path appearing different within IC616 and IC618

Jay,

I don't see any difference in the display between these two releases. Both are a bit odd because it's sort of reentrant, but they look identical to me. I've not heard of any compatibility issues either (I'd be very surprised).

Perhaps you can share screenshots of the difference so it's clearer as to what  you're seeing? Also, which specific subversion of each tool you're using would help (I tried the base IC616 IC6.1.6.101 against the latest hotfix of IC618, IC6.1.8-64b.500.9).

Andrew.

Jay,

I don't see any difference in the display between these two releases. Both are a bit odd because it's sort of reentrant, but they look identical to me. I've not heard of any compatibility issues either (I'd be very surprised).

Perhaps you can share screenshots of the difference so it's clearer as to what  you're seeing? Also, which specific subversion of each tool you're using would help (I tried the base IC616 IC6.1.6.101 against the latest hotfix of IC618, IC6.1.8-64b.500.9).

Andrew.

Hi Andrew,

Thank you for your support.

I use IC06.18.030 and IC06.16.132

This path was created in IC616, however when I opened it in IC618, it got the single point touch as the picture shows.

Jay

It turns out I'd not set the path width as you'd specified, but even then it's OK in the versions I tried.

I checked the version you are using of IC618 (ISR3) and I see the problem with the triangular fillet from IC6.1.8-64b.500.1 until IC6.1.8-64b.500.7 - in the earlier IC618 base release and IC6.1.8-64b.500.8 or IC6.1.8-64b.500.9 it's OK - so it appears we made a temporary error in handling these (in my opinion) malformed paths. Here's what it looks like now (and in the older version you were starting from):

The reason why the stream changes too is because the stream interface converts paths to polygons in certain situations - such as having a segment with a length less than half the path width (it puts a message in the stream log file to tell you this is happening). We do this because interpretation of such reentrant shapes is very likely to be tool-dependent so we try to remove the ambiguity by converting to an unambiguous polygon.

To be honest, if you have shapes like this in your layout, I really would just fix them. It's asking for trouble. I didn't find the precise report of the issue (I found something which might have been it, but such shapes are pretty unlikely to be in any regression suite so it was probably stumbled across by somebody in product validation - if indeed the report I found was the same issue).

Andrew.

Thank you for advice.

Jay