docs: M1.5 handoff for the port codegen defects - #74
Merged
Conversation
Reviewing PR #71 found one urdr fail-open and, separately, that `make conformance` was already red. That led to three pre-existing code-generation defects in the Shen ports -- two in shen-lua, two in shen-cl -- none caused by #71, which was merely the first change deep enough to trip them. All are fixed and merged (shen-lua#53, shen-cl#12). The part worth a document is what the fixes do NOT address: all four ports carry a byte-identical core.kl, and the same source expands to the same 3457-node, depth-435 KL form on every one of them. The kernel's pattern compiler re-derives the full hd/tl spine per pattern element, so every port receives the same pathological KL and they differ only in host tolerance. Both port fixes are coping mechanisms; the shared cause is still there and a deeper pattern will find it again. Records the shen-cl factoriser blowup separately because it is not an urdr problem at all -- code grew 2^groups and a 33-clause dispatch function was fatal, which any moderately large dispatch table would have hit. Also records the ADR 0007 corrections, the open debts (verification item 5, the core.kl expansion, shen-lua's other reader productions), and two method notes: never pipe a gate run through head/tail, and never hand-write a build stamp. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds
docs/status/m1-5-port-codegen-handoff.md, following them1-port-perf-handoff.mdconvention.Documents the three pre-existing Shen port code-generation defects
surfaced while reviewing #71 (all fixed and merged: pyrex41/shen-lua#53,
pyrex41/shen-cl#12), and — the reason this is a document rather than a
commit message — the fact that their root cause is shared and still
unfixed.
All four ports carry a byte-identical
core.kl, and the same sourceexpands to the same 3457-node / depth-435 KL form on every one of them.
The kernel's pattern compiler re-derives the full
hd/tlaccessorspine per pattern element; every port receives the same pathological KL
and they differ only in host tolerance. Both port fixes are coping
mechanisms in the backends.
Also records:
dispatch function was fatal at 446s / 938MB) — not an urdr problem,
any large dispatch table would have hit it
shen/run/run.shen's 8092-byte header comment is what trippedshen-lua's reader, which makes this a live constraint on the
repository's own house style
Met mid-review while item 5 remains a standalone debt
every repo, so
make conformancemust be run by hand before mergeshead/tail(SIGPIPEtruncation reads as a suite that stopped early), and never hand-write
a
.urdr-build-stampDocs-only.
make fmt-checkpasses.🤖 Generated with Claude Code