One thing that would make it much easier to experiment with SIMH in scenarios like this is if its build system wasn't horribly redundant. It genuinely looks like someone looked at make and said "How can I turn this into a procedural scripting system?" and then wrote the SIMH makefile in that style.
It should only take a couple seconds to build SIMH but instead it takes a couple minutes, rebuilds a ton of things redundantly, and runs all sorts of testing as a side-effect (instead of having that under a separate target). At one point I worked out that most of the preprocessor macros fall into just a couple of buckets so building SIMH could be separated into building just a couple of libsimh libraries from the same sources (one for 32-bit simulated pointers, one for 64-bit simulated pointers) and then most of the rest of the targets could *just* be the target-specific sources plus the right libsimh. Unfortunately (1) I can't contribute back that change if I make it without jumping through a lot of bureaucratic hoops (employment agreement) and (2) the current maintainer appears to not want to hear any criticism whatsoever of SIMH's build system. -- Chris
