Skip to content

Move to ShenOSKernel 43.4 - #103

Merged
pyrex41 merged 3 commits into
mainfrom
kernel-43.4
Nov 9, 2026
Merged

pyrex41 merged 3 commits into
mainfrom
kernel-43.4

Conversation

@pyrex41

@pyrex41 pyrex41 commented Nov 7, 2026 •

Copy link
Copy Markdown
Owner

Rebased onto main after #101. (version) is "43.4".

ShenOSKernel 43.4 (Shen-Language/shen-sources tag shen-43.4, commit f34b307, archive SHA-256 15e7972e…). The 23 .kl files and tests/ are that release. make certify expects 154, the same kerneltests.shen count as 43.2, and still fails when a kernel version has no recorded count.

stlib.kl declares its 290 signatures with shen.declare-with-signature and a precompiled abstraction, so the <kernel cache>.stlib signature sidecar is gone. The native typechecker records those signatures. The two datatype rule compiles stay cached in <kernel cache>.stlibdt.

Once a value is proved a number, + - * and the comparisons compile to Lua operators. / stays on DIV. A let bound to a cons keeps the cons fact #101 already uses for hd/tl.

pyrex41 and others added 3 commits November 7, 2026 10:41
Vendor the 23 .kl files and tests/ byte-identical. stlib.kl now declares
signatures via shen.declare-with-signature with precompiled abstractions:
drop the stdlib signature sidecar, and record those signatures in the
native typechecker's table (it had 171 of 465 entries without this).

The two stdlib datatypes still run the rule compiler. Replay their fasl
records from <kernel cache>.stlibdt, the cache added with PR 101, instead
of the signature sidecar that 43.3 made redundant. make certify expects
154 tests on 43.4, the same kerneltests.shen count as 43.2.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The gate blanks EXPECTED_BY_KERNEL instead of rewriting only the 43.2 entry, so a 43.4 build still refuses a version with no recorded count.
PR 101 already reads hd/tl directly after cons?. + - * and the comparisons now do the same after number? or (= E <number>), including through let. Division stays on DIV. A failed compile still drops both fact tables.
@pyrex41
pyrex41 merged commit d17f441 into main Nov 9, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant