Skip to content

proposal: a declared reopen returns the source when its target is retired - #48

Merged
delchev merged 3 commits into
mainfrom
spec/reopen-on-retire
Aug 24, 2026
Merged

proposal: a declared reopen returns the source when its target is retired#48
delchev merged 3 commits into
mainfrom
spec/reopen-on-retire

Conversation

@delchev

@delchev delchev commented Aug 21, 2026

Copy link
Copy Markdown
Contributor

Adds proposals/0025-declared-reopen-on-retire.md.

Follow-up to 0024 (a retired target stops blocking its source), which left one combination unreachable.

The gap. An event-driven create-from may declare a completion hook — the status the SOURCE moves to once the target exists. It exists precisely so the guard-claimed source stops matching its own trigger. 0024 then made a retired target stop blocking, which frees the source's slot. Put the two together and the door is unlocked with nobody able to knock: the source stands at the post-generation status, a lifecycle graph declares no edge back, so the source never transitions through the qualifying status again and no qualifying event is ever published. Where a button is also declared a person can click it; an event-only rule — the shape the event axis exists for — has no reissue path at all, and nothing says so.

What this proposes. The hook's inverse on the same rule: where the source returns when a target it produced is retired. Retiring the target returns the source, which is a real transition announced on the source's own channel, so the ordinary trigger re-fires, the guard steps over the retired target, and the replacement is minted. The reissue is the ordinary path — assembled from rules already in the language, not a second creation route with semantics of its own. Nothing new declares what "retired" means: it is the same stage: classification the guard reads, asked from the other end.

Idempotence is by state, not a marker: the reopen acts only while the source still stands at that rule's own completion status. And whether the replacement is immediate stays the trigger's decision — returning the source to an earlier status (a draft, for correction) makes the reissue reviewable rather than automatic.

The two alternatives are recorded in the proposal: having the target's retirement re-deliver the source's qualifying event (fabricates a transition nobody made, and re-fires every other consumer of that channel), and documenting-plus-warning (honest, but leaves the automation inexpressible).

Numbered 0025 — 0021–0024 are taken by open PRs; renumber as you see fit.

Reference implementation: eclipse-dirigible/dirigible#6868.

…ired

A retired target stops blocking its source (proposal 0024), so the source's
one slot is freed. Where the create-from also declares a completion hook,
nothing can refill it: the hook exists to move the source OFF the status its own
trigger qualifies on, and a lifecycle graph declares no edge back - so the source
never re-qualifies and no event ever fires again. An event-only rule has no
reissue path at all, and nothing says so.

This proposes the hook's inverse on the same rule - where the source RETURNS when
a target it produced is retired - so the reissue becomes the ordinary path:
a real transition of the source, announced on the source's own channel, which
the trigger already listens to.

Nothing new declares what "retired" means: it is the same stage classification
the guard reads, asked from the other end.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
delchev and others added 2 commits August 21, 2026 15:56
A retirement is delivered at-least-once, so it can arrive again after the
replacement already exists - and by then the source is standing at sourceStatus
once more, because the reissue put it there. The standing-status test alone would
re-open a source with a live target against it. The reopen now runs exactly when
a creation would be allowed through.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A create-from triggered only by a button carries no at-most-once guard, so
nothing blocks a replacement - the button is itself the reissue, and there is no
trigger for a reopen to re-fire.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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