Discriminate generics inside lists

[patrick91]

Still WIP, should fix #3432, but right now it only fixes
the union part 😊

[codecov]

Codecov Report

Merging #3463 (8ad5789) into main (3edf95d) will increase coverage by 0.01%.
The diff coverage is 98.70%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3463      +/-   ##
==========================================
+ Coverage   96.50%   96.51%   +0.01%     
==========================================
  Files         510      511       +1     
  Lines       32761    32859      +98     
  Branches     5430     5451      +21     
==========================================
+ Hits        31615    31713      +98     
  Misses        914      914              
  Partials      232      232              

[botberry]

Apollo Federation Subgraph Compatibility Results

Federation 1 Support	Federation 2 Support
_service 🟢 @key (single) 🟢 @key (multi) 🟢 @key (composite) 🟢 repeatable @key 🟢 @requires 🟢 @provides 🟢 federated tracing 🔲	@link 🟢 @shareable 🟢 @tag 🟢 @override 🟢 @inaccessible 🟢 @composeDirective 🟢 @interfaceObject 🟢

Learn more:

• Apollo Federation Subgraph Specification
• Compatibility Tests

[patrick91]

Random thought we might want to add support for something like this:

return BlockRowType[str](total=3, items=[]),

because when we have 0 items when don't know which type we should use, so for now we fallback to the first

This is interesting, maybe it's useful: https://github.com/graphql-python/graphql-core/blob/3cf0d267b096ba92e7289d009800083166a5b7d9/src/graphql/execution/execute.py#L1947C1-L1957C16

[codspeed-hq]

CodSpeed Performance Report

Merging #3463 will not alter performance

_{Comparing feature/fix-unions-with-similar-fields (8ad5789) with main (3edf95d)}

Summary

✅ 12 untouched benchmarks

[patrick91]

/pre-release

[botberry]

Pre-release

👋

Releasing commit [cddafae] to PyPi as pre-release! 📦

[botberry]

Pre-release

👋

Pre-release 0.228.0.dev.1713643365 [cddafae] has been released on PyPi! 🚀
You can try it by doing:

poetry add strawberry-graphql==0.228.0.dev.1713643365

[botberry]

Thanks for adding the RELEASE.md file!

Here's a preview of the changelog:

---

This release improves our support for generic types, now using the same the same
generic multiple times with a list inside an interface or union is supported,
for example the following will work:

import strawberry


@strawberry.type
class BlockRow[T]:
    items: list[T]


@strawberry.type
class Query:
    @strawberry.field
    def blocks(self) -> list[BlockRow[str] | BlockRow[int]]:
        return [
            BlockRow(items=["a", "b", "c"]),
            BlockRow(items=[1, 2, 3, 4]),
        ]


schema = strawberry.Schema(query=Query)

Here's the tweet text:

🆕 Release (next) is out! Thanks to @patrick91 for the PR 👏

Get it here 👉 https://strawberry.rocks/release/(next)

[sourcery-ai]

Hey @patrick91 - I've reviewed your changes and they look great!

Here's what I looked at during the review

• 🟡 General issues: 3 issues found
• 🟢 Security: all looks good
• 🟡 Testing: 2 issues found
• 🟢 Complexity: all looks good
• 🟢 Docstrings: all looks good

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment to tell me if it was helpful.}

[patrick91]

@Speedy1991 let me know if you have time to try this!

[Speedy1991]

I'm not 100% sure that my code is correct, but this example fails: https://play.strawberry.rocks/?gist=d7d4da4f8b8fe245e645736ba5708b83

€dit: Nvm, I mixed up Interface with Union query. Looks like it works with unions:
https://play.strawberry.rocks/?gist=b4f3ffc37b21f9bd0a4d56525d4dad78

There is the obvious type collision on the client site (items of type str/int) so you are enforced to use an alias. This is the correct behaviour IMO

[Speedy1991]

Random thought we might want to add support for something like this:

return BlockRowType[str](total=3, items=[]),

Could this lead to problems on the client site? E.g. Apollo uses cache normalization (not sure about relay - didn't used it for years). I guess this should be ok as long as the return type matches the possible types for this field - just to make sure :)

[jibujacobamboss]

Adding the bug introduced as part of the PR