Vern

On 3/20/2011 9:06 PM, Joe Pluta wrote:

On 3/20/2011 8:25 PM, Vern Hamberg wrote:

As to side effects - actually, the RPG engine is protecting you pretty
well there, I believe. All the handler does is set some values that the
RPG engine uses.

I don't think I was explaining myself very well. A CHAIN can only do
one thing - read the database. It cannot affect anything else. Without
a trigger or a handler, a CHAIN can never affect anything else. Thus,
if I need to debug a situation where a file is getting updated, I don't
have to worry about any program that only CHAINs to files... until I
start adding handlers. Because now, every program that CHAINs to a file
that has a handler can now potentially be updating anything, because now
it is in effect calling another program. Suddenly my analysis of what
used to be a read-only program now changes to having to check the
handler as well.

That's not necessarily horrible, and I have to do the same thing for
program calls, but at least with a CALL I know it's a CALL! With a
handler, I have to check the F-spec to make sure it doesn't have a
handler attached.

Minor, sure, but one more thing to check. Is that a little clearer?

Joe