Skip to content

fix(run): skip IDL step for non-lez-framework projects#162

Open
weboko wants to merge 1 commit into
masterfrom
weboko/fix-dogfooding
Open

fix(run): skip IDL step for non-lez-framework projects#162
weboko wants to merge 1 commit into
masterfrom
weboko/fix-dogfooding

Conversation

@weboko
Copy link
Copy Markdown
Collaborator

@weboko weboko commented May 14, 2026

Description

While running the full DOGFOODING.md scenarios end-to-end, three first-success-path issues surfaced. This PR addresses them.

What changed

src/commands/run.rs — Gate the Building IDL… step on framework.kind == "lez-framework". build_idl_for_current_project bails on default-template projects rather than no-op'ing, so lgs run on the default template previously failed at step 2/5 with build idl is only supported for lez-framework projects. Step labels ([i/n]) now renumber dynamically so the user sees consecutive markers in both the default (4-step) and lez-framework (5-step) pipelines.
src/doctor_checks.rs + src/commands/doctor.rs — New risc0 r0vm and logos-blockchain-circuits doctor rows. Both are sequencer runtime prereqs that live outside the project tree; without them the sequencer panics on its first block-settlement (zk signature) or first risc0 program execution with opaque ProgramExecutionFailed / Main loop exited unexpectedly errors. Surfacing them at lgs doctor time with actionable install hints replaces a mid-run crash with a clean preflight failure.
src/cli_help.rs — Drop the logos-scaffold create my-app --cache-root … example from create/new help. The flag was never wired up; the non-vendored cache root is XDG-resolved.
DOGFOODING.md — Add the runtime prereqs (rzup, risc0 rust, r0vm, logos-blockchain-circuits) to "Shared Preconditions". Capture the lez-framework-vs-LEZ-pin compile mismatch as a "Known Gap" under L1 so L1–L4 evidence is recorded the same way across runs. Rewrite E2 against the real new/create flag surface (no --cache-root; document XDG_CACHE_HOME instead).

Test plan

D1–D7 pass end-to-end against a fresh lgs new dogfood-default project (including lgs run with and without [run].post_deploy hooks).
lgs doctor reports the two new rows with PASS when r0vm/circuits are installed and FAIL with the remediation strings when they're absent.
Existing unit tests in commands::run (42) and doctor_checks (3) pass; no new tests added since these are display/wiring changes.

@weboko weboko requested a review from a team May 14, 2026 20:24
@weboko weboko force-pushed the weboko/fix-dogfooding branch from c05855b to baa1bdd Compare May 14, 2026 20:46
…verage; document sequencer prereqs

The `lgs run` pipeline previously called `build_idl_for_current_project`
unconditionally, but that function bails on non-lez-framework projects
rather than silently no-op'ing. The default-template scenario in
DOGFOODING.md (D7) therefore failed at step 2/5 with
`build idl is only supported for lez-framework projects`. Gate the IDL
step on the framework kind and renumber the printed step labels
dynamically so the user sees consecutive `[i/n]` markers.

While running the dogfooding scenarios end-to-end I also hit two
external-dependency gaps that produce opaque sequencer crashes
(`ProgramExecutionFailed: No such file or directory`,
`Main loop exited unexpectedly`). Add doctor rows for `risc0 r0vm` and
`logos-blockchain-circuits` with actionable remediation hints, and
document the sequencer's runtime prereqs in the dogfooding runbook so
they're caught at `lgs doctor` time instead of mid-run.

Also remove the `--cache-root` flag from the create/new help examples
(the flag was never wired up; the non-vendored cache root is XDG-resolved)
and capture the current lez-framework-vs-LEZ-pin compile mismatch as a
known gap in DOGFOODING.md so L1-L4 evidence is recorded the same way
across runs.
@weboko weboko force-pushed the weboko/fix-dogfooding branch from baa1bdd to 7d7b178 Compare May 14, 2026 20:55
Comment thread DOGFOODING.md

### Known Gap (current pins)

The `lez-framework` pin (`jimmy-claw/lez-framework` rev `1e146970…`) referenced
Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@vpavlin is it time to fix it?

Copy link
Copy Markdown
Collaborator

@fryorcraken fryorcraken left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would say no, that scaffold is only here to support project that use spel.

@weboko
Copy link
Copy Markdown
Collaborator Author

weboko commented May 19, 2026

I would say no

no to what exactly?

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.

3 participants