Should merge suppress errors

Richard Gaskin ambassador at fourthworld.com
Wed Sep 21 18:03:31 EDT 2016


Monte Goulding wrote:

 >> On 22 Sep 2016, at 6:51 AM, Richard Gaskin wrote:
 >>
 >> Interesting.  If merge is indeed thwarting normal error handling
 >> then that may be related.
 >
 > Has merge ever thrown errors on invalid expressions? It seems to be
 > deliberately implemented so it doesn’t and I don’t recall it ever
 > doing otherwise.

I'd have to go back among versions farther than I currently have 
installed to be sure, but at least as far back as v6.7 merge doesn't 
report errors.

This may or may not be a bug per se, which is why I initially posed it 
as a question.  If there's a benefit to having code execution encounter 
an error but provide no notification I'd be happy to learn what it is so 
I could enjoy it.

But then when I tried the "try" variant, it's now off into the weeds of 
unpredictability, at least for my modest mind.  Perhaps the behavior 
patterns make sense to someone smarter, and hopefully they'll be able to 
explain the logic behind what we're seeing.

Or it may just be a bug that's been there so long that no one noticed, 
everyone's just accepted that merge works like it does and we stopped 
asking for anything more.

Heck, I've used merge long enough you'd think I would have found this 
mystifying earlier too. :)

So maybe it's not a bug, or maybe it wasn't a bug that became a bug when 
"try" was added.

Either way, seems worth exploring, either refining the behavior of the 
token's documentation.


Extra bonus points (though I recognize this is quite unrelated): it 
would be super-cool if merge worked more like the implicit merge of LC 
Server, in which we could have conditional blocks in non-contiguous 
portions of the text.

-- 
  Richard Gaskin
  Fourth World Systems
  Software Design and Development for the Desktop, Mobile, and the Web
  ____________________________________________________________________
  Ambassador at FourthWorld.com                http://www.FourthWorld.com




More information about the use-livecode mailing list