Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Refactor argument-handling to use a struct #103

Merged
merged 5 commits into from
Sep 23, 2021

Conversation

benjaminjkraft
Copy link
Collaborator

@benjaminjkraft benjaminjkraft commented Sep 17, 2021

Summary:

In this commit I refactor the argument-generation logic to move most of
the code out of the template and into the type-generator. This logic
predates #51, and I didn't think to update it there, but I think it
benefits from similar treatment, for similar reasons.

Specifically, the main change is to treat variables as another struct
type we can generate, rather than handling them inline as a
map[string]interface{}. Users still pass them the same way, but
instead of putting them into a map[string]interface{} and JSONifying
that, we generate a struct and put them there.

This turns out to simplify things quite a lot, because we already have a
lot of code to generate types. Notably, the omitempty code goes from a
dozen lines to basically two, and fixes a bug (#43) in the process,
because now that we have a struct, json.Marshal will do our work for
us! (And, once we have syntax for it (#14), we'll be able to handle
field-level omitempty basically for free.) More importantly, it will
simplify custom marshalers (#38, forthcoming) significantly, since we do
all that logic at the containing-struct level, but will need to apply it
to arguments.

It does require two breaking changes:

  1. For folks implementing the graphql.Client API (rather than just
    calling NewClient): we now pass them variables as an interface{}
    rather than a map[string]interface{}. For most callers, including
    Khan/webapp, this is basically a one-line change to the signature of
    their MakeRequest, and it should be a lot more future-proof.
  2. genqlient's handling of the omitempty option has changed to match
    that of encoding/json, in particular it now never considers structs
    "empty". The difference was never intentional (I just didn't realize
    that behavior of encoding/json); arguably our behavior was more
    useful but I think that's outweighed by the value of consistency with
    encoding/json as well as the simpler and more correct
    implementation (fixing omitempty generates broken code for structs #43 is actually quite nontrivial otherwise).
    Once we have custom unmarshaler support (Allow mapping directly to a type but not using its standard JSON serialization #38), users will be able to
    map a zero value to JSON null if they wish, which is mostly if not
    entirely equivalent for GraphQL's purposes.

Issue: #38
Issue: #43

Test plan:

make check

@benjaminjkraft
Copy link
Collaborator Author

Hmm, I realized this actually makes some more breaking changes, because my implementation of omitempty had some inconsistencies with JSON around structs. Let me think about this some more.

Copy link
Member

@StevenACoffman StevenACoffman left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So that was my concern... that it is "breaking"-ish, which I guess if fine before our blog post and initial release version, but I'd like to do that sooner rather than later, and be more sure we don't need to again if we do this.

@benjaminjkraft
Copy link
Collaborator Author

Yeah, I think if we had more users I'd be more inclined to live with the existing behavior, but now's a good time to fix the inconsistency before anyone cares. (FWIW, this appears to be something that yaml changed from v1 to v2!) In practice it's not hard to search for potential problems (just find struct-typed fields with # genqlient(omitempty: true)), and webapp has none.

benjaminjkraft added a commit that referenced this pull request Sep 17, 2021
This is useful if you want to bind to a type you don't control (or use
for other things) but need different serialization than its default.
This is a feature gqlgen has and we've found it very useful.  For
example, in webapp we want to bind `DateTime` to `time.Time`, but its
default serialization is not compatible with Python, so currently we
have to bind to a wrapper type and cast all over the place, which is
exactly the sort of boilerplate genqlient is supposed to avoid.

For unmarshaling, the implementation basically just follows the existing
support for abstract types; instead of calling our own generated
helper, we now call your specified function.  This required some
refactoring to abstract the handling of custom unmarshalers generally
from abstract types specifically, and to wire in not only the
unmarshaler-name but also the `generator` (in order to compute the right
import alias).

For marshaling, I had to implement all that stuff over again; it's
mostly parallel to unmarshaling (and I made a few minor changes to
unmarshaling to make the two more parallel).  Luckily, after #103 I at
least only had to do it once, rather than implementing the same
functionality for arguments and for input-type fields.  It was still
quite a bit of code; I didn't try to be quite as completionist about the
tests as with unmarshal but still had to add a few.

Issue: #38

Test plan:
make check

Reviewers: marksandstrom, steve, adam, jvoll, miguel, mahtab
Copy link
Contributor

@dnerdy dnerdy left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nice. I don't feel like we need to worry about breaking anyone at this point, plus the change is well documented.

{{.GoName}} {{.GoType}},
{{- if .Input -}}
{{- range .Input.Fields -}}
{{/* the GraphQL name is here the user-specified variable-name */ -}}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

typo? "is here" => "here is"

// Typically, GraphQL APIs will accept a JSON payload of the form
// `query myQuery { myField }`. variables contains a JSON-marshalable
// value containing the variables to be sent along with the query,
// or may be nil if there are none. Typically, GraphQL APIs wills
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
// or may be nil if there are none. Typically, GraphQL APIs wills
// or may be nil if there are none. Typically, GraphQL APIs will

{{if not .Options.GetOmitempty -}}
"{{.GraphQLName}}": {{.GoName}},
{{- if .Input -}}
{{/* __input is not a legal variable name in GraphQL */ -}}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It might be helpful to be more explicit in this comment. Something like "__input won't clash with arguments to this function since all argument names are valid GraphQL variable names and __input isn't a legal variable name in GraphQL".

Comment on lines 164 to 165
// note prefix is ignored here (see generator.typeName), as is
// selectionSet (for input types we use the whole thing).
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree with the comment in the code below that it's a bit confusing to pass around arguments that aren't used. It may be helpful to explicitly say that selection sets don't apply to input objects, and input types may only be scalars, enums or (recursively) other input objects. A mention of ast.InputObject in convertDefinition might also be helpful.

Also, I think generator.typeName is a dangling reference? (Here and below in convertDefinition.)

In this commit I refactor the argument-generation logic to move most of
the code out of the template and into the type-generator.  This logic
predates #51, and I didn't think to update it there, but I think it
benefits from similar treatment, for similar reasons.

Specifically, the main change is to treat variables as another struct
type we can generate, rather than handling them inline as a
`map[string]interface{}`.  Users still pass them the same way, but
instead of putting them into a `map[string]interface{}` and JSONifying
that, we generate a struct and put them there.

This turns out to simplify things quite a lot, because we already have a
lot of code to generate types.  Notably, the omitempty code goes from a
dozen lines to basically two, and fixes a bug (#43) in the process,
because now that we have a struct, `json.Marshal` will do our work for
us! (And, once we have syntax for it (#14), we'll be able to handle
field-level omitempty basically for free.)  More importantly, it will
simplify custom marshalers (#38, forthcoming) significantly, since we do
all that logic at the containing-struct level, but will need to apply it
to arguments.

It does require one breaking change, for folks implementing the
`graphql.Client` API (rather than just calling `NewClient`): we now pass
them variables as an `interface{}` rather than a
`map[string]interface{}`.  For most callers, including Khan/webapp, this
is basically a one-line change to the signature of their `MakeRequest`.

Issue: #38
Issue: #43

Test plan:
make check

Reviewers: marksandstrom, steve, jvoll, mahtab, adam, miguel
@benjaminjkraft benjaminjkraft force-pushed the benkraft.argument-refactor branch from d344208 to 7794c1b Compare September 23, 2021 00:13
benjaminjkraft added a commit that referenced this pull request Sep 23, 2021
This is useful if you want to bind to a type you don't control (or use
for other things) but need different serialization than its default.
This is a feature gqlgen has and we've found it very useful.  For
example, in webapp we want to bind `DateTime` to `time.Time`, but its
default serialization is not compatible with Python, so currently we
have to bind to a wrapper type and cast all over the place, which is
exactly the sort of boilerplate genqlient is supposed to avoid.

For unmarshaling, the implementation basically just follows the existing
support for abstract types; instead of calling our own generated
helper, we now call your specified function.  This required some
refactoring to abstract the handling of custom unmarshalers generally
from abstract types specifically, and to wire in not only the
unmarshaler-name but also the `generator` (in order to compute the right
import alias).

For marshaling, I had to implement all that stuff over again; it's
mostly parallel to unmarshaling (and I made a few minor changes to
unmarshaling to make the two more parallel).  Luckily, after #103 I at
least only had to do it once, rather than implementing the same
functionality for arguments and for input-type fields.  It was still
quite a bit of code; I didn't try to be quite as completionist about the
tests as with unmarshal but still had to add a few.

Issue: #38

Test plan:
make check

Reviewers: marksandstrom, steve, adam, jvoll, miguel, mahtab
@benjaminjkraft benjaminjkraft merged commit 5995653 into main Sep 23, 2021
@benjaminjkraft benjaminjkraft deleted the benkraft.argument-refactor branch September 23, 2021 00:16
benjaminjkraft added a commit that referenced this pull request Sep 23, 2021
This is useful if you want to bind to a type you don't control (or use
for other things) but need different serialization than its default.
This is a feature gqlgen has and we've found it very useful.  For
example, in webapp we want to bind `DateTime` to `time.Time`, but its
default serialization is not compatible with Python, so currently we
have to bind to a wrapper type and cast all over the place, which is
exactly the sort of boilerplate genqlient is supposed to avoid.

For unmarshaling, the implementation basically just follows the existing
support for abstract types; instead of calling our own generated
helper, we now call your specified function.  This required some
refactoring to abstract the handling of custom unmarshalers generally
from abstract types specifically, and to wire in not only the
unmarshaler-name but also the `generator` (in order to compute the right
import alias).

For marshaling, I had to implement all that stuff over again; it's
mostly parallel to unmarshaling (and I made a few minor changes to
unmarshaling to make the two more parallel).  Luckily, after #103 I at
least only had to do it once, rather than implementing the same
functionality for arguments and for input-type fields.  It was still
quite a bit of code; I didn't try to be quite as completionist about the
tests as with unmarshal but still had to add a few.

Issue: #38

Test plan:
make check

Reviewers: marksandstrom, steve, adam, jvoll, miguel, mahtab
benjaminjkraft added a commit that referenced this pull request Sep 24, 2021
## Summary:
This is useful if you want to bind to a type you don't control (or use
for other things) but need different serialization than its default.
This is a feature gqlgen has and we've found it very useful.  For
example, in webapp we want to bind `DateTime` to `time.Time`, but its
default serialization is not compatible with Python, so currently we
have to bind to a wrapper type and cast all over the place, which is
exactly the sort of boilerplate genqlient is supposed to avoid.

For unmarshaling, the implementation basically just follows the existing
support for abstract types; instead of calling our own generated
helper, we now call your specified function.  This required some
refactoring to abstract the handling of custom unmarshalers generally
from abstract types specifically, and to wire in not only the
unmarshaler-name but also the `generator` (in order to compute the right
import alias).

For marshaling, I had to implement all that stuff over again; it's
mostly parallel to unmarshaling (and I made a few minor changes to
unmarshaling to make the two more parallel).  Luckily, after #103 I at
least only had to do it once, rather than implementing the same
functionality for arguments and for input-type fields.  It was still
quite a bit of code; I didn't try to be quite as completionist about the
tests as with unmarshal but still had to add a few.

Issue: #38

## Test plan:
make check


Author: benjaminjkraft

Reviewers: StevenACoffman, dnerdy, benjaminjkraft, aberkan, jvoll, mahtabsabet, MiguelCastillo

Required Reviewers: 

Approved By: StevenACoffman, dnerdy

Checks: ✅ Test (1.17), ✅ Test (1.16), ✅ Test (1.15), ✅ Test (1.14), ✅ Lint, ✅ Test (1.17), ✅ Test (1.16), ✅ Test (1.15), ✅ Test (1.14), ✅ Lint

Pull Request URL: #104
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

omitempty generates broken code for structs
3 participants