<!-- Thank you for submitting this PR! :) -->
## Description
We want to test `where` clauses over nested selects, so first let's get
a working test for selecting everything. Adds new test tables and
updates them in the test metadata etc. Functional no-op.
V3_GIT_ORIGIN_REV_ID: 8ebc6b27379b930eaf2220061c2ec58cd1650167
Add a test case for input type infinite recursion bug (https://github.com/hasura/v3-engine/pull/676). I confirmed the
test failure when the fix was removed.
V3_GIT_ORIGIN_REV_ID: b9d0a9b2a65328aa52f5ec057dcba6470c4e3956
## Description
This PR implements argument presets for `DataConnectorLink`, which can
be used to forward request headers as function/procedure arguments to
NDC. This PR implements the execution part.
**Note**: response header forwarding is not implemented yet.
V3_GIT_ORIGIN_REV_ID: ff69129aee3e3052367ca42acdec3922cbc2cb0c
<!-- Thank you for submitting this PR! :) -->
## Description
We'd like not to tie our boolean expressions to one particular data
connector, we were doing this in relationships and had no need to.
Functional no-op.
V3_GIT_ORIGIN_REV_ID: 0af16ad92117a45ac1f424b49d9f13184787f63f
## Description
This PR implements argument presets and response headers for
`DataConnectorLink`, which can be used to forward request headers as
function/procedure arguments to NDC, and also forward response headers
from NDC back to the client. This PR implements only the OpenDD changes
and metadata resolving part.
V3_GIT_ORIGIN_REV_ID: 2b7f20af7c2f192c0103553bccb000114dcba916
<!-- Thank you for submitting this PR! :) -->
## Description
This PR is a refactoring which is a functional no-op.
<!--
Questions to consider answering:
1. What user-facing changes are being made?
2. What are issues related to this PR? (Consider adding `(close
#<issue-no>)` to the PR title)
3. What is the conceptual design behind this PR?
4. How can this PR be tested/verified?
5. Does the PR have limitations?
6. Does the PR introduce breaking changes?
-->
## Changelog
- Add a changelog entry (in the "Changelog entry" section below) if the
changes
in this PR have any user-facing impact. See
[changelog
guide](https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide).
- If no changelog is required ignore/remove this section and add a
`no-changelog-required` label to the PR.
### Product
_(Select all products this will be available in)_
- [ ] community-edition
- [ ] cloud
<!-- product : end : DO NOT REMOVE -->
### Type
<!-- See changelog structure:
https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide#structure-of-our-changelog
-->
_(Select only one. In case of multiple, choose the most appropriate)_
- [ ] highlight
- [ ] enhancement
- [ ] bugfix
- [ ] behaviour-change
- [ ] performance-enhancement
- [ ] security-fix
<!-- type : end : DO NOT REMOVE -->
### Changelog entry
<!--
- Add a user understandable changelog entry
- Include all details needed to understand the change. Try including
links to docs or issues if relevant
- For Highlights start with a H4 heading (#### <entry title>)
- Get the changelog entry reviewed by your team
-->
_Replace with changelog entry_
<!-- changelog-entry : end : DO NOT REMOVE -->
<!-- changelog : end : DO NOT REMOVE -->
V3_GIT_ORIGIN_REV_ID: 625475ef23ef7fd2f72921e951fa1a957807516a
Previously we required an operator called `_eq` to be called `_eq` in
all data connectors that used it.
Now we can give them nice names, and map each one to a different
operator name per data connector.
V3_GIT_ORIGIN_REV_ID: 1d2f678e25d932e5ea45d34999cf5709ef0038b4
<!-- Thank you for submitting this PR! :) -->
## Description
Very occasionally these tests fail with the wrong duplicated graphql
name. Not sure why, I assume some unpredictable HashMap ordering
somewhere. In lieu of changing all HashMaps to BTreeMaps in engine
(which isn't immediately possible because of missing `Ord` instances for
`serde_json::Value`), let's make only one set of failing items in these
tests so the same one goes wrong each time.
Functional no-op.
V3_GIT_ORIGIN_REV_ID: 6dff998ae66aedbcca05dd05dc6fcb12e6d704c3
## Description
JIRA: https://hasurahq.atlassian.net/browse/V3ENGINE-180
Fix an edge case where the function
`build_annotations_from_input_object_type_permissions` produce infinite
recursions
when an input object type contains fields with self-referential types.
Note: This PR doesn't contain any tests. They'll be added in a separate
PR ([JIRA](https://hasurahq.atlassian.net/browse/V3ENGINE-188)). This
fix needs to be done quickly and merged as other teams (graphql
connector) are blocked by this.
V3_GIT_ORIGIN_REV_ID: 4cc09cb96871804d4daf83d7e49a6f4529c1943d
The original Aggregates RFC is missing some descriptions fields in its
specification of the Open DD metadata, which means that those particular
items currently don't get a description in GraphQL.
This PR adds the missing descriptions to the RFC document.
V3_GIT_ORIGIN_REV_ID: a93d5d92c427fe95f85ed3889660d2c22bc08f5a
<!-- Thank you for submitting this PR! :) -->
## Description
This makes the absolute happiest path `BooleanExpressionType` work.
Caveats:
- We ignore operator mappings, so we assume that the OpenDD operator
name (ie, `_eq`) is the same as the data connector's name for it.
- We have not tried to implement `comparisonRelationships` in any way
yet.
- Have not tried nested objects yet.
All functionality here is behind a feature flag so has no user-facing
impact.
V3_GIT_ORIGIN_REV_ID: 5edb706c6d8b03b9fa59433ce24f05f37a69729a
## Description
There is a bug in NDC validation of command where it returns early
without validating. This is because we perform the NDC validation right
after resolving the command source, but haven't attached the "resolved
command source" to the command yet. As a result the command's source is
`None`, and the validation does an early return.
This PR fixes it by taking the command source directly. If there is no
source to resolve, we don't have to perform any NDC validation of the
command.
Note: this is not the case with models. Model's source resolving
attaches the "resolved source", then performs the validation.
## Changelog
- Add a changelog entry (in the "Changelog entry" section below) if the
changes
in this PR have any user-facing impact. See
[changelog
guide](https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide).
- If no changelog is required ignore/remove this section and add a
`no-changelog-required` label to the PR.
### Product
_(Select all products this will be available in)_
- [x] community-edition
- [x] cloud
<!-- product : end : DO NOT REMOVE -->
### Type
<!-- See changelog structure:
https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide#structure-of-our-changelog
-->
_(Select only one. In case of multiple, choose the most appropriate)_
- [ ] highlight
- [ ] enhancement
- [x] bugfix
- [ ] behaviour-change
- [ ] performance-enhancement
- [ ] security-fix
<!-- type : end : DO NOT REMOVE -->
### Changelog entry
<!--
- Add a user understandable changelog entry
- Include all details needed to understand the change. Try including
links to docs or issues if relevant
- For Highlights start with a H4 heading (#### <entry title>)
- Get the changelog entry reviewed by your team
-->
Fix skipping of NDC validation of commands due to wrong assumption.
<!-- changelog-entry : end : DO NOT REMOVE -->
<!-- changelog : end : DO NOT REMOVE -->
V3_GIT_ORIGIN_REV_ID: 5b9f4c922ee83394f1b6fe840cd934b6e138e2e9
<!-- Thank you for submitting this PR! :) -->
## Description
We were using `String` for these, and making some assumptions about NDCs
and OpenDD operator names being the same. This uses newtypes and
documents where these assumptions are being made. Functional no-op.
V3_GIT_ORIGIN_REV_ID: 195e55db6fadf48c956684130e23f647eaa17fb1
Calling `.to_owned()` on a reference, `.to_vec()` on a vector reference,
etc. are just synonyms for `.clone()` which are less explicit about
cloning. Let's be explicit.
This also removes some unnecessary clones.
V3_GIT_ORIGIN_REV_ID: 1bc00c4106f0346303d73e4268c89030c0ce93fc
## Description
Tiny PR to add some comments and clarifications to the data connector
resolve types.
Also, rearranges the types and functions to make the flow clearer.
V3_GIT_ORIGIN_REV_ID: db5b18cbd7c79670c9dfc68c80950ccf5105f9ce
1. Use `map_or(…, …)` instead of `.map(…).unwrap_or(…)`.
2. Use `.is_some_and(…)` instead of `.map(…).unwrap_or_default(…)`.
3. Nest `|` patterns where possible.
4. Be more specific about match patterns.
I found I could also simplify `typecheck_qualified_type_reference`
considerably.
V3_GIT_ORIGIN_REV_ID: 6a3b1a4c525c0187c2fdb6df0c979ca0b7b3016c
This means we can use the cache far more often.
Pull requests will also use "dev" so as to avoid rebuilding
unnecessarily. Only builds on the `main` branch and tags will set a
version (using the short Git commit hash).
V3_GIT_ORIGIN_REV_ID: 37acaec3d369eef6df6487f9612f6d2da3e3ba5a
<!-- Thank you for submitting this PR! :) -->
## Description
We pass around `scalar_type_name` but never need it, removing it makes
boolean expressions easier to implement, so split into a fresh PR.
Functional no-op.
V3_GIT_ORIGIN_REV_ID: 0e4df8da4cb15d90a7f7be068b40c47936b70734
## Description
We can use Nix to build Docker images, which gives us a few advantages:
1. the images will be cached a little better
2. aarch64 builds become easy
3. Samir is happy because the Nixification continues
## Changelog
- Add a changelog entry (in the "Changelog entry" section below) if the
changes
in this PR have any user-facing impact. See
[changelog
guide](https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide).
- If no changelog is required ignore/remove this section and add a
`no-changelog-required` label to the PR.
### Product
_(Select all products this will be available in)_
- [x] community-edition
- [ ] cloud
<!-- product : end : DO NOT REMOVE -->
### Type
<!-- See changelog structure:
https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide#structure-of-our-changelog
-->
_(Select only one. In case of multiple, choose the most appropriate)_
- [ ] highlight
- [x] enhancement
- [ ] bugfix
- [ ] behaviour-change
- [ ] performance-enhancement
- [ ] security-fix
<!-- type : end : DO NOT REMOVE -->
### Changelog entry
<!--
- Add a user understandable changelog entry
- Include all details needed to understand the change. Try including
links to docs or issues if relevant
- For Highlights start with a H4 heading (#### <entry title>)
- Get the changelog entry reviewed by your team
-->
The v3 engine and dev-auth-webhook Docker images are now published for
both x86_64 (`amd64`) and aarch64 (`arm64`) architectures.
<!-- changelog-entry : end : DO NOT REMOVE -->
<!-- changelog : end : DO NOT REMOVE -->
Co-Authored-By: Philip Carlsen <philip@hasura.io>
Co-Authored-By: Gil Mizrahi <gil@hasura.io>
V3_GIT_ORIGIN_REV_ID: ae6fec45dee62a21f03b5258b57d841a16542c72
This PR refactors the schema crate to enable overriding the
interpretation of roles at runtime.
By providing a `NamespacedGetter` that ignores roles entirely (i.e.
`GDSNamespacedGetterAgnostic`) to schema generation functions it is
possible to derive a GraphQL schema that ignores permissions and
presets.
This PR leaves back a few opportunities to improve code clarity, which I
hope to address in follow-up PRs.
For instance, it seems superfluous to make sense for generic code
parametrised `<S: SchemaContext>` to carry around both a `S::Namespace`
and a `S::NamespaceGetter`, when all that code could ever do with that
`S::Namespace` value anyway is apply it to the `S::NamespaceGetter`.
Taking this line of reasoning to its conclusion suggests that we can
reduce the complexity of `trait SchemaContext` and its associated types
and kit.
---------
Co-authored-by: Samir Talwar <samir.talwar@hasura.io>
V3_GIT_ORIGIN_REV_ID: c84555fd88279582670919faa5a62bbd00b714bd
1. Always name the struct when calling `default()`.
2. Sort construction according to the definitions.
3. Approve allowing `struct_field_names` because it doesn't seem to be
helpful.
4. Enable `manual_string_new`; nothing seems to be triggering it now.
V3_GIT_ORIGIN_REV_ID: 868742114b0bf27bc3ea03cdf1e63a0f710ebe33
<!-- Thank you for submitting this PR! :) -->
## Description
Broke this out of new bool exps work, breaks this big function into
smaller functions with clearer dependencies. Functional no-op.
V3_GIT_ORIGIN_REV_ID: 4f85132f9c381977b1cd406bb71e278c45866691
<!-- Thank you for submitting this PR! :) -->
## Description
This somewhat resolves the `BooleanExpressionType` metadata type.
There is a lot that is missing, but everything is behind a feature so I
want to merge what is here before moving onto getting it working with
the later parts of the pipeline. We'll start with a lot of duplication
and then work to remove as much as we can.
Functional no-op for users, the feature flag is not exposed to the
actual binary.
How to review this PR:
- Check that any changes to existing code are cosmetic and naming-only
to ensure this is a no-op
- Look through the new resolve step and compare it to the
[RFC](https://github.com/hasura/v3-engine/blob/main/rfcs/open-dd-expression-type-changes.md)
and see if it is doing roughly what you might expect.
- Don't get too bogged down in missing things / style stuff, there are
definitely missing things and bad style.
V3_GIT_ORIGIN_REV_ID: 24e3b3e72e62d0094db3b461cb8c6d359982755d
Rather than allowing the `cast_precision_loss` and `inline_always`
warnings everywhere, we just suppress them in the few places they're
already used.
V3_GIT_ORIGIN_REV_ID: c28350a21efa029c8c6aae2a602eec2d75f42216
<!-- Thank you for submitting this PR! :) -->
## Description
The field mappings for an object type are the mappings between the
OpenDD field name (`FieldName`) and whatever the NDC column name
(`DataConnectorColumnName`) are.
Type mappings is a map from OpenDD type names
(`Qualified<CustomTypeName>`) to the field mappings for that type.
Previously, Boolean Expressions were attached to a single data connector
and object type. Now we are generalising them so that they can be used
across any of the multiple data connectors that their attached
ObjectType might have type mappings for.
To do this, we stop including them in the resolved
`ObjectBooleanExpressionType`, and instead fetch the relevant type
mappings from whichever thing is using the boolean expression type,
usually the `ModelSource` for a `Model`.
Functional no-op.
V3_GIT_ORIGIN_REV_ID: 4a251473b4aecedf4ede0e79aa7b577e25207cb8
<!-- Thank you for submitting this PR! :) -->
## Description
We'd like to be able to test new WIP experimental features. This adds a
`UNSTABLE_FEATURES` env var / command line arg that can be passed a
comma separated list of names.
Currently we accept `UNSTABLE_FEATURES=enable-boolean-expression-types`
but in future users could pass
`UNSTABLE_FEATURES=some-fancy-feature,other-feature,great`.
<!--
Questions to consider answering:
1. What user-facing changes are being made?
2. What are issues related to this PR? (Consider adding `(close
#<issue-no>)` to the PR title)
3. What is the conceptual design behind this PR?
4. How can this PR be tested/verified?
5. Does the PR have limitations?
6. Does the PR introduce breaking changes?
-->
## Changelog
- Add a changelog entry (in the "Changelog entry" section below) if the
changes
in this PR have any user-facing impact. See
[changelog
guide](https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide).
- If no changelog is required ignore/remove this section and add a
`no-changelog-required` label to the PR.
### Product
_(Select all products this will be available in)_
- [x] community-edition
- [x] cloud
<!-- product : end : DO NOT REMOVE -->
### Type
<!-- See changelog structure:
https://github.com/hasura/graphql-engine-mono/wiki/Changelog-Guide#structure-of-our-changelog
-->
_(Select only one. In case of multiple, choose the most appropriate)_
- [ ] highlight
- [x] enhancement
- [ ] bugfix
- [ ] behaviour-change
- [ ] performance-enhancement
- [ ] security-fix
<!-- type : end : DO NOT REMOVE -->
### Changelog entry
<!--
- Add a user understandable changelog entry
- Include all details needed to understand the change. Try including
links to docs or issues if relevant
- For Highlights start with a H4 heading (#### <entry title>)
- Get the changelog entry reviewed by your team
-->
Add `UNSTABLE_FEATURES` environment variable
<!-- changelog-entry : end : DO NOT REMOVE -->
<!-- changelog : end : DO NOT REMOVE -->
V3_GIT_ORIGIN_REV_ID: 3562c1341d4ea3a512626110dbd2b055425c1d60
Add the ability to render field input-arguments and default values in
sdl.
Also add a small CLI tool to help inspect the GraphQL schemas that
result from a piece of engine metadata.
V3_GIT_ORIGIN_REV_ID: ddb7315f12d24eafbf57782f37f840ca44d94789
## Description
I want to be able to use a `where` clause for Artists and Albums when
running v3-engine with the tests ndc-postgres schema.
V3_GIT_ORIGIN_REV_ID: 50824decb794f41f745268e7ca297a5a1c04c808
Sometimes you just want to run a specific test.
The arguments are passed straight to the test runner: `cargo nextest` if
you have it installed, and `cargo test` if you do not. These accept
different arguments so the user will need to know which one they are
using.
V3_GIT_ORIGIN_REV_ID: 9cd6f9d770899e0bb3aa4537008eedba052818d7
Just because it's fewer lines of code.
1. Invert `if`/`else` blocks with negative conditions.
2. Unwrap redundant `else` blocks.
3. Simplify a few branches to `let … else`.
4. Replace a `match` with `if let`.
V3_GIT_ORIGIN_REV_ID: 4f10730b688d21c1fc86a45ee5fb4adf008b3d94
Sometimes they will be borrowed and therefore must be cloned, but
sometimes they'll be owned, and we can use that.
V3_GIT_ORIGIN_REV_ID: 99f40f303c42acc0f44e4ce3081a1c9dce1802b1
