dev: require the pinned toolchain, and clear the QEMU-era leftovers

The host gate accepted a WATCOM environment override and skipped when nothing
was found. Both are wrong for what it is: a system-wide Open Watcom is a
different version reporting different diagnostics, and a gate that skips is a
gate that is not running, which is the exact shape of the problem this file was
added to close. It now uses .dosboxx/watcom only and fails with the setup
command when that is absent, matching how dosboxx.py already behaves.

Nothing else in the project reaches for a system install: the DOS session sets
WATCOM=W: before calling BUILD.BAT, so the C:\DEVEL\WATCOMC fallback inside it
is unreachable.

fec/test-dos.bat was tracked but dead -- the runner generates RUN.BAT and only
copies build-dos.bat -- so it goes, along with the comment claiming it drives
the build and the three fixture READMEs that still pointed at it. The registry
decides what runs now.

(.qemu/ is untracked local debris from the same era and is left alone.)
This commit is contained in:
2026-08-17 02:26:02 +09:00
parent 32ad50b14a
commit d5ba699744
6 changed files with 19 additions and 309 deletions
+1 -1
View File
@@ -21,6 +21,6 @@ Fail cases:
- `pubpriv`: a public signature cannot expose a private nominal type.
- `errnom`: nominal error cannot flow into `core.Error` via `try`.
The current M5 `TEST-DOS.BAT` is intentionally unchanged. M8 should add procedural checks for `.fei` creation/hash invalidation and deterministic `fe_errors.h` using these fixtures.
They are not registered in `src/ferrolang_vm/registry.py` yet. M8 should add procedural checks for `.fei` creation/hash invalidation and deterministic `fe_errors.h` using these fixtures.
The M8 DOS gate must additionally construct two separate `-I` roots containing the same canonical unit and require an ambiguity error; repeat the case with two paths to the same canonical file and require deduplication. It must also verify that `std.*` resolves only from the built-in std root, ordinary user units never do, and that changing a private non-generic implementation preserves the dependent interface-cache hit. Those checks need temporary roots/cache inspection and deliberately remain procedural rather than encoding host paths in fixtures.