I paused my experiments long enough to run the Windows experiments.
The tests *mostly* work, except for test-sysdef-asdf which fails as follows:
TEST ABORTED: Subprocess with command ("make.bat" "driver_files") exited with error code 1
AFAICT this just echoes a list of files that's hardcoded into make.bat:
uiop\package.lisp + uiop\common-lisp.lisp + uiop\utility.lisp + uiop\os.lisp + uiop\pathname.lisp + uiop\filesystem.lisp + \ uiop\stream.lisp + uiop\image.lisp + uiop\run-program.lisp + uiop\lisp-build.lisp + uiop\configuration.lisp + uiop\backward-driver.lisp + u\ iop\driver.lisp
So I don't see why we are running the subprocess instead of hard-coding the list of files into test-sysdef-asdf.
The test claims to be testing that the makefile and .asd agree on the order of files, but it's not obvious how this happens.
For now I'm simply disabling this test on windows, but if I can do something better than that, please LMK.
Thanks, r
OK, I think I see the solution. In ordinary CL I can read the vanilla Makefile, find the line with the list of files, and compare that against what's in the .asd file.
That seems to be the intent of this test.