Problem was caching which did not take into account the array nature
of instances.
This fix also moves the cache one level below so it is effective also
when instance tree traversal happens. This might speed up things too.
Needs testing.
One unit test was failing because the netlist compare did not
properly consider dropped pins:
* A severe bug ("g1" should be "g2")
* Incomplete detection of dropped pins upwards in the hierarchy
The general pin and net mapping scheme has been enhanced so that
net mapping to "0" is valid (this will happen in case of dropped
pins) and this condition is used to detect pins without match
requirement.
The implementation uses extrapolation of strings in the
"Expressions" framework.
There is how:
* $(tech_name) -> substituted by the technology name
* $(tech_dir) -> substituted by the directory the technology file is stored in
* $(tech_file) -> substituted by the absolute path to the tech file
* $(appdata_path) -> substituted by KLayout's home directory (e.g. ~/.klayout)
* $(env('X')) -> substituted by the environment variable $X
The problem is solved by always producing subcircuits for cell
instances, even if there are no connections.
The netlist comparer had to be adjusted too because subcircuits
without pins were used for representing "unknown" subcircuit pairing.
In addition, this patch should lead to a better matching of
parallel subcircuit configurations where two different subcircuits
are entirely parallel.
Fake pins: pins that happen because something connects to a cell at an
unexpected position. Such a pin is difficult to find. The solution is
to keep those nets and nur purge them so these nets can be identified
in the layout.
Here: is_floating? will be true only if there are no pins. Hence
nets with pins are not removed. is_passive is introduced for nets -
passive nets are such that don't have elements, but a pin.
Circuits are purged if they only have passive nets.
The problem was that with the floating test case, the
ambiguity resolution sometimes assigned the wrong pins
and floating pins/connected pins were swapped.
One option is to make the ambiguity resolver consider
the pin connection state when tenatively evaluating
nodes.
Another option is to put more emphasis on net names
and use them for ambiguity resolution. This has helped
here.
1.) is_floating is now only true if there is no device
and no subcircuit on a net. This means we only purge
nets if they are really floating. So far we purged
nets without pins which lead to the mismatch:
Before purge:
Layout: (net) <--> DEVICE.TERMINAL
Schematic: PIN <--> DEVICE.TERMINAL
After purge:
Layout: (null) <--> DEVICE.TERMINAL
Schematic: PIN <--> DEVICE.TERMINAL
(null does not match any net)
2.) circuit pin matching was a bit picky. Only when
one circuit did not have pins, matching was sloppy.
In real cases however, circuits may have unconnected
pins:
- top level pins without a counterpart (no label)
- subcircuits pins which are not used
We catch both cases by refining the match: if a pin
is not used, it does not need to match against
any other pin. It's reported as "matching against null"
though.
Matthias Koefferlein