Parameter passing hygiene

Most design kits are normally CDF-based these days, and so you should not be using NLP expressions but using pPar("a") - that wouldn't suffer from the loop.

The annotation setup can show the evaluated expressions (this is the "Display Type" column I think from memory - you can set it to show "Full" evaluation). That would show the resolved values as you push down the hierarchy.

Regards,

Andrew.

Hi Andrew. Many thanks for your reply. I think in trying to distill my question into it's most basic form I left out some important info. Here's some additional info:

I have defined the following procedure:

procedure( myProc(ilInst)
if(cdfParseFloatString(ilInst~>c))>1 then
sprintf(nil "%g" cdfParseFloatString(ilInst~>c)))
)
)

On the child cell I have set it's c parameter to pPar("c") so that it receives the value of c from above, and on the child symbol I placed an ILLabel:
myProc(ilInst)
I want the child symbol to show the value of the parameter c passed down from above, however myProc isn't receiving the value of c from above, it's receiving the text pPar("c").

How can I fix this to do what I want?

Try using the abIPar function in this other forum post for your label display - that is something I wrote a few years ago to do precisely this.

Regards,

Andrew.

This is works, but this onion has more layers :-). I'll make the passing comment though that I would surely have thought your solution would have been a built-in function - it's such an obvious thing to do! I also noticed that the value that made it's way into the child parameter did not subsequently get processed by the child's callback code. I think now what I want is for the parameter to be set in the parent, it passes down to child, goes through the child's callback and gets modified accordingly, and then that result makes its way back up to the parent i.e. I want the parent to behave just like I'm editing the child.

I could replicate the child's callback details in the cdf of the parent, but when I tried that it didn't work (the function was returning some errors that I can provide if you want). Not sure if the scope of the callbacks in the child (which is in a different library) are valid in my library. Do you have any more magic up your sleeve :-)?

I think I'm at the point where I'm not appreciating the complexity of what I'm asking for...

I think you need to read my article The Dangers of CDF Callbacks. CDF callbacks do not (and cannot) work the way you are expecting them to.

It is also already possible to show the evaluated value of any CDF parameter via the cdsParam() labels using the annotation setup - you can ask it to fully evaluate any expression and show what it would resolve to - so the abIPar isn't really necessary in that case. I just wrote it to provide a convenient means of showing the evaluated label if that was what you wanted all the time.

Regards,

Andrew.

I finally got around to using the abIPar function you wrote and it pretty much solved all of my problems! Many thanks for your assistance and I'm really happy with the way the cells that use it turned out.

Hi Andrew. I spoke too soon about abIPar fixing ALL of my problems. If I set my parent parameter to a number then abIPar nicely shows the final value on the symbol instead of the string pPar("whatever"). However, if I set the parent parameter to a string (e.g. "myVar") then abIPar just returns nil (instead of just the original string) and my symbol doesn't show anything useful. I tried to be daring and edit abIPar to cater for it receiving a non-number, but it's written in Lisp and I just couldn't get deep enough into it before drowning in my own vomit :-). Is there any chance you could tweak abIPar for this condition, or maybe even convert it to skill and I'll take over from there. Sorry about this hassle...

I originally wrote it over 14 years ago, so I'd have to construct some test data to test it again in this situation.

My guess is that you'd need to change the line:

(setq evaluated (aelNumber (aelEnvExecute compiled)))

to:

(setq evaluated  (aelEnvExecute compiled))

Note that the code is written in SKILL. It's just in LISP syntax rather than the C-style. You can always use pp(abIPar) to print out in C-style but the formatting isn't pretty (a shame for a pretty printer) and also you lose all the comments.

I'll have a look next week (may not have time over the weekend) to check the behaviour with string-valued parameters.

Regards,

Andrew.

Wow - that was quick! Note that there is a weird exception to the string rule in that I might give assign the string parameter "Ibias" (which yields a string in cdfParseFloatString") and on another instance I might assign the string parameter "Ibias*2" (which also yields a string in cdfParseFloatString") or but I could have used the string parameter "2*Ibias" (which yields a nil in cdfParseFloatString"). In similar code I've written, I just check to see if cdfParseFloatString is a number (via the numberp function) and if it's not then I just interpret it as a string.

So is this a string or the name of a design variable that you set in ADE? The code gets design variable values from the top cell (if my memory is correct) so you may just neee to do Variables->Copy to cellview in ADE to makthe this work. 

This is one of those things where it would be better if I can see the actual data that isn’t working. If the above doesn’t fix it can you log a case on support.cadence.com and ask the AE to contact me?

Andrew