* Copies guidelines files into docs/ for website
* Copying manual/CHAPTER_Prog for new docs
* Copying manual/APPNOTE_011... for new docs
Also adding faketime to list of packages for website build.
Co-authored-by: KrystalDelusion <krystinedawn@yosyshq.com>
The previous approach only initialized past_d and past_ad while for FST
cosim we also need to initialize the other past values like past_clk,
etc. Also to properly initialize them, we need to run a combinational
update step in case any of the wires feeding into the FF are private or
otherwise not part of the FST.
When writing VCDs smtbmc replaces square brackets with angle brackets to
avoid the issues with VCD readers misinterpreting such signal names.
For memory addresses it also uses angle brackets and hexadecimal
addresses, while other tools will use square brackets and decimal
addresses.
Previously the code handled both forms of memory addresses, assuming
that any signal that looks like a memory address is a memory address.
This is not the case when the user uses regular signals whose names
include square brackets _or_ when the verific frontend generates such
names to represent various constructs.
With this change all angular brackets are turned into square brackets
when reading the trace _and_ when performing a signal lookup. This means
no matter which kind of brackets are used in the design or in the VCD
signals will be matched. This will not handle multiple signals that are
the same apart from replacing square/angle brackets, but this will cause
issues during the VCD writing of smtbmc already.
It still uses the distinction between square and angle brackets for
memories to decide whether the address is hex or decimal, but even if
something looks like a memory and is added to the `memory_to_handle`
data, the plain signal added to `name_to_handle` is used as-is, without
rewriting the address.
This last change is needed to successfully match verific generated
signal names that look like memory addresses while keeping memories
working at the same time. It may cause regressions when VCD generation
was done with a design that had memories but simulation is done with a
design where the memories were mapped to registers. This seems like an
unusual setup, but could be worked around with some further changes
should this be required.
Otherwise the final values will not show up in gtkwave waveforms
when looking at the generated traces.
Signed-off-by: Claire Xenia Wolf <claire@clairexen.net>