9 Commits
Author SHA1 Message Date
awkawbandClaude Sonnet 5 2c9915699d nixos-module: run db migrations on every activation, not just once
The .initial-migration marker file prevented `migrate --force` from ever
running again after the first activation, so migrations shipped by a later
invoiceninja version upgrade would never apply to an already-deployed
instance. `migrate` is idempotent (Laravel tracks applied migrations in the
DB), so just run it unconditionally on every activation instead.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-10-02 13:15:00 -04:00
awkawbandClaude Sonnet 5 bd112541e5 nixos-module: use playwright's chromium-headless-shell instead of pkgs.chromium for PDF generation
snappdf only ever shells out to `$SNAPPDF_CHROMIUM_PATH --headless --print-to-pdf=...`,
so the full browser UI/GPU stack pkgs.chromium ships is unnecessary. Google's official
"Chrome for Testing" headless-only build (packaged by nixpkgs' playwright-driver,
selected via `browsers.override` with every other browser/ffmpeg disabled) does the
same job for a ~1.45GB smaller closure (431MB vs 1.88GB), needs no allowUnfree, and
PDF generation is noticeably faster in the VM test (~2.5s vs ~10s per request).

Drops the now-unnecessary allowUnfree wiring from the interactive VM and automated
test configs, which existed solely to build pkgs.chromium.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-30 18:19:38 -04:00
awkawbandClaude Sonnet 5 b883e04303 package: strip composer-repository store path from runtime closure
composer.lock and vendor/composer/installed.json record each package's
"dist.url" as the store path of the build-time composer-repository
mirror, which drags that whole ~570MB fixed-output derivation into the
runtime closure even though it's just install provenance -- neither
Laravel's package manifest (name/extra only) nor Composer's runtime
InstalledVersions (reads installed.php, not this file) consult it.
Cuts the package closure from ~1.75GB to ~1.18GB.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-30 17:39:14 -04:00
awkawbandClaude Sonnet 5 8216a723f0 tests: add client/invoice/payment end-to-end check
Extracts small api_get/api_post helpers to cut the repeated curl/header/
json.loads boilerplate, then uses them for a new subtest that creates a
client, invoices it, marks the invoice sent, and pays it in full - checking
the invoice comes back paid with a zero balance. Runs after the PDF preview
check, since PreviewController::blankEntity only builds its mock PDF when
the company has no invoices/invitations yet.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-30 17:15:26 -04:00
awkawbandClaude Sonnet 5 ef08e0d7fc package: update invoiceninja to v5.13.43, ui to 28.09.2026.1
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-28 20:28:11 -04:00
awkawbandClaude Sonnet 5 330ec8c503 tests: fix APP_URL port mismatch in the interactive VM
`just boot-vm` forwards the guest's port 80 to host port 8080
(QEMU_NET_OPTS in the justfile), but the backend builds absolute URLs
(e.g. the stored company logo path returned by the upload API) from
APP_URL, which defaults to plain "http://localhost" with no port. The
SPA's own relative URLs are fine - they inherit whatever port the
browser is already on - but any backend-generated absolute URL then
points at host port 80, where nothing is listening, instead of the
forwarded 8080. That gave a silent net::ERR_CONNECTION_REFUSED for the
logo image specifically, even though the upload itself succeeded.

Verified directly: with APP_URL overridden to "http://localhost:8080",
a freshly uploaded logo's URL is
"http://localhost:8080/storage/.../....png" instead of missing the
port. Automated test checks are unaffected (checks derivation hash
unchanged) since they curl the guest's own localhost directly, with no
host-side port remap involved.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-26 16:47:23 -04:00
awkawbandClaude Sonnet 5 b1fe86e9d5 tests: give the interactive VM enough RAM to not OOM under real use
Root-caused a "logo upload hangs, connection refused in browser
console" report: the qemu-vm module's defaults (1 core, 1024MB) are
too tight for this stack (MySQL + Redis + PHP-FPM + chromium for PDF
generation, each PDF request spawning a full multi-process browser
instance) under real concurrent load, and this VM boots with
panic_on_oom - so exhausting memory doesn't degrade gracefully, the
whole VM goes down and every connection starts refusing.

Reproduced directly: at the 1024MB default, even 2 concurrent PDF
generation requests triggered `redis-server invoked oom-killer` and a
kernel panic. Bumped to 2 cores / 2560MB, which comfortably survives
that same load (~1.4GB still free afterward) - verified via the exact
concurrent-request scenario that crashed the 1024MB config, not just
absence of errors.

virtualisation.cores/memorySize aren't options on a plain nixosSystem
in this nixpkgs (only nixosTest nodes get the qemu-vm module
automatically), so this also adds an explicit import of it.

Documented the underlying resource cost (chromium is a full browser
per PDF request, pm.max_children=8 means up to 8 can run concurrently)
in CLAUDE.md as an operational note for real deployments, not just
this test VM.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-26 16:36:22 -04:00
awkawbandClaude Sonnet 5 ccf4d01a34 tests: add end-to-end PDF generation check
Exercises the chromium/snappdf pipeline via POST /api/v1/preview with
no entity, which falls back to PdfMock's self-contained mock PDF -
verified this needs no client/invoice test data, since a fresh account
has neither. Asserts the response starts with the %PDF magic bytes.

This was previously the biggest untested path in the module: an actual
binary runtime dependency (chromium) rather than just config, and the
kind of thing that silently breaks in a minimal NixOS closure.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-26 15:50:24 -04:00
awkawbandClaude Sonnet 5 fd055f0cf9 nixos-module: remove nginx recommendedProxySettings, keep recommendedOptimisation
recommendedProxySettings only adds proxy_* directives (proxy_redirect,
proxy_connect_timeout, proxy_set_header Connection, etc.), which only
take effect on locations using proxy_pass. This vhost only ever uses
fastcgi_pass for PHP, so the setting was inert here.

recommendedOptimisation (sendfile/tcp_nopush/tcp_nodelay/keepalive)
does apply regardless of proxy_pass vs fastcgi_pass, so it's kept.
Verified via the full automated test suite (both proxy backends).

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-26 15:33:43 -04:00
5 changed files with 162 additions and 52 deletions
+31 -15
View File
@@ -34,23 +34,27 @@ This project uses `just` (see `justfile`) and Nix directly.
- `default.nix` — the interactive VM config (`nixosConfigurations.test-nginx`/`test-caddy` in `flake.nix`), used
for manual, click-around verification via `just build-vm [proxy]` / `just boot-vm`. It's also a function taking
`{ proxyServer ? "nginx" }`, forwarded straight through to `common.nix`, so `flake.nix` can instantiate one
`nixosConfigurations` entry per proxy backend. Adds the qemu-guest profile, a root login, and
`nixpkgs.config.allowUnfree = true` (required to build `pkgs.chromium`). It also sets a dummy `fileSystems."/"`
and disables `boot.loader.grub` purely to satisfy NixOS's built-in assertions when `nix flake check` evaluates
`nixosConfigurations` entry per proxy backend. Adds the qemu-guest profile and a root login. It also sets a
dummy `fileSystems."/"` and disables `boot.loader.grub` purely to satisfy NixOS's built-in assertions when
`nix flake check` evaluates
`config.system.build.toplevel` directly — the qemu-vm module overrides both with a tmpfs root and direct kernel
boot when actually building `config.system.build.vm` (what `just build-vm` uses), so these values are otherwise
unused.
unused. Also raises `virtualisation.cores`/`memorySize` to 2/2560 (qemu-vm's defaults, 1 core/1024MB, are too
tight for real interactive use — see the module architecture notes below on PDF generation's memory cost).
- `nixos-test.nix` — the automated test (`pkgs.testers.nixosTest`, exposed as `checks.<system>.invoiceninja` in
`flake.nix`). Builds its own `pkgs` (with `allowUnfree` baked in — `nixosTest` forbids setting `nixpkgs.config`
on a node once `pkgs` is externally supplied) and boots two nodes from `common.nix` — one with
`proxyServer = "nginx"`, one with `proxyServer = "caddy"` — so both proxy backends get exercised. The shared
`flake.nix`). Builds its own `pkgs` externally (needed regardless of config, since `nixosTest` itself is called
as `pkgs.testers.nixosTest`) and boots two nodes from `common.nix` — one with `proxyServer = "nginx"`, one with
`proxyServer = "caddy"` — so both proxy backends get exercised. The shared
assertion suite (factored into a `check_invoiceninja(machine, proxy_unit)` function in the test script) runs
against each node: waits for mysql/redis/phpfpm/queue-worker/proxy to be active, checks the
`invoiceninja-data-setup` marker files (`.initial-migration`, `.db-seeded`, `.admin-created`) to confirm the
`invoiceninja-data-setup` marker files (`.db-seeded`, `.admin-created`) to confirm the
one-time setup actually completed (not just that the oneshot unit exited), curls `/` for the app shell, exercises
the `invoiceninja-manage` wrapper, and logs in via `/api/v1/login` using the test admin credentials as an
end-to-end check that migrations/seeding/account creation all worked. This is the only automated test in the
repo — there is no separate unit-test suite.
the `invoiceninja-manage` wrapper, logs in via `/api/v1/login` using the test admin credentials as an
end-to-end check that migrations/seeding/account creation all worked, checks the full language list is
available (see the `cache:forget languages` note above), and exercises PDF generation end-to-end (POSTs to
`/api/v1/preview` with no entity, which falls back to a self-contained mock PDF built via the same
chromium/snappdf pipeline real invoices use, and checks for the `%PDF` magic bytes — no client/invoice setup
needed). This is the only automated test in the repo — there is no separate unit-test suite.
## Architecture
@@ -95,8 +99,11 @@ expect Nix to report the correct hash on a failed build (or use `nix-prefetch`/`
contents directly on disk in `dataDir`.
- `invoiceninja-data-setup.service` (oneshot, runs on every activation) does all first-run and idempotent setup:
clears bootstrap/opcode cache, rebuilds `.env`, rsyncs static storage into `dataDir`, symlinks the runtime
`app.php`, and guards one-time actions (`migrate`, `db:seed`, admin account creation) behind marker files
(`.initial-migration`, `.db-seeded`, `.admin-created`) in `dataDir`. `phpfpm-invoiceninja.service` and
`app.php`, runs `migrate --force` unconditionally on every activation (safe/idempotent — Laravel tracks
applied migrations in the DB and only runs new ones; this must NOT be marker-gated, since a later invoiceninja
version upgrade can ship new migrations that need to run on an already-migrated install), and guards the
remaining one-time actions (`db:seed`, admin account creation) behind marker files (`.db-seeded`,
`.admin-created`) in `dataDir`. `phpfpm-invoiceninja.service` and
`invoiceninja-queue-worker.service` both depend on it. It also unconditionally runs
`invoiceninja-manage cache:forget languages` on every activation, working around an upstream bug: migration
`2025_08_31_224356_additional_languages_af_ca_id.php` runs during `migrate` (before `db:seed`'s
@@ -122,8 +129,17 @@ expect Nix to report the correct hash on a failed build (or use `nix-prefetch`/`
regardless of `Host`). When `hostname == "localhost"`, the caddy vhost adds `serverAliases` for
`127.0.0.1:80`/`[::1]:80` too — otherwise visiting the VM via the IP literal instead of `localhost` matches no
vhost and Caddy silently returns an empty `200 OK` (a blank page with no error anywhere).
- Chromium is required at runtime for PDF generation (`PDF_GENERATOR=snappdf`, `SNAPPDF_CHROMIUM_PATH`) and is
added to the PHP-FPM and data-setup services' `path`.
- A headless browser is required at runtime for PDF generation (`PDF_GENERATOR=snappdf`,
`SNAPPDF_CHROMIUM_PATH`) and is added to the PHP-FPM and data-setup services' `path`. This is
`nixpkgs.playwright-driver`'s `chromium-headless-shell` component (Google's official "Chrome for Testing"
headless-only build, selected via `playwright-driver.browsers.override { ... }` with every other
browser/ffmpeg disabled) rather than `pkgs.chromium` — snappdf only ever shells out to
`$SNAPPDF_CHROMIUM_PATH --headless --print-to-pdf=...`, so the full browser UI/GPU stack `pkgs.chromium` ships
is dead weight; the headless-shell closure is ~1.45GB smaller and needs no `allowUnfree`. Each PDF request
still launches a full instance of the underlying browser engine (not a lightweight renderer), and PHP-FPM's
default `pm.max_children = 8` means up to 8 of these can run concurrently — on a memory-constrained deployment
this can exhaust RAM (confirmed: even a couple of concurrent PDF requests OOM-killed a 1024MB VM). Size
deployments accordingly, or lower `phpfpm.settings."pm.max_children"`.
### `flake.nix`
+21 -5
View File
@@ -47,8 +47,22 @@ let
extraConfig = "memory_limit = 1024M";
};
# Chromium is required for PDF invoice generation
chromium = pkgs.chromium;
# A headless browser is required for PDF invoice generation (PDF_GENERATOR=snappdf). snappdf
# just shells out to `$SNAPPDF_CHROMIUM_PATH --headless --print-to-pdf=... <url>`, so a full
# browser UI/GPU stack isn't needed -- Google's own "Chrome for Testing" headless-only build
# (as packaged by nixpkgs' playwright-driver) does the same job for a ~1.45GB smaller closure
# than pkgs.chromium, with no allowUnfree requirement.
chromiumHeadlessShellBrowsers = pkgs.playwright-driver.browsers.override {
withChromium = false;
withFirefox = false;
withWebkit = false;
withFfmpeg = false;
withChromiumHeadlessShell = true;
};
chromium = pkgs.runCommand "chromium-headless-shell" { } ''
mkdir -p $out/bin
ln -s ${chromiumHeadlessShellBrowsers}/chromium_headless_shell-*/chrome-headless-shell-linux64/chrome-headless-shell $out/bin/chromium
'';
extraPrograms = [ chromium ];
# Management script
@@ -486,7 +500,6 @@ in
recommendedTlsSettings = (if (cfg.hostname == "localhost") then false else true);
recommendedGzipSettings = true;
recommendedProxySettings = true;
recommendedOptimisation = true;
clientMaxBodySize = cfg.maxUploadSize;
@@ -689,8 +702,11 @@ in
# We cannot link the cache folder only because bootstrap folder needs to be writeable.
ln -sf ${invoiceninja}/bootstrap-static/app.php ${cfg.runtimeDir}/app.php
# Perform the first migration
[[ ! -f ${cfg.dataDir}/.initial-migration ]] && invoiceninja-manage migrate --force && touch ${cfg.dataDir}/.initial-migration
# Run migrations on every activation, not just the first: `migrate` is idempotent
# (Laravel tracks applied migrations in the DB and only runs new ones), and gating this
# behind a marker file meant migrations shipped by a later invoiceninja version upgrade
# would never run.
invoiceninja-manage migrate --force
# Seed database with records
# Necessary for languages, currencies, countries, etc.
+22 -8
View File
@@ -1,7 +1,9 @@
{ buildNpmPackage
{ lib
, buildNpmPackage
, npmHooks
, fetchFromGitHub
, fetchNpmDeps
, jq
, php82
, nodejs_22
, npm-lockfile-fix
@@ -12,13 +14,13 @@
let
pname = "invoiceninja";
version = "5.13.19";
version = "5.13.43";
src = fetchFromGitHub {
name = "${pname}";
owner = pname;
repo = pname;
rev = "v${version}";
hash = "sha256-pvZNTiGGX6OqOEqcDrDqqpLD1Ohc/TA4nIRjm2jEp74=";
hash = "sha256-F8lEyVi2EFRnkxWSDRIqsqrKmtCEUEee8lDDUZXoo0c=";
postFetch = ''
# add missing integrity fields to lockfile
@@ -26,13 +28,13 @@ let
'';
};
uiVersion = "03.05.2026.1";
uiVersion = "28.09.2026.1";
uiSrc = fetchFromGitHub {
name = "${pname}-ui";
owner = pname;
repo = "ui";
tag = uiVersion;
hash = "sha256-BTaWNHTE+9NvavUYs56DdJHmqXy36N5k/crj3rU2Npg=";
hash = "sha256-RLirh0Q5lhH2O7AEwoR4MdlELAgB00f9w6FfiBNhoho=";
postFetch = ''
# add missing integrity fields to lockfile
@@ -48,7 +50,7 @@ let
nodejs = nodejs_22;
npmDepsHash = "sha256-JA5TfXeg7iHVjQdjeU6SSD2JFSLISad8hPowaR1roQw=";
npmDepsHash = "sha256-7Boz+o4pQkPNMJREWgUMeKQa7avChqd+GpZoaXO0LqE=";
preConfigure = ''
sed -i 's/VITE_IS_TEST=true/VITE_IS_TEST=false/' .env.example
@@ -73,12 +75,13 @@ php82.buildComposerProject (finalAttrs: {
pname = "${pname}-composer";
nativeBuildInputs = [
jq
nodejs_22
npmHooks.npmConfigHook
npmHooks.npmBuildHook
];
vendorHash = "sha256-2CyWKlyCyoCt/WdY7Ta0oEXW2KIkGwhilHtzpBJ5Wnk=";
vendorHash = "sha256-zTo+unMIsv2XObjwiFpxTdUxVfZoXbMNviDrLMVMJI0=";
# Upstream composer.json has invalid license, webpatser/laravel-countries package is pointing
# to commit-ref, and php required in require and require-dev
@@ -89,7 +92,7 @@ php82.buildComposerProject (finalAttrs: {
npmDeps = fetchNpmDeps {
name = "${pname}-${version}-npm-deps";
inherit src;
hash = "sha256-WCadsQVj9eusYWJEOaEp5DIkVcbmPFGe6bmKi9NZDDg=";
hash = "sha256-uHbvhf8EWOBiyPQ9ieA4ynIUkIXWJEXZ7ZVsr4NXWtU=";
};
preConfigure = ''
@@ -103,6 +106,17 @@ php82.buildComposerProject (finalAttrs: {
# Remove JS/CSS build artifacts
rm -rf $out/node_modules
# composer.lock and vendor/composer/installed.json record each package's "dist.url" as the
# store path of the composer-repository mirror it was installed from, which otherwise drags
# that whole (500+MB) fixed-output derivation into the runtime closure for no reason -- it's
# provenance metadata, not something Laravel's package manifest (which only reads
# name/extra) or Composer\InstalledVersions (which reads vendor/composer/installed.php, not
# this file) consult at runtime.
jq '(.packages[]?.dist.url) |= "removed"' $out/vendor/composer/installed.json > $out/vendor/composer/installed.json.tmp
mv $out/vendor/composer/installed.json.tmp $out/vendor/composer/installed.json
jq '(.packages[]?.dist.url, .["packages-dev"][]?.dist.url) |= "removed"' $out/composer.lock > $out/composer.lock.tmp
mv $out/composer.lock.tmp $out/composer.lock
rm -r $out/bootstrap/cache
rm -rf $out/public/storage
+26 -3
View File
@@ -4,12 +4,12 @@
{
imports = [
(modulesPath + "/profiles/qemu-guest.nix")
# Not imported by default for a plain nixosSystem in this nixpkgs (only nixosTest nodes get it
# automatically) - needed explicitly so virtualisation.cores/memorySize below are real options.
(modulesPath + "/virtualisation/qemu-vm.nix")
(import ./common.nix { inherit proxyServer; })
];
# Required to build pkgs.chromium, which the module pulls in for PDF generation.
nixpkgs.config.allowUnfree = true;
# `nix flake check` evaluates config.system.build.toplevel directly, which asserts a real root
# filesystem and bootloader are configured. `just build-vm` builds config.system.build.vm
# instead, where the qemu-vm module overrides both of these with a tmpfs root and direct
@@ -26,4 +26,27 @@
};
services.resolved.enable = true;
# `just boot-vm` forwards the guest's port 80 to host port 8080 (QEMU_NET_OPTS in the justfile),
# but the backend builds absolute URLs (e.g. the stored company logo path) from APP_URL, which
# defaults to plain "http://localhost" with no port. The SPA's own relative URLs are fine (they
# inherit whatever port the browser is already on), but any backend-generated absolute URL then
# points at host port 80 - where nothing is listening - instead of the forwarded 8080, giving a
# silent net::ERR_CONNECTION_REFUSED for things like the logo image. Override to match the actual
# forwarded port.
services.invoiceninja.settings.APP_URL = "http://localhost:8080";
# The qemu-vm module's defaults (1 core, 1024MB) are too tight for real interactive use: MySQL +
# Redis + PHP-FPM + chromium (PDF generation, spawned per request - each is a full multi-process
# browser, not a lightweight renderer) running concurrently can exhaust 1GB fast, and this VM
# boots with panic_on_oom - so it's not a graceful slowdown, the whole VM goes down. Confirmed by
# reproducing a "logo upload hangs" report this way: at the 1024MB default, even a couple of
# concurrent PDF-generation requests triggered `redis-server invoked oom-killer` and a kernel
# panic. 2 cores / 2560MB comfortably handles 2 concurrent PDF generations alongside normal
# traffic (verified: ~1.4GB still free afterward); a deliberate flood of many concurrent PDF
# requests can still exhaust even this, since each is a full chromium instance - if that's a
# concern for your deployment, consider capping `phpfpm.settings."pm.max_children"` below its
# default of 8, or provisioning more RAM.
virtualisation.cores = 2;
virtualisation.memorySize = 2560;
}
+62 -21
View File
@@ -1,13 +1,7 @@
{ nixpkgsPath, system }:
let
# nixosTest forbids setting `nixpkgs.config` on a node once `pkgs` is externally supplied,
# so allowUnfree (required to build pkgs.chromium) has to be baked into this pkgs instance
# up front rather than set from within common.nix like the interactive VM does.
pkgs = import nixpkgsPath {
inherit system;
config.allowUnfree = true;
};
pkgs = import nixpkgsPath { inherit system; };
adminEmail = "admin@email.com";
adminPassword = pkgs.lib.strings.trim (builtins.readFile ./invoice_ninja_test_password);
@@ -23,6 +17,18 @@ pkgs.testers.nixosTest {
testScript = ''
import json
def api_get(machine, path, token):
return json.loads(machine.succeed(f"curl -sf http://localhost{path} -H 'X-API-TOKEN: {token}'"))
def api_post(machine, path, payload, token=None):
auth_header = f"-H 'X-API-TOKEN: {token}' " if token else ""
return json.loads(machine.succeed(
f"curl -sf -X POST http://localhost{path} "
+ auth_header
+ "-H 'Content-Type: application/json' "
+ f"-d '{json.dumps(payload)}'"
))
def check_invoiceninja(machine, proxy_unit):
machine.wait_for_unit("mysql.service")
machine.wait_for_unit("redis-invoiceninja.service")
@@ -33,9 +39,10 @@ pkgs.testers.nixosTest {
# phpfpm-invoiceninja/invoiceninja-queue-worker both Require= invoiceninja-data-setup,
# so systemd already waited for it to finish above. Confirm the one-time setup steps
# it guards actually completed, rather than just that the unit exited.
with subtest("one-time setup (migrate, seed, admin account) completed"):
machine.succeed("test -f /var/lib/invoiceninja/.initial-migration")
# it guards actually completed, rather than just that the unit exited. (Migrations run
# unconditionally on every activation rather than being marker-gated, so there's no
# marker file to check for them here.)
with subtest("one-time setup (seed, admin account) completed"):
machine.succeed("test -f /var/lib/invoiceninja/.db-seeded")
machine.succeed("test -f /var/lib/invoiceninja/.admin-created")
@@ -60,12 +67,9 @@ pkgs.testers.nixosTest {
machine.succeed("systemctl start --wait invoiceninja-scheduler.service")
with subtest("admin account can log in via the API"):
login_response = machine.succeed(
"curl -sf -X POST http://localhost/api/v1/login "
+ "-H 'Content-Type: application/json' "
+ "-d '{\"email\":\"${adminEmail}\",\"password\":\"${adminPassword}\"}'"
)
assert "${adminEmail}" in login_response, f"login did not succeed: {login_response}"
login_response = api_post(machine, "/api/v1/login", {"email": "${adminEmail}", "password": "${adminPassword}"})
assert "${adminEmail}" in json.dumps(login_response), f"login did not succeed: {login_response}"
token = login_response["data"][0]["token"]["token"]
with subtest("full language list is available, not just the af/ca/id migration's 3"):
# Regression check: migration 2025_08_31_224356_additional_languages_af_ca_id.php runs
@@ -73,13 +77,50 @@ pkgs.testers.nixosTest {
# does Cache::forever('languages', Language::all()) - permanently caching just the 3
# languages it sees at that moment. invoiceninja-data-setup works around this with
# `invoiceninja-manage cache:forget languages`; this confirms the workaround is in place.
token = json.loads(login_response)["data"][0]["token"]["token"]
statics = machine.succeed(
f"curl -sf http://localhost/api/v1/statics -H 'X-API-TOKEN: {token}'"
)
language_count = len(json.loads(statics)["languages"])
language_count = len(api_get(machine, "/api/v1/statics", token)["languages"])
assert language_count > 40, f"expected ~45 languages, got {language_count}"
with subtest("PDF generation works (chromium/snappdf)"):
# /api/v1/preview with no entity falls back to a self-contained mock PDF (no invoice/client
# setup needed), exercising the same chromium/snappdf pipeline real invoice PDFs use.
# This fallback only triggers when the company has no invoices/invitations at all yet
# (PreviewController::blankEntity), so this must run before the client/invoice/payment
# subtest below creates one.
machine.succeed(
"curl -sf -X POST http://localhost/api/v1/preview "
+ f"-H 'X-API-TOKEN: {token}' -H 'Content-Type: application/json' "
+ "-d '{}' -o /tmp/preview.pdf"
)
magic_bytes = machine.succeed("head -c 4 /tmp/preview.pdf")
assert magic_bytes == "%PDF", f"expected a PDF, got: {magic_bytes!r}"
with subtest("a client can be created, invoiced, and pay that invoice in full"):
client = api_post(machine, "/api/v1/clients", {"name": "Test Client"}, token)["data"]
invoice = api_post(machine, "/api/v1/invoices", {
"client_id": client["id"],
"line_items": [
{"quantity": 1, "cost": 100, "product_key": "test-product", "notes": "test line item"},
],
}, token)["data"]
assert invoice["amount"] == 100, f"unexpected invoice amount: {invoice}"
# Mirrors the normal workflow of sending an invoice to a client before it's paid,
# rather than relying on the fact that draft invoices also happen to be payable.
api_get(machine, f"/api/v1/invoices/{invoice['id']}/mark_sent", token)
payment = api_post(machine, "/api/v1/payments", {
"client_id": client["id"],
"amount": invoice["amount"],
"invoices": [{"invoice_id": invoice["id"], "amount": invoice["amount"]}],
}, token)["data"]
assert payment["applied"] == invoice["amount"], f"payment not fully applied: {payment}"
paid_invoice = api_get(machine, f"/api/v1/invoices/{invoice['id']}", token)["data"]
# status_id 4 == STATUS_PAID
assert paid_invoice["status_id"] == "4", f"invoice not marked paid: {paid_invoice}"
assert paid_invoice["balance"] == 0, f"invoice balance not zero: {paid_invoice}"
nginx.start()
caddy.start()