You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The big question is: do we need to bump a major version in GraphQLite when upgrading Webonyx?
GraphQLite acts as a wrapper around Webonyx and in theory, we could have minor versions of GraphQLite using different major versions of Webonyx. In practice, there is one breaking change (Strict coercion of scalar types) that might impact GraphQLite users.
I think any breaking change of stable features should follow https://graphqlite.thecodingmachine.io/docs/semver otherwise that promise means nothing. If the stable public API of GraphQLite has a breaking change when updating the webonyx library, then it should bump up a major version. If the webonyx library has a new update with breaking changes that don't affect the public API of GraphQLite, then I would make it a minor update.
Reacted by Marcin Wawrzyniak, Dmitry Grachikov and Andrii Maslov
@oojacoboo
The Webonyx's breaking change about strict coercion of scalar types has a great chance of affecting GraphQLite users, even though it would be a rare use case. Considering this, we should bump up the version to 5.0 with this update and I believe it worth doing so in the nearest future.
BTW there's a PR for this already and looks like it needs some work. The unit tests are failing there. I can take a look into it closer to the weekend. #327
I think we should hold off on dropping Doctrine Annotations in version 5. We should deprecate Doctrine annotations first, for at least one full version, before dropping them. PHP 8 hasn't even been released for a year yet. I think we need to give this more time before making such a change.
+1, forcing PHP8 would prevent a lot of people using it that can't upgrade yet because of other dependencies, or local dependencies on older PHP versions.
I do agree with the major 5.0 version update for updating Webonyx.
Activity
MattBred commentedon Aug 25, 2020
Hey @moufmouf any movement on this, or need any help with it?
moufmouf commentedon Aug 27, 2020
Hey @MattBred ,
Thanks for asking. Actually, I haven't started working on this yet. I'd like to get this PR merged before: #269 and it is quite massive.
I'd definitely accept some help with the migration to Webonyx 14 though!
MattBred commentedon Aug 27, 2020
Have you started any kind of plan/document on what would need changing yet?
moufmouf commentedon Aug 28, 2020
Not really.
I believe it should be relatively easy, according to the migration documentation of Webonyx: https://github.com/webonyx/graphql-php/blob/master/UPGRADE.md
The big question is: do we need to bump a major version in GraphQLite when upgrading Webonyx?
GraphQLite acts as a wrapper around Webonyx and in theory, we could have minor versions of GraphQLite using different major versions of Webonyx. In practice, there is one breaking change (Strict coercion of scalar types) that might impact GraphQLite users.
What do you think?
MattBred commentedon Sep 1, 2020
I think any breaking change of stable features should follow https://graphqlite.thecodingmachine.io/docs/semver otherwise that promise means nothing. If the stable public API of GraphQLite has a breaking change when updating the webonyx library, then it should bump up a major version. If the webonyx library has a new update with breaking changes that don't affect the public API of GraphQLite, then I would make it a minor update.
oojacoboo commentedon Mar 29, 2021
@moufmouf @devmaslov Should we do this and target a 5.0 release now that #269 has been merged in?
devmaslov commentedon Mar 30, 2021
@oojacoboo
The Webonyx's breaking change about strict coercion of scalar types has a great chance of affecting GraphQLite users, even though it would be a rare use case. Considering this, we should bump up the version to 5.0 with this update and I believe it worth doing so in the nearest future.
There's one more thing. Based on the docs, we should drop the support of Doctrine Annotations in 5th version, which means as well PHP version less then 8. It might be too early for this. What do you think?
https://graphqlite.thecodingmachine.io/docs/next/doctrine-annotations-attributes#docsNav
BTW there's a PR for this already and looks like it needs some work. The unit tests are failing there. I can take a look into it closer to the weekend.
#327
oojacoboo commentedon Mar 30, 2021
@devmaslov thanks for looking into this.
I think we should hold off on dropping Doctrine Annotations in version 5. We should deprecate Doctrine annotations first, for at least one full version, before dropping them. PHP 8 hasn't even been released for a year yet. I think we need to give this more time before making such a change.
devmaslov commentedon Mar 30, 2021
Totally agree
MattBred commentedon Apr 1, 2021
+1, forcing PHP8 would prevent a lot of people using it that can't upgrade yet because of other dependencies, or local dependencies on older PHP versions.
I do agree with the major 5.0 version update for updating Webonyx.