another off-grid problem (pathSeg cannot be rounded to grid)

More about this, the script actually does something really unexpected and bad IMO, it sets the pathSeg extension to zero, so that a wire pathSeg that starts out like this, note the extension of 3.4um

turns into this

Note extension is now zero

We don't know what do do about this. 

Would it be possible to suggest a fix to this part of code so that it modifies the existing pathSeg to make sure all properties are preserved?

here is the code in text

("pathSeg"

pathSegId=dbCreatePathSeg(cv obj~>lpp car(path) car(last(path)) obj~>width obj~>beginStyle obj~>endStyle)

if(obj~>figGroup then dbAddFigToFigGroup(obj~>figGroup pathSegId))

geSelectObject(pathSegId)

dbDeleteObject(obj)

) ;pathSeg

Probably it would be enough if the script would round the pathSeg extension values to the grid as well as the vertices

I have also encountered problems in fixing "pathSeg", could you please help me upgrade the code?

I would suggest that you either create a case via support.cadence.com or give feedback on the article you reference; that's the best way of it getting back to the article author (it's not me...)

I would also wonder why you are getting the off-grid points in the first place; that's very odd, given that pathSegs are constrained to be octilinear. You might want to raise a case for that alone.

Andrew

Well like I said, the tens of thousands of offgrid errors resulted from a bogus offgrid placement of wiring done in a temporary copy of the top layout. When this wiring cell got placed back to the top layout it was offgrid, which should not have been allowed to occur. We did not notice and wired up lots more stuff. Now we are faced with the big job of fixing all these errors. A working offgrid fixer script would be very useful, since the current ones fail on some pathSeg wiring objects and they fail without any reporting of which ones they fail on, leaving us in a strange state of trying to find these objects by divide by two approach. Wrapping this pathSeg code in errset and reporting which objects failed (or somehow marking them) would be useful. 

And there is still the mystery of why we cannot manually edit the pathSeg object shown above to fix it in the properties dialog. 

It would be very nice to have a robust version of the CCSputOnGridHier script with bindkey installation to make life easier. We are not good at this coding but maybe will try it later after the tapeout.

Thanks as always for the quick responses Andrew, you are great!

 BTW, the script CCSputOnGridHier.il did this to one of our buses, shorting out a bunch of signals

I suspect that the code has not been that well-tested with pathSegs (by the looks of things). That's why contacting support makes sense as that way the author can improve it (sorry, I'd love to re-write it, but don't have the bandwidth to do this - I'd sooner the original author improves it themselves).

Thanks,

Andrew

Change is needed at two places in the code. One while storing the values (first foreach loop) and one while recreating (case statement).

1. Storing the beginExt and endExt initially in the foreach loop…

 ("pathSeg"

           path = list(list(obj~>beginExt obj~>endExt)  obj~>endPt  obj~>beginPt)

         )

The order in the list is utilized during recreation of pathSeg, so stick to it.

 2. Using the values during recreation of pathSeg in case statement.

("pathSeg"

             geSelectObject(dbCreatePathSeg(cv obj~>lpp car(path) cadr(path) obj~>width obj~>beginStyle obj~>endStyle caddr(path)))

             dbDeleteObject(obj)

           )

I have sent the modified file to Tobi's mail. Once he confirms that it works for him, I will upload the file in the following article as well.

https://support.cadence.com/apex/ArticleAttachmentPortal?id=a1Od0000000nWVlEAM&pageName=ArticleContent

Thanks for this great responsivity. I have shared back to Raj my edits to the code to make it more robust, so that it doesn't simply abort when it encounters a pathSeg it cannot modify (for unknown reason we had a couple of these). My modified script prints out more information about what it changed,, it only puts things back to grid that are offgrid, and it leaves selected those objects that it cannot fix by catching these exceptions with errset. My SKILL coding skills are basically nonexistent so maybe Raj can make it better.

BTW, I found it very useful to add a default grid setting and make a bindkey, then invoke it with key to run the script with the desired grid, in this case putting these lines to the CIW or in .cdsinit makes it easier to run

load("../skil/CCSputOnGridPins.il")

hiSetBindKey("Layout" "<Key>5" "putOnGridCB(\".005\")")