Wilson Snyder

Point taken. The port connection depends on a correct data type of the child module, and that data type incorrectly (or more accurately not yet correctly) has a length. I...

BTW a complication with what I said above is we'd need to clone a module, determine parameters, and only then would know that the deparam'ed module is a duplicate. At...

> force the config file to be first in the command (?) Hmm. Presently .vlt files are handled same as .v files. This in theory allows the controls to be...

Thanks for debugging, f you run with --debug in the .tree file you can see the levels. (L and number) ``` 1: MODULE 0x555557065680 {f45ai} __024root L1 [P] [1ps] 1:...

Looking at #5708, perhaps that is related, the level is calculated based on recursion need not what the user observes. Also if an interface is used at several different levels,...

For similar in test_regress format: ``` interface AXIS_IF ( input logic aclk ); logic [127:0] tdata; logic [ 31:0] tuser; logic tvalid, tready; modport master(input aclk, output tdata, tuser, tvalid,...

Implemented as described in comments.

Notes on reproducing: ``` test_regress/t/t_udp_sequential.py --vltmt --debug ( cd test_regress ; while obj_vltmt/t_udp_sequential/Vt_udp_sequential; do :; done ``` in another window load up the system, should fail in a minute ```...

int vs uint32_t (vs size_t in C++) inconsistencies are rampant, and not really worth cleaning up as this won't happen with real code. Might you make a pull to warn...

BTW part of the problem will always be that IEEE says part select indices are signed 32-bits, while C++ is size_t - 64 bit unsigned.