specs/port/fpga: ported FPGA tests are read again (#5728) - #5729
Merged
Merged
Conversation
Four ported FPGA specs wrote their BDD tests in Zig syntax (glued
`test_name`, `T{ .f = v }`, `[0] ** N`, tuple destructuring, bare enum
literals). The parser discarded 312 + 219 + 435 + 422 tokens, so the
generated Zig kept only the first few tests. The tests are now t27.
The port bugs they expose are fixed against the original Verilog: the
e8m0 capture bit packing and the shift direction, and the w_code bit order
in ternary_mac_demo_top. Three wrong expectations are corrected, each
with a comment. The four parse-no-discard ledger entries are removed and
max_entries goes from 134 to 130.
Closes #5728
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
Closes #5728 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
PR DashboardGenerated at: 2026-10-03 10:28:27 UTC
Summary
Seal Status
|
This was referenced Oct 3, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #5728
t27c parse-complete --shownow printsnothing discardedfore8m0_jtag(was 312 tokens),link_node(219),ternary_mac_demo_top(435) anduart_echo_top(422).link_node_process), 3→9 and 3→15.t27c test-reportgives 18/18, 16/16, 9/9 and 15/15 (master: 2/3, 16/16, 2/3, 2/3).e8m0_jtag, the capture word now places the flag bits correctly and the shift register shifts right (sr <= {tdi, sr[31:1]}). Internary_mac_demo_top,w_codefor2'b01is[1, 0]. Three wrong expectations are corrected, each with a comment: the TDO bit of 0xA5A531BC, and the one-step counter ticks internary_mac_demo_topanduart_echo_top. No check was deleted.zig ast-checkis clean, on master and here. None of these specs has a seal.parse-no-discardentries are removed andmax_entriesdrops by four, to 126 on top of specs/numeric: GF8/20/24/32 invariants check concrete points instead of prose the parser drops #5702 (which had already lowered it to 130; master is merged in). A localt27c suite --ratchet --corpus-onlyreportsRATCHET: CLEAN.check_assertionless_spec_tests.pyis OK.gft_dup2_jtag.t27from the same batch. Its tests loop 2^24 steps and assume the beat counts raw steps rather than slow-clock ticks, so it needs a test redesign, not a syntax repair.🤖 Generated with Claude Code