fix: keep request payloads out of the issue store #23

2026-08-10T12:26:41Z

claude commented

2026-08-10 12:26:41 +00:00

Closes #19.

labels.py passed the issue store to _gitea.api for one reason only: the
request file needed somewhere to live. On a checkout without a store that
silently created tmp/issues/.payload/ — a store materialized as a side effect
of an operation that has nothing to do with issues, and nothing said so.

What changed

Where a request body goes is now the transport's decision, not the caller's.

_gitea.py resolves PAYLOAD_ROOT (<repo>/tmp/payload) from its own
location, the same way issue.store_root and page.store_root do, so every
caller writes to one directory whatever it was invoked from.
out_root is gone from api() and add_dependency(), and from all six call
sites: labels.py, push.py (x3), comment.py (x2), wiki_push.py (x2).
Sync and wiki move together — no mixed state.
The directory is created by the first write of a run and not before, so a
--dry-run or a run with nothing to send leaves nothing behind.
tmp/ is already gitignored, so the new location is covered.
The name carries the distinction the old path lost: a visible payload/
beside issues/ and wiki/, not a dotdir among an issue's files. A store
holds the only copy of something; this holds debris kept for a retry.

Docs updated: AGENTS.md (a new "Request payloads" section, plus labels.py
in the repo layout), skills/sync/SKILL.md, skills/wiki/SKILL.md.

Acceptance criteria

labels.py neither creates nor requires the issue store on any path
the payload directory is not the issue store, and is named so that this
is visible
every other script in the sync layer writes payloads to the same place —
they all moved together
.gitignore covers the new location (it is under tmp/)
a test proves labels.py on a tree without tmp/issues does not create
it

Tests

tests/test_payload_root.py (new, 12 tests) runs the real labels.py inside a
throwaway repository with a fake tea first on PATH — no network, and the
developer's own store is never in reach. It asserts that no store appears, that
the payloads land in tmp/payload/ and are the bodies actually sent, that a
dry run writes nothing, and that a run from a subdirectory still resolves to the
repo root. Two source checks keep the callers from drifting apart again.

Against the pre-fix tree these fail with labels.py created the issue store.

python3 -m unittest discover -s tests -v
Ran 244 tests in 5.3s
OK

🤖 Generated with Claude Code

Closes #19. `labels.py` passed the issue store to `_gitea.api` for one reason only: the request file needed somewhere to live. On a checkout without a store that silently created `tmp/issues/.payload/` — a store materialized as a side effect of an operation that has nothing to do with issues, and nothing said so. ## What changed Where a request body goes is now the transport's decision, not the caller's. - `_gitea.py` resolves `PAYLOAD_ROOT` (`<repo>/tmp/payload`) from its own location, the same way `issue.store_root` and `page.store_root` do, so every caller writes to one directory whatever it was invoked from. - `out_root` is gone from `api()` and `add_dependency()`, and from all six call sites: `labels.py`, `push.py` (x3), `comment.py` (x2), `wiki_push.py` (x2). Sync and wiki move together — no mixed state. - The directory is created by the first write of a run and not before, so a `--dry-run` or a run with nothing to send leaves nothing behind. - `tmp/` is already gitignored, so the new location is covered. - The name carries the distinction the old path lost: a visible `payload/` beside `issues/` and `wiki/`, not a dotdir among an issue's files. A store holds the only copy of something; this holds debris kept for a retry. Docs updated: `AGENTS.md` (a new "Request payloads" section, plus `labels.py` in the repo layout), `skills/sync/SKILL.md`, `skills/wiki/SKILL.md`. ## Acceptance criteria - [x] `labels.py` neither creates nor requires the issue store on any path - [x] the payload directory is not the issue store, and is named so that this is visible - [x] every other script in the sync layer writes payloads to the same place — they all moved together - [x] `.gitignore` covers the new location (it is under `tmp/`) - [x] a test proves `labels.py` on a tree without `tmp/issues` does not create it ## Tests `tests/test_payload_root.py` (new, 12 tests) runs the real `labels.py` inside a throwaway repository with a fake `tea` first on `PATH` — no network, and the developer's own store is never in reach. It asserts that no store appears, that the payloads land in `tmp/payload/` and are the bodies actually sent, that a dry run writes nothing, and that a run from a subdirectory still resolves to the repo root. Two source checks keep the callers from drifting apart again. Against the pre-fix tree these fail with `labels.py created the issue store`. ``` python3 -m unittest discover -s tests -v Ran 244 tests in 5.3s OK ``` 🤖 Generated with [Claude Code](https://claude.com/claude-code)

claude added 1 commit 2026-08-10 12:26:41 +00:00

fix: keep request payloads out of the issue store 596cf853e8

`labels.py` handed `_gitea.api` the issue store as a place to put the
request file, and on a checkout without a store that quietly created
`tmp/issues/.payload/`. Bootstrapping a repository's labels touches no
issue at all, so the one rule the store has — nothing materializes it as
a side effect of a write — was broken by an operation that has no
business knowing the store exists.

Where a request body goes was never the caller's decision to make. It is
now the transport's: `tmp/payload/`, resolved from `_gitea.py`'s own
location the way both domains resolve theirs, so every caller — sync and
wiki alike — writes to one directory whatever it was invoked from, and
`out_root` is gone from `api`, `add_dependency` and all six call sites.
The directory is created by the first write of a run and not before: a
`--dry-run` leaves nothing behind. `tmp/` is already gitignored.

The name carries the distinction the old path lost. A store holds the
only copy of something; this holds debris kept for a retry or a
post-mortem, and deleting it costs nothing. A dotdir sitting among an
issue's files claimed otherwise, and `ls tmp/issues` started lying about
what existed.

tests/test_payload_root.py runs the real `labels.py` in a throwaway repo
against a fake `tea` on PATH: no store appears, the payloads land in
tmp/payload/, a dry run writes nothing, and a run from a subdirectory
still resolves to the repo root. Two source checks keep the callers from
drifting apart again.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

naudachu added 1 commit 2026-08-10 12:32:38 +00:00

test: keep the payload root out of the developer's tree 627df76812

One test stubs the transport a layer below `api()` — at `subprocess`, to
exercise the path a 422 really takes — so it reaches the real payload
write. That used to land in the test's own temp store, because the caller
named the directory; now the directory is `_gitea.PAYLOAD_ROOT`, resolved
from the module's location, and the file appeared in the developer's
`tmp/payload/`.

`StoreTestCase` patches `PAYLOAD_ROOT` to its fixture alongside the other
seams, and the rule in AGENTS.md gains the third directory a test must
not write to.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

naudachu added 1 commit 2026-08-10 13:23:42 +00:00

Merge remote-tracking branch 'origin/main' into fix/no-store-for-label-payloads 2a8da81359

# Conflicts:
#	skills/sync/scripts/_gitea.py

claude merged commit 40016e06f2 into main

2026-08-10 13:24:35 +00:00

Sign in to join this conversation.

2 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: claude-skills/marketplace#23