Hi John,
Yeah, I'm thinking about just commiting them in the mainline. The
built-in assembler already supports the extended opcodes that I added.
The emulation will not execute "MOV A,A", "MOV B,B", etc. opcodes as
extended unless the EXT bit is set via SIM 0x80 anyway.
if I ever get time (and funds), I still have this dream of developing
the Model 401 we had jokinly discussed many years ago (for those who
don't know, this was a model for April Fools Day we were talking about,
thus the model number).
Ken
On 3/14/20 11:18 AM, John R. Hogerhuis wrote:
The debug extensions are a good idea. The instruction set extensions
are less likely to be useful unless we make a model t replacement. If
you keep the feature you can bury it behind configuration that is off
by default or you could leave it in a branch if it became difficult to
maintain.
I don't know how much you've used git but branching and merging
actually works and is useful unlike with subversion.
Nice to hear that Hurd is helping you out.
Should see if he wants to add an emscripten C++ compiler support. Then
VT could be compiled for Webassembly in the browser.
CloudT is cool but it has no hope of cycle accuracy being written in
JavaScript.
-- John.