**Breaking changes**
The `outDir` has been deprecated and wont be used anymore, instead all the path are computed internally following these rules
- if admin is not `disabled` and the `build` command is run without the `--admin-only` flag, then the admin output dir will be `.medusa/server/public/admin` and it will be served from that same location from the medusa instance.
- if admin is not `disabled` and the `build` command is run with the `--admin-only` flag, then the admin output dir will be `.medusa/admin` with the purpose of deploying the admin separately. ⚠️ (expect to receive a warning log)
- if the admin is `disabled` and the `build` command is run with the `--admin-only` flag, then fallback to rule number 2
| admin enabled | medusa build --admin-only | output dir |
|---|---|---|
| true | true | `.medusa/admin` ⚠️ (expect to receive a warning log) |
| true | false | `.medusa/server/public/admin` |
| false | true | `.medusa/admin` |
| false | false | none |
```diff
// medusa-config.ts
{
// ...
admin: {
- outDir: 'some/path'
}
}
```
cc @kasperkristensen @sradevski @olivermrbl
**What**
When resolving the modules through an array, if the module does not have a joiner config to rely on (not `queryable`, not using the `Module` util), then the key should be provided to register the module in the container. Eventually, the module author should provide that key to be used.
The only exception is for the medusa modules where we have a map that allows us to resolve that key automatically from the path of the resolve value
* chore: add compare_at_unit_price when price list price is retrieved
* chore: add test for update item + more fixes along the way
* chore: fix tests
* chore: add refresh spec
* Apply suggestions from code review
Co-authored-by: Adrien de Peretti <adrien.deperetti@gmail.com>
* chore: use undefined checker
* chore: switch to map
---------
Co-authored-by: Adrien de Peretti <adrien.deperetti@gmail.com>
**What**
- validate that variants are unique with respect to options on product update/create and variant update/create
- validate that the product has options upon creation
- ensure variants have the same number of option values as the product has options
- admin error handling
- update tests
---
FIXES FRMW-2707 CC-556
**What**
- remove cascade delete of inventory items on product delete
- implement inventory deletion in product/variant delete workflows with checks:
- product/variant cannot be deleted if there are reservations associated with their inventory items
- inventory item will be cascade deleted if it's not used by other variants (that are not being deleted in the current flow)
---
FIXES CC-581 CC-582
**What**
The module service name case has changed and the hash generation was performed on the non lower cased version of it while after the hash generation everything is based on the lower case version of the generated table name form the service names leading to different hash computation. This pr update the link migration table to adjust the to/from module value with its new value as well as generating the hash from the lower case version of the computed table name
Co-authored-by: Carlos R. L. Rodrigues <37986729+carlos-r-l-rodrigues@users.noreply.github.com>
Fixes: FRMW-2743
This PR extracts and removes the `additional_data` from the workflow input before calling the steps and hence the `additional_data` is not passed down to the service layer.
However, this bug has made me discover one inconsistency in the input structure of certain workflows.
✅ **Following is the input structure of the `updateProductsWorkflow`**. Here, we accept the `products` and the `additional_data` as two separate top-level properties.
```ts
// Shape
export type UpdateProductsWorkflowInputSelector = {
selector: ProductTypes.FilterableProductProps
update: Omit<ProductTypes.UpdateProductDTO, "variants"> & {
sales_channels?: { id: string }[]
variants?: UpdateProductVariantWorkflowInputDTO[]
}
} & AdditionalData
// Calling the workflow
const { result } = await updateProductsWorkflow(req.scope).run({
input: {
selector: { id: req.params.id },
update,
additional_data,
},
})
```
❌ **Following in the input structure of the `updateCartWorflow`**. In this case, we are accepting the cart properties at the top-level, hence the `additional_data` is merged within those properties, increasing the chance of passing it down to the service layer by mistake.
```ts
// Shape
WorkflowData<UpdateCartWorkflowInputDTO & AdditionalData>
// Calling the workflow
await workflow.run({
input: {
...req.validatedBody // Additional data is part of the validatedBody,
id: req.params.id,
},
})
```
Ideally, the input of `updateCartWorkflow` should look as follows.
```ts
WorkflowData<{ cart: UpdateCartWorkflowInputDTO } & AdditionalData>
```
**What**
- Adds the ability to do global searches from cmd + k in the admin.
- The solution is temporary, until we have a proper search API.
**Note**
I have deviated a bit from the design, due to the constraints of this temporary solution:
- We don't have nested items, such as showing variants under a product (don't think having a proper search API will make this any easier, and not entirely sure how we would handle this for cases where a query returns multiple products, which is the only case that is designed)
- I have added a "Load {{count}} more" button instead of doing infinite scrolling, I am assuming the later is the intended behaviour based on the design file, but with 20+ sources of data changing so often it was resulting in some weird behaviours, so settled for the simpler approach for this temporary solution.
- Removed the "Details" label on search results as it seemed a bit repetitive
- I haven't added icons for the different types of search results, as there are only a couple of examples in the design doc, and I wasn't sure what to pick for all the different types of results. If we want to add icons, then I think it's something we can add when we revisit this later, but think its fine to omit, as each group of results is labeled, so they are easy to tell apart.
Resolves CC-574
**What**
- fix `prepareConfirmInventory` to account for inventory kit items
- _note: this step is reused in the complete cart and all RMA flows_
- properly remove reservations for items that are removed from the order edit
- invalidate inventory/reservations cache when order edit is confirmed
---
https://github.com/user-attachments/assets/f12e9198-0718-4c08-bd81-efc536eca146
---
FIXES CC-565
Fixes: FRMW-2742
In this PR, we fix the build output of the backend source code, which eliminates a lot of magic between the development and production environments.
Right now, we only compile the source files from the `src` directory and write them within the `dist` directory.
**Here's how the `src` directory with a custom module looks like**
```
src
├── modules
│ └── hello
│ ├── index.ts
```
**Here's the build output**
```
dist
├── modules
│ └── hello
│ ├── index.js
```
Let's imagine a file at the root of your project (maybe the `medusa-config.js` file) that wants to import the `modules/hello/index` file. How can we ensure that the import will work in both the development and production environments?
If we write the import targeting the `src` directory, it will break in production because it should target the `dist` directory.
## Solution
The solution is to compile everything within the project and mimic the file structure in the build output, not just the `src` directory.
**Here's how the fixed output should look like**
```
dist
├── src
│ ├── modules
│ │ └── hello
│ │ ├── index.js
├── medusa-config.js
├── yarn.lock
├── package.json
```
If you notice carefully, we also have `medusa-config.js`, `yarn.lock`, and `package.json` within the `dist` directory. We do so to create a standalone built application, something you can copy/paste to your server and run without relying on the original source code.
- This results in small containers since you are not copying unnecessary files.
- Clear distinction between the development and the production code. If you want to run the production server, then `cd` into the `dist` directory and run it from there.
## Changes in the PR
- Breaking: Remove the `dist` and `build` folders. Instead, write them production artefacts within the `.medusa` directory as `.medusa/admin` and `.medusa/server`.
- Breaking: Change the output of the `.medusa/server` folder to mimic the root project structure.
- Refactor: Remove `Symbol.for("ts-node.register.instance")]` check to find from where to load the source code.
- Refactor: Use `tsc` for creating the production build. This ensures we respect `tsconfig` settings when creating the build and also perform type-checking.
Co-authored-by: Adrien de Peretti <25098370+adrien2p@users.noreply.github.com>
**What**
The extra serialization check hapen to break the serialization chain by reusing already serialized entities when they have been serialized from a different parents sequence
What
- Store result of cart-completion workflow for three days by default
- This enables the built-in idempotency mechanism to kick-in, provided the same transaction ID is used on workflow executions
- Return order from cart-completion workflow if the cart has already been completed
- In case transaction ID is not used on workflow executions, we still only want to complete a cart once
**What**
If the country on the shipping address changes to a country without an associated tax region, we clear the tax lines on shipping methods and line items.
**What**
On cart creation:
- If region only has one country -> create cart with country code
On cart updates:
- If shipping address country code is provided in input ->
- If cart region doesn't include that country -> throw
- If cart includes the country -> update shipping address
- If region is provided in input and is different from the one currently on the cart ->
- If there is a shipping address on the cart -> clear the address
- If the region only has one country -> set country code of address
- If there is not a shipping address on the cart ->
- If the region only has one country -> set country code of address
Closes CC-545
