sharukhan phase-0: scaffold SDD artifacts - #319
Closed
dcasota wants to merge 1 commit into
Closed
Conversation
Creates the subproject skeleton so it can receive specs, following the
convention already live in this repository (tdnf-depgraph, docsystem,
vCenter-CVE-drift-analyzer): specs/{README,prd,adr,features,tasks,findings},
sdd/<initiative>-phase-N-<slug> branches, and
'<subproject> phase-N task NNN: <summary>' commit subjects.
ARCHITECTURE.md records why this is one binary rather than more shell: the
tooling it replaces resolved inputs implicitly (two downstream-fixes.patch
copies drifted 78 lines and 8-vs-27 files apart, and the build silently used
whichever sat beside the invoked script), had checks that could not fail
(40-check-staging.sh never exits non-zero), and rested on pipelines that can
be silently vacuous (/usr/bin/grep is toybox non-interactively and has no -a,
returning zero matches on a NUL-bearing serial log).
It also records the axis model that makes 34 permutations cost 4 ISO builds:
ISO type and installer version are build-time, while STIG, filesystem and
kickstart-vs-UI are free because isoInstaller reads guestinfo.kickstart.data
via vmtoolsd, which is present in the installer initrd.
AGENTS.md is authored by hand and says so. Neither nominated SDD source
contains a constitution file; the AGENTS.md that SDD-book-tracking-app
references is generated by an Azure-centric APM toolchain and is not
committed there, so nothing here is inherited or generated.
Signed-off-by: Daniel Casota <dcasota@gmail.com>
dcasota
pushed a commit
that referenced
this pull request
Sep 9, 2026
…facts PRs #319/#320 proposed staging/sharukhan/ while the CLI lives at staging/sharukhan-cli/ - two directories that were never reconciled, which is why both carried a PRD. This lands their content in the CLI's own tree and adopts the convention the four sibling SDD subprojects already share (tdnf-depgraph, docsystem, vCenter-CVE-drift-analyzer, photon-gating-conflict-detection): specs/adr/NNNN-<slug>.md, plus features/, tasks/, findings/, archive/. specs/adr-0001-run-stop-watch.md -> specs/adr/0001-run-stop-watch.md specs/prd.md (2863 b) -> specs/archive/prd-2026-08-initial.md #320's PRD replaces it as specs/prd.md. It is the SDD-template form the siblings use, and it is closer to the shipped tool - it names `run` 15 times against 1, `build` 9 against 2, and mentions verify and canister at all. The emoji was dropped from its title to match the house style. The short one is archived rather than deleted; it is the earlier sketch, not wrong. schema/memory.sql is NOT #320's copy. That one predates the schema: 6 tables, no indexes, no views, no next_step. This is generated from the live database, which matters more than it looks - src/memory.rs opens the database READ-ONLY and never creates it, so until now nothing in the repository could recreate a lost memory.db. Verified by applying it to an empty file: 14 objects, an identical object set to the live database. MEMORY.md is likewise regenerated rather than imported. #320's rendering shows 0 rows in every table; the live one has 121 runs, 3374 check results and 52 findings. vm-lab (#316) is archived under superseded-vm-lab/ rather than dropped. ARCHITECTURE.md names it as one of the two toolkits sharukhan replaces, and mission-control's bash is already preserved the same way. ARCHIVE.md records why it went, quoting the architecture's own verdict: 40-check-staging.sh never exits non-zero, which is fine for an inspection tool and useless as a gate. Path references in the imported prose were repointed to where the files now actually live. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JW73JTCUGRcaNTUEQcAMtf
Owner
Author
|
Landed by another route — closing rather than merging. The content of this PR is in
Why the move. This PR targeted The one pre-existing deviation was fixed at the same time — Path references inside the imported prose were repointed to where the files actually live. |
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.
Phase 0 of a new SDD-tracked subproject: a single standalone Rust CLI,
sharukhan, replacing the shell tooling instaging/vm-labandstaging/mission-control.Per
specs/README.mdthe phases gate each other via merged PRs, so this PR only makes the subproject ready to receive specs. Nosrc/in this PR — implementation is blocked until the PRD merges.What is here
specs/README.mdARCHITECTURE.mdAGENTS.mdConvention followed, not invented
Layout and workflow mirror
tdnf-depgraph/specs/— the sameprd.md/adr/NNNN-<slug>.md/features/<slug>.md/tasks/NNN-task-<slug>.md/findings/set,sdd/<initiative>-phase-N-<slug>branches, and<subproject> phase-N task NNN: <imperative>commit subjects. The quality gates are restated in Rust terms (cargo test,cargo clippy -- -D warnings,cargo fmt --check, ≥80% coverage — the in-repo precedent) because those differ by toolchain.One thing worth flagging
Neither nominated SDD source contains a constitution file. There is no
constitution.md, nomemory/, no.specify/in eithervCenter-CVE-drift-analyzerorsitoader/SDD-book-tracking-app. TheAGENTS.mdthe latter references is generated by an Azure-centric APM toolchain (apm install && apm compile) and is not committed there.So the
AGENTS.mdin this PR was authored by hand for Rust, and both it andspecs/README.mdsay so explicitly. I would rather state that plainly than present an invented file as an inherited standard.Why one binary
ARCHITECTURE.mdrecords three observed problems in the tooling being replaced, not hypothesised ones:runPh5_normal.shresolves its patch relative to its own location; the two copies on this host had drifted 78 lines and 8-vs-27 files apart, and the build used whichever sat beside the invoked script. It surfaced aspatch does not applyagainst a spec, which reads like a rebase problem.40-check-staging.shnever exits non-zero./usr/bin/grepis toybox in a non-interactive shell and has no-a, returning zero matches on a NUL-bearing serial log; interactively the same name isugrep.sed \Uandgrep -Pare GNU-only and absent here.It also records the axis model that makes 34 permutations cost 4 ISO builds: ISO type and installer version are build-time, while STIG, filesystem and kickstart-vs-UI are free, because
isoInstallerreadsguestinfo.kickstart.datathroughvmtoolsdandvmtoolsdis present in the installer initrd.Next:
specs/prd.mdonsdd/sharukhan-phase-1-prd.