Skip to content

armgit / spree

Public Imported from GitHub
0

Open Source Platform for DTC, B2B Commerce, Marketplaces & Omnichannel. REST APIs, TypeScript SDKs, and production-ready Next.js storefront. Self-host it. Own your data. No vendor lock-in. Zero platform fees.

Address PR #14123 review comments

Damian Legawiecd8c8d64ab1a6
diff --git a/.claude/settings.local.json b/.claude/settings.local.json
index 3812559a68..b6d79e6d03 100644
--- a/.claude/settings.local.json
+++ b/.claude/settings.local.json
@@ -11,12 +11,9 @@
"WebSearch",
"Task",
"mcp__*",
- "Read(//private/tmp/**)",
- "Bash(cp server/.env /tmp/server-env-backup)",
"Bash(rm -rf server)",
"Bash(git clone *)",
- "Bash(rm -rf server/.git server/.gitignore)",
- "Bash(cp /tmp/server-env-backup server/.env)"
+ "Bash(rm -rf server/.git server/.gitignore)"
],
"deny": [
"Bash(sudo *)",
diff --git a/docs/developer/cli/quickstart.mdx b/docs/developer/cli/quickstart.mdx
index 85aaed0655..427a649ce6 100644
--- a/docs/developer/cli/quickstart.mdx
+++ b/docs/developer/cli/quickstart.mdx
@@ -54,6 +54,16 @@ Stop all services.
spree stop
```
+### `spree restart`
+
+Restart `web` and `worker` in place — same image, same volumes, fresh Rails process. Good for `config/initializers/*.rb` changes or anything Zeitwerk doesn't auto-reload.
+
+```bash
+spree restart
+```
+
+Not appropriate for Gemfile changes (use `spree bundle install` then `spree stop && spree dev`), Dockerfile / `.ruby-version` changes (use `spree build`), or compose file changes (use `spree stop && spree dev`).
+
### `spree update`
Pull the latest Spree Docker image and recreate containers. Migrations run automatically on startup.
@@ -62,6 +72,34 @@ Pull the latest Spree Docker image and recreate containers. Migrations run autom
spree update
```
+### `spree build`
+
+Rebuild the dev image after Dockerfile or `.ruby-version` changes. Only relevant after `spree eject`.
+
+```bash
+spree build
+spree build --reset-bundle # Also drop the bundle_cache volume so gems re-seed
+spree build --yes # Skip confirmation prompts (for CI)
+```
+
+`spree bundle add` does NOT require a rebuild — gems land in the `bundle_cache` volume which survives image rebuilds.
+
+`--reset-bundle` only drops `bundle_cache`. Postgres, Redis, Meilisearch, and storage volumes are preserved.
+
+### `spree upgrade`
+
+Walk a Spree version upgrade end-to-end inside the dev stack. Runs `bundle update`, applies pending migrations, then walks the version-specific data backfills from the upgrade manifest.
+
+```bash
+spree upgrade # Detect target, run the full sequence interactively
+spree upgrade --plan # Print the plan without executing (DRY_RUN=1)
+spree upgrade --to 5.5 # Cap the walk at a specific target version
+spree upgrade --step <id> # Re-run a single step idempotently (skips bundle + migrate)
+spree upgrade --yes # Skip the per-step prompts
+```
+
+Production upgrades only need the third stage — `bundle install` and `db:migrate` happen in your deploy pipeline. The CLI runs all three for local development. See [Upgrades](/developer/upgrades) for the manual production path.
+
### `spree eject`
Switch from the prebuilt Docker image to building from your local `backend/` directory. This lets you customize the Rails app — add gems, override models, add migrations, etc.
@@ -135,3 +173,113 @@ spree api-key create --name "Storefront" --type publishable # Store API key
spree api-key create --name "Admin" --type secret # Admin API key
spree api-key revoke <id> # Revoke a key
```
+
+## Dev workflow commands
+
+These mirror the Rails commands you'd run if you were running the app directly on your host — but each is routed through `docker compose exec` against the running stack.
+
+### `spree generate`
+
+Run a Rails generator inside the container. Same arguments as `bin/rails generate`.
+
+```bash
+spree generate model Brand name:string slug:string
+spree generate migration AddPositionToProducts position:integer
+spree generate spree:api_resource Brand name:string slug:string # The Spree v3-conformant generator
+```
+
+See [Generators](/developer/customization/generators) for the Spree-specific generators (`spree:model`, `spree:api_resource`) that extend Rails' built-in ones with Spree conventions.
+
+### `spree migrate`
+
+Install pending Spree migrations from gems, then run `db:migrate`.
+
+```bash
+spree migrate
+```
+
+Equivalent to `spree rails spree:install:migrations db:migrate` — a shortcut for the canonical post-update sequence.
+
+### `spree db:reset`
+
+Drop, recreate, and seed the database. **Destructive** — wipes everything. Asks for confirmation.
+
+```bash
+spree db:reset
+spree db:reset --yes # Skip confirmation (for CI)
+```
+
+### `spree db:console`
+
+Open a database console (`bin/rails dbconsole`) connected to the dev Postgres.
+
+```bash
+spree db:console
+```
+
+### `spree routes`
+
+Show Rails routes. All flags pass through to `bin/rails routes`.
+
+```bash
+spree routes
+spree routes -g products # Filter by pattern
+spree routes -c Spree::Api::V3::Store::ProductsController # Filter by controller
+spree routes --expanded
+```
+
+## Running things inside the container
+
+These are passthrough commands — anything after the subcommand reaches the inner Rails/Bundle/Rake invocation as-is. Useful for ad-hoc commands that don't have a dedicated wrapper.
+
+### `spree exec`
+
+Run an arbitrary command inside the `web` container.
+
+```bash
+spree exec ls -la
+spree exec env
+spree exec ruby -v
+```
+
+### `spree rails`
+
+Run a `bin/rails` command.
+
+```bash
+spree rails c # Console (alias of `spree console`)
+spree rails routes -g products # Alias of `spree routes -g products`
+spree rails runner "puts Spree::Product.count"
+spree rails new_subcommand arg1 arg2
+```
+
+### `spree bundle`
+
+Run a `bundle` command. Gems land in the `bundle_cache` volume (no image rebuild needed).
+
+```bash
+spree bundle install
+spree bundle add stripe
+spree bundle update spree_core
+spree bundle outdated
+```
+
+### `spree rake`
+
+Run a Rake task.
+
+```bash
+spree rake -T # List tasks
+spree rake spree:upgrade # Run the version-specific data backfills
+spree rake spree:upgrade DRY_RUN=1 # Same, plan only
+```
+
+### `spree task`
+
+Shortcut for `bin/rails <task>` (Spree-style rake tasks registered as Rails tasks).
+
+```bash
+spree task load_sample_data
+spree task spree:price_history:seed
+spree task spree:price_history:prune
+```
diff --git a/docs/developer/upgrades/5.4-to-5.5.mdx b/docs/developer/upgrades/5.4-to-5.5.mdx
index 5d5c024e6d..16d22f87c6 100644
--- a/docs/developer/upgrades/5.4-to-5.5.mdx
+++ b/docs/developer/upgrades/5.4-to-5.5.mdx
@@ -7,7 +7,7 @@ description: Step-by-step guide to upgrading a Spree 5.4 application to Spree 5.
Before proceeding to upgrade, please ensure you're at [Spree 5.4](/developer/upgrades/5.3-to-5.4).
</Info>
-The upgrade is usually
+The upgrade is usually completed in five steps:
1. **Update the Ruby gems** which power Spree API
2. **Run database migrations** — to migrate your existing schema to the new version
diff --git a/docs/plans/5.5-agent-skills.md b/docs/plans/5.5-agent-skills.md
index 64dea76990..45e2bb91a7 100644
--- a/docs/plans/5.5-agent-skills.md
+++ b/docs/plans/5.5-agent-skills.md
@@ -35,7 +35,7 @@ This is **for Spree users, not monorepo contributors.** Skills reference `backen
### Standalone repo layout (github.com/spree/agent-skills)
-```
+```text
agent-skills/
├── .claude-plugin/
│ └── plugin.json # Claude Code plugin manifest (name: "spree", version, description)
@@ -136,7 +136,7 @@ The scaffold **does not bundle skills** any more. Instead:
2. Removes `src/skills.ts` and the `writeClaudeArtifacts(projectDir)` call from `src/scaffold.ts`.
3. Adds a post-scaffold hint to the success message:
- ```
+ ```text
Agent skills (optional)
npx skills add spree/agent-skills
# Adds 23 Spree skills to whichever AI agent(s) you use
diff --git a/docs/plans/spree-dev-cli-and-generators.md b/docs/plans/spree-dev-cli-and-generators.md
index a4beb88bca..25513d013d 100644
--- a/docs/plans/spree-dev-cli-and-generators.md
+++ b/docs/plans/spree-dev-cli-and-generators.md
@@ -8,7 +8,7 @@
`create-spree-app` scaffolds:
-```
+```text
my-spree-app/
├── backend/ # the Rails app (cloned from spree-starter)
│ ├── Dockerfile
@@ -399,7 +399,7 @@ Today's data tasks are *implicitly* version-scoped — `spree/core/lib/tasks/cha
The mapping uses **directory convention + tiny per-version manifest**, with the rake tasks themselves unchanged:
-```
+```text
core/lib/spree/upgrades/
├── 5_4_to_5_5/
│ ├── manifest.yml
@@ -457,7 +457,7 @@ Why a YAML manifest rather than a Ruby DSL or rake task metadata: the manifest i
### CLI surface
-```
+```bash
spree upgrade # detect from → to from Gemfile.lock and run the manifest interactively
spree upgrade --to 5.5 # explicit target (when you want to step rather than jump)
spree upgrade --plan # print the plan, don't execute (good for PR description / runbook)
@@ -506,7 +506,7 @@ The CLI's `spree generate` proxies to `bin/rails g spree:<name>`. The actual sca
### Generator surface
-```
+```text
spree:api_resource NAME [field[:type[:index|:uniq|:null|:default=…]]…] [options]
```
@@ -532,7 +532,7 @@ spree:api_resource NAME [field[:type[:index|:uniq|:null|:default=…]]…] [opti
For `spree generate api_resource Brand name:string:uniq active:boolean:default=true`:
-```
+```text
core/app/models/spree/brand.rb (owned-once)
core/db/migrate/<ts>_create_spree_brands.rb (append-only)
@@ -749,6 +749,6 @@ These were open and now aren't, written down so future-me doesn't reopen them:
1. **~~`db/schema.rb` write permissions from the dev container.~~ Resolved.** Originally framed as a real problem — turns out it isn't. Every CLI command goes through `docker compose exec web …`, which means file writes happen as the container's default user (UID 1000, `rails`). On macOS, Docker Desktop and OrbStack do UID mapping at the VM boundary, so bind-mount writes always appear as the host user regardless of container UID. On Linux, the host user is typically UID 1000 (Ubuntu/Debian/Fedora default), so the UIDs match. The bundle/tmp/storage volumes are docker-managed (not bind-mounted), so the host never touches them — initial UID-1000 ownership inside the container is exactly what's wanted. The compose stays simple: no `user:` override, no init service, no entrypoint chown. The CLI exporting `HOST_UID`/`HOST_GID` is only needed for the Linux multi-user edge case (host user not UID 1000) and is deferred until someone actually hits it.
-2. **`spree-starter` PR.** The new compose ships from `spree-starter`'s `backend/` template; the local clone is at `/Users/damian/github/spree-starter`. A sibling PR lands the new `docker-compose.dev.yml` there. (The monorepo `server/` directory is dev-only for the gems and doesn't need separate consideration.)
+2. **`spree-starter` PR.** The new compose ships from `spree-starter`'s `backend/` template. A sibling PR lands the new `docker-compose.dev.yml` there. (The monorepo `server/` directory is dev-only for the gems and doesn't need separate consideration.)
3. **Idempotency helper for backfill tasks.** The existing `backfill_*` family will get `where.not(<col>: nil)` guards as part of the Track 2c prep audit. Whether to extract a shared helper (`Spree::Upgrades::Idempotent.backfill(scope, &block)`) or keep guards per-task is style preference; lean per-task — explicit skip conditions read better than a clever wrapper for something that runs once per upgrade.
diff --git a/package.json b/package.json
index 077fd7b7c0..9d3c3882a9 100644
--- a/package.json
+++ b/package.json
@@ -16,10 +16,10 @@
"server:start": "SPREE_PATH=$PWD docker compose -f server/docker-compose.dev.yml -f scripts/docker-compose.edge.yml up -d --force-recreate web worker",
"server:restart": "SPREE_PATH=$PWD docker compose -f server/docker-compose.dev.yml -f scripts/docker-compose.edge.yml restart web worker",
"server:stop": "SPREE_PATH=$PWD docker compose -f server/docker-compose.dev.yml -f scripts/docker-compose.edge.yml down",
- "server:console": "cd server && pnpm exec spree console",
- "server:logs": "cd server && pnpm exec spree logs",
- "server:seed": "cd server && pnpm exec spree rails db:seed",
- "server:load_sample_data": "cd server && pnpm exec spree task load_sample_data"
+ "server:console": "cd server && spree console",
+ "server:logs": "cd server && spree logs",
+ "server:seed": "cd server && spree rails db:seed",
+ "server:load_sample_data": "cd server && spree task load_sample_data"
},
"devDependencies": {
"@biomejs/biome": "^2.4.14",
diff --git a/packages/cli/src/commands/build.ts b/packages/cli/src/commands/build.ts
index a6e7477249..a330f240d7 100644
--- a/packages/cli/src/commands/build.ts
+++ b/packages/cli/src/commands/build.ts
@@ -2,20 +2,27 @@ import fs from 'node:fs'
import path from 'node:path'
import * as p from '@clack/prompts'
import type { Command } from 'commander'
+import { execa } from 'execa'
import pc from 'picocolors'
import { detectProject } from '../context.js'
import { dockerCompose } from '../docker.js'
-// Rebuild the dev image. Needed after Dockerfile or .ruby-version changes;
-// `spree bundle add` does NOT require this — gems land in the bundle_cache
-// volume which survives image rebuilds.
-//
-// --reset-bundle wipes the bundle_cache volume so the new image's gem
-// baseline gets re-seeded. Use when you bumped .ruby-version (gems compiled
-// against the old Ruby won't load against the new one).
-//
-// Targets docker-compose.dev.yml explicitly when present, so this works
-// before `spree eject` too (eject collapses the two files into one).
+/**
+ * Register the `spree build` command on the CLI program.
+ *
+ * Rebuilds the dev image. Needed after Dockerfile or .ruby-version changes;
+ * `spree bundle add` does NOT require this — gems land in the bundle_cache
+ * volume which survives image rebuilds.
+ *
+ * `--reset-bundle` wipes the bundle_cache volume so the new image's gem
+ * baseline gets re-seeded. Use when you bumped .ruby-version (gems compiled
+ * against the old Ruby won't load against the new one).
+ *
+ * Targets docker-compose.dev.yml explicitly when present, so this works
+ * before `spree eject` too (eject collapses the two files into one).
+ *
+ * @param program - The Commander CLI program to register the command on.
+ */
export function registerBuildCommand(program: Command): void {
program
.command('build')
@@ -41,10 +48,20 @@ export function registerBuildCommand(program: Command): void {
const s = p.spinner()
s.start('Wiping bundle_cache volume...')
- // `docker compose down -v` is portable (no project-name guessing) but
- // also tears down all containers. Acceptable: rebuilding the image
- // means the user is recreating containers anyway.
- await dockerCompose([...composeArgs, 'down', '-v'], ctx.projectDir, { stdio: 'ignore' })
+ // Stop containers without -v: `down -v` would also wipe postgres,
+ // redis, meilisearch, and storage volumes — destroying dev data
+ // and uploads. We only want to drop bundle_cache.
+ await dockerCompose([...composeArgs, 'down'], ctx.projectDir, { stdio: 'ignore' })
+ // Resolve the compose project name so we can target the namespaced
+ // volume directly. Compose derives the project name from the dir
+ // (or COMPOSE_PROJECT_NAME); `compose ls --format json` returns
+ // whatever it actually uses for THIS compose file.
+ const projectName = await resolveComposeProjectName(composeArgs, ctx.projectDir)
+ await execa('docker', ['volume', 'rm', `${projectName}_bundle_cache`], {
+ cwd: ctx.projectDir,
+ stdio: 'ignore',
+ reject: false, // Volume may not exist on first build — that's fine
+ })
s.stop('bundle_cache volume wiped.')
}
@@ -74,3 +91,30 @@ function composeFileArgs(projectDir: string): string[] {
const devCompose = path.join(projectDir, 'docker-compose.dev.yml')
return fs.existsSync(devCompose) ? ['-f', 'docker-compose.dev.yml'] : []
}
+
+// Returns the compose project name (e.g. `my-spree-app` or whatever
+// COMPOSE_PROJECT_NAME / dir basename resolves to). Used to target the
+// `<project>_bundle_cache` named volume without guessing.
+//
+// Falls back to the project directory's basename if the compose ls call
+// doesn't find an entry — matches Compose's own default behavior.
+async function resolveComposeProjectName(
+ composeArgs: string[],
+ projectDir: string,
+): Promise<string> {
+ try {
+ const { stdout } = await execa(
+ 'docker',
+ ['compose', ...composeArgs, 'config', '--format', 'json'],
+ { cwd: projectDir },
+ )
+ const parsed = JSON.parse(stdout) as { name?: string }
+ if (parsed.name) return parsed.name
+ } catch {
+ // Fall through to basename fallback.
+ }
+ return path
+ .basename(projectDir)
+ .toLowerCase()
+ .replace(/[^a-z0-9_-]/g, '')
+}
diff --git a/packages/cli/src/commands/restart.ts b/packages/cli/src/commands/restart.ts
index b6ea27315e..4d545a8d2d 100644
--- a/packages/cli/src/commands/restart.ts
+++ b/packages/cli/src/commands/restart.ts
@@ -23,7 +23,14 @@ export function registerRestartCommand(program: Command): void {
const s = p.spinner()
s.start('Restarting web + worker...')
- await dockerCompose(['restart', 'web', 'worker'], ctx.projectDir)
- s.stop('Services restarted.')
+ try {
+ await dockerCompose(['restart', 'web', 'worker'], ctx.projectDir)
+ s.stop('Services restarted.')
+ } catch (error) {
+ // Without this, the spinner stays "live" when the top-level
+ // handler prints the error — output lands under a stale prompt.
+ s.stop('Restart failed.')
+ throw error
+ }
})
}
diff --git a/packages/cli/src/commands/routes.ts b/packages/cli/src/commands/routes.ts
index ea44e3297e..a383ce0483 100644
--- a/packages/cli/src/commands/routes.ts
+++ b/packages/cli/src/commands/routes.ts
@@ -2,10 +2,17 @@ import type { Command } from 'commander'
import { detectProject } from '../context.js'
import { dockerComposeExec } from '../docker.js'
-// `bin/rails routes [-g pattern] [-c controller] [--expanded]` passthrough.
-// spree routes
-// spree routes -g products
-// spree routes -c Spree::Api::V3::Store::ProductsController
+/**
+ * Register the `spree routes` command on the CLI program.
+ *
+ * Passthrough to `bin/rails routes [-g pattern] [-c controller] [--expanded]`.
+ *
+ * spree routes
+ * spree routes -g products
+ * spree routes -c Spree::Api::V3::Store::ProductsController
+ *
+ * @param program - The Commander CLI program to register the command on.
+ */
export function registerRoutesCommand(program: Command): void {
program
.command('routes')
diff --git a/packages/cli/src/commands/upgrade.ts b/packages/cli/src/commands/upgrade.ts
index 20ce91ad48..3d75fa5b76 100644
--- a/packages/cli/src/commands/upgrade.ts
+++ b/packages/cli/src/commands/upgrade.ts
@@ -4,31 +4,38 @@ import pc from 'picocolors'
import { detectProject } from '../context.js'
import { dockerComposeExec } from '../docker.js'
-// `spree upgrade` is a dev-friendly sequencer around the real upgrade flow:
-//
-// 1. bundle update (universal — bumps spree_core etc.)
-// 2. db:migrate (universal — applies the new gem's migrations)
-// 3. spree:upgrade (version-specific — runs the rake-only manifest)
-//
-// On production, step 3 is what your deploy pipeline runs after `bundle
-// install` + `db:migrate` happen there. Steps 1 and 2 are handled by your
-// platform (Heroku release phase, K8s init container, Render auto-migrate,
-// Capistrano deploy hook), not by this CLI.
-//
-// Local dev: this command runs all three for you. On a clean stack the
-// total time is dominated by `bundle update` (network + native ext build).
-//
-// Flags map to env vars on the inner rake task so the same args work in
-// both surfaces:
-// --plan DRY_RUN=1 (also skips steps 1 + 2)
-// --step <id> STEP=<id>
-// --to <version> TO=<version>
-// --yes no flag — skips the per-step prompt
-//
-// What this CLI deliberately does NOT do: schedule cron jobs, review
-// breaking changes, tune reservation TTLs. Those are in
-// docs/developer/upgrades/<v>.mdx and the manifest's `notes` text; we
-// print the docs URL at the end so the operator knows where to look.
+/**
+ * Register the `spree upgrade` command on the CLI program.
+ *
+ * Dev-friendly sequencer around the real upgrade flow:
+ *
+ * 1. bundle update (universal — bumps spree_core etc.)
+ * 2. db:migrate (universal — applies the new gem's migrations)
+ * 3. spree:upgrade (version-specific — runs the rake-only manifest)
+ *
+ * On production, step 3 is what your deploy pipeline runs after `bundle
+ * install` + `db:migrate` happen there. Steps 1 and 2 are handled by your
+ * platform (Heroku release phase, K8s init container, Render auto-migrate,
+ * Capistrano deploy hook), not by this CLI.
+ *
+ * Local dev: this command runs all three for you. On a clean stack the
+ * total time is dominated by `bundle update` (network + native ext build).
+ *
+ * Flags map to env vars on the inner rake task so the same args work in
+ * both surfaces:
+ *
+ * --plan DRY_RUN=1 (also skips steps 1 + 2)
+ * --step <id> STEP=<id>
+ * --to <version> TO=<version>
+ * --yes no flag — skips the per-step prompt
+ *
+ * What this CLI deliberately does NOT do: schedule cron jobs, review
+ * breaking changes, tune reservation TTLs. Those are in
+ * docs/developer/upgrades/<v>.mdx and the manifest's `notes` text; we
+ * print the docs URL at the end so the operator knows where to look.
+ *
+ * @param program - The Commander CLI program to register the command on.
+ */
export function registerUpgradeCommand(program: Command): void {
program
.command('upgrade')
@@ -59,7 +66,9 @@ export function registerUpgradeCommand(program: Command): void {
// Closing reminder: the manifest only covers rake-runnable steps.
// Anything human (cron scheduling, breaking-change review) is in
// the upgrade doc — point there so the operator doesn't forget.
- printPostUpgradeReminder()
+ // Skipped on --plan because nothing actually ran — the "next steps"
+ // reminder would be misleading.
+ if (!flags.plan) printPostUpgradeReminder()
})
}
diff --git a/packages/cli/src/index.ts b/packages/cli/src/index.ts
index 4d678fa35d..8505285fd5 100644
--- a/packages/cli/src/index.ts
+++ b/packages/cli/src/index.ts
@@ -28,7 +28,7 @@ import { registerUserCommand } from './commands/user.js'
const program = new Command()
.name('spree')
.description('CLI for managing Spree Commerce projects')
- .version('2.0.0-beta.6')
+ .version('2.0.0')
// Required by passThroughOptions on subcommands (exec/rails/bundle/rake/task)
// so flags like `ls -la` or `bin/rails routes -g foo` reach the inner command
// instead of being parsed as options of the spree subcommand.
diff --git a/spree/core/lib/generators/spree/api_resource/api_resource_generator.rb b/spree/core/lib/generators/spree/api_resource/api_resource_generator.rb
index ddb029a5ec..74e09401fd 100644
--- a/spree/core/lib/generators/spree/api_resource/api_resource_generator.rb
+++ b/spree/core/lib/generators/spree/api_resource/api_resource_generator.rb
@@ -326,7 +326,14 @@ module Spree
# the namespace's opening line.
opening = " namespace :#{namespace} do\n"
new_block = "#{BEGIN_MARKER}\n#{route_line}\n#{END_MARKER}\n\n"
- File.write(file, content.sub(opening, opening + new_block))
+ updated = content.sub(opening, opening + new_block)
+ if updated == content
+ # The namespace anchor wasn't found — likely a non-default
+ # routes.rb. Warn rather than silently claim success.
+ say_status :skip, "routes.rb (#{namespace}: namespace block not found — add manually)", :yellow
+ return
+ end
+ File.write(file, updated)
say_status :inject, "routes.rb (#{namespace}: sentinels + #{route_line.strip})", :green
end
end
diff --git a/spree/core/lib/generators/spree/model/model_generator.rb b/spree/core/lib/generators/spree/model/model_generator.rb
index 76e35d17ea..6909dd63ef 100644
--- a/spree/core/lib/generators/spree/model/model_generator.rb
+++ b/spree/core/lib/generators/spree/model/model_generator.rb
@@ -15,10 +15,15 @@ module Spree
[File.expand_path('templates', __dir__), *superclass.source_paths]
end
+ # Default is the literal expression `Spree.base_class`, not a class
+ # constant — Spree.base_class is the swappable lookup that lets apps
+ # override the inheritance root via config. The template emits this
+ # verbatim. Override with --parent=Spree::SomeOtherClass when you
+ # genuinely need to inherit from something else.
class_option :parent,
type: :string,
- default: 'Spree::Base',
- desc: 'The parent class for the generated model'
+ default: 'Spree.base_class',
+ desc: 'The parent class expression for the generated model'
class_option :id_prefix,
type: :string,
@@ -34,7 +39,7 @@ module Spree
default: false,
desc: 'Include Spree::Metafields and Spree::Metadata concerns'
- desc 'Creates a new Spree model with prefixed IDs and Spree::Base parent'
+ desc 'Creates a new Spree model with prefixed IDs and Spree.base_class parent'
# Override to prevent module file from being created
def create_module_file
diff --git a/spree/core/lib/generators/spree/model/templates/model.rb.tt b/spree/core/lib/generators/spree/model/templates/model.rb.tt
index d3465bbcd1..f7af89dded 100644
--- a/spree/core/lib/generators/spree/model/templates/model.rb.tt
+++ b/spree/core/lib/generators/spree/model/templates/model.rb.tt
@@ -1,5 +1,5 @@
module Spree
- class <%= class_name.demodulize %> < <%= parent_class_name.classify %>
+ class <%= class_name.demodulize %> < <%= parent_class_name %>
<% if metafields? -%>
include Spree::Metafields
include Spree::Metadata
@@ -12,7 +12,11 @@ module Spree
has_prefix_id :<%= id_prefix %>
<% attributes.select(&:reference?).each do |attribute| -%>
- belongs_to :<%= attribute.name %><%= ", polymorphic: true" if attribute.polymorphic? %>, class_name: 'Spree::<%= attribute.name.camelize %>'
+<% if attribute.polymorphic? -%>
+ belongs_to :<%= attribute.name %>, polymorphic: true
+<% else -%>
+ belongs_to :<%= attribute.name %>, class_name: 'Spree::<%= attribute.name.camelize %>'
+<% end -%>
<% end -%>
<% attributes.select(&:rich_text?).each do |attribute| -%>
has_rich_text :<%= attribute.name %>
@@ -37,8 +41,8 @@ module Spree
<% end -%>
<% end -%>
<% ransackable = attributes.reject { |a| a.reference? || a.attachment? || a.attachments? || a.rich_text? || a.token? || a.password_digest? }.map(&:name) -%>
-<% if ransackable.any? -%>
self.whitelisted_ransackable_attributes = %w[<%= ransackable.join(' ') %>]
-<% end -%>
+ self.whitelisted_ransackable_associations = %w[]
+ self.whitelisted_ransackable_scopes = %w[]
end
end