The package name has never matched the repository it lives in
(gitea.awkawb.cloud/awkawb/invoiceninja-sdk, mirrored to
github.com/awkawb1/invoiceninja-sdk) — both pointed at `invoiceninja-sdk`
while `Cargo.toml` declared `invoiceninjars`. Lines everything up on the
name the repo has always used.
Cargo derives the crate's Rust identifier from the package name by
lowercasing hyphens to underscores, so every `use invoiceninjars::...`
becomes `use invoiceninja_sdk::...` — in doctests across `src/`, the four
examples, all four integration test files, and `README.md`. `Cargo.lock`'s
own self-entry picked up the new name on the next `cargo check`, which is
also when the rename gets its real check: every doctest and every test file
has to resolve the new path to compile at all, not just the ones a grep
happens to catch.
`flake.nix`'s `pname` and description follow, so `nix build` produces
`invoiceninja-sdk-0.1.0.drv` rather than a name that no longer matches
anything. `GOALS.md`'s two mentions — the hypothetical CLI's name and the
`cargo install` example — follow too, for the same reason.
Verified clean on both the dev toolchain and the 1.88 MSRV shell: full test
suite (209 tests), clippy, fmt, and `nix build`/`nix flake check`.
Not yet published to crates.io (checked: the name 404s there), so this
lands before the old name was ever claimed — no one's `Cargo.toml` breaks
today, but it's recorded as a Changed entry since `GOALS.md` asks for
breaking changes to be legible in the changelog rather than discovered at
compile time, and the crate is reachable as a git dependency regardless of
crates.io status.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
`nix build` now works, using `rustPlatform.buildRustPackage` built with the
pinned dev toolchain rather than whatever rustc nixpkgs happens to carry, so
it compiles with the same compiler `nix develop` and CI do. `cargoLock`
points at the repo's own `Cargo.lock` rather than a separately maintained
hash.
The crate has no `[[bin]]` target — see the "No CLI, and no binary at all"
non-goal just added to `GOALS.md` — so the resulting store path comes out
empty: `cargoInstallHook` only copies files matching a `[[bin]]` target out
of `target/release`, and there is nothing to copy. Verified rather than
assumed; the first version of this comment claimed the output held "an rlib
and docs", which is not what actually lands in the store path.
What `nix build` is good for regardless: it compiles with the pinned
toolchain and runs the full suite, doctests included, inside the Nix
sandbox with no network — matching the "no credentials needed" property
`CLAUDE.md` already claims for `cargo test`, now exercised a second way.
Treat it as a build-and-test gate, not a source of an installable artifact.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Running the password tier against a disposable 5.13.50 instance for the
first time failed its last check. The purge itself was fine — the row really
was gone — but the assertion looked for a 404 and the server answers 400.
`Handler.php` maps Laravel's `ModelNotFoundException` to a 400 carrying
`No query results for model [App\Models\X].` and reserves 404 for a route
that does not exist, so a not-found check written against 404 never fires.
The assertion now matches the status together with the message, so a 400
raised for some other reason still fails, and `Error::Api` says so for
callers, who would otherwise write the same dead branch.
Also adds jq to the devshell, which the MSRV recheck in CLAUDE.md pipes
through but which was only ever resolving from the host.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The dev shell tracked `stable.latest` and CI used `@stable`, so nothing
pinned a version and nothing declared a floor. A consumer on an older
toolchain hit a confusing error deep in a dependency rather than cargo
saying the version was too old, and a new stable release could turn a
lint into an unrelated red build.
- `rust-version = "1.88"` in Cargo.toml. The floor comes from the
dependency graph, not this crate's code: reqwest pulls in url -> idna
-> idna_adapter -> icu_*, and the icu crates require 1.88. Verified by
running the full suite against 1.88, not inferred.
- flake.nix pins the dev toolchain to 1.99.1 and adds an `msrv` devShell,
so the floor can be rechecked with
`nix develop .#msrv --command cargo test`.
- CI pins the main job to the same 1.99.1 and adds a job building on
1.88. Lint and format stay on the pinned toolchain only, since their
output varies by release.
Note the tradeoff: with CI pinned, a future stable that breaks the build
will not be noticed until the pin moves. Worth adding a non-blocking
`stable` job if that matters.
CLAUDE.md records the three places that must agree when either version
moves, and how to recheck the floor after a dependency bump.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Covers Clients, Invoices, Quotes, Credits, Payments, Products, Expenses,
Projects, Tasks, and Vendors, each with list, list_all, get, create,
update, and delete, generated by the resource_methods! macro.
Every model carries a #[serde(flatten)] extra map so fields the API
returns but the crate does not model survive a get/update round-trip
instead of being silently dropped.
Includes a Nix dev shell (Rust via rust-overlay), a GitHub Actions CI
workflow running build/test/fmt/clippy, and round-trip tests for the
extra-preservation invariant.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
awkawbandClaude Sonnet 5