Skip to content

8.3.697

Choose a tag to compare

@github-actions github-actions released this 11 Nov 08:03
· 25 commits to master since this release

commit 8.3.697

"Tagging version 8.3.697"

See documentation at https://github.com/RTimothyEdwards/magic/blob/e400017f/appimage/7/README.md
See documentation at https://github.com/RTimothyEdwards/magic/blob/e400017f/appimage/8/README.md
See documentation at https://github.com/RTimothyEdwards/magic/blob/e400017f/appimage/9/README.md
See documentation at https://github.com/RTimothyEdwards/magic/blob/e400017f/appimage/10/README.md

Change Log (since previous tag):

e400017f Noticed that ResNoMergeMask[] in the resis/ code is never initialized, and while it is a global variable and should always be zero, it is not useful, and so it has been removed everywhere it occurs.
d1e58944 On further reflection, changed the behavior in the last commit such that only two *port* labels of the same name are considered the same node, while multiple *non-port* labels are not, and the name of a sub-net takes the ".n" or ".t" suffix.  This is generally consistent with common usage.
19ed660d Claude Opus 5.5 resolved two issues with R-C extraction resistor network simplification.  For most layouts, having the same label in multiple places generally implies (like it would for a standard cell) multiple points of contact from hierarchy above, any of which would be valid.  That means that they must be treated effectively as a single point.  The code changes force such multiple labels to be treated as a single node, merging them as needed.  Also:  A sink-point routine for hierarchical extraction was causing device nodes to get their type reassigned away from the device type, resulting in the device becoming "lost" to the R-C extraction. This has been fixed.  Also:  An early loop break which has not been observed to cause an error but potentially could was also fixed.
e7c7dd2e I got Claude Opus to track down a segfault reported by Torleif Skaar (github magic issue #573);  this turned out to be an error condition in the "extresist" code that had become exposed by an earlier correction made about a week ago.  In between, I found another unrelated error in the continuous DRC check code that occurs when the same provided reproducible example is run in magic's interactive mode.  Both errors have been fixed.