The apollo-cache-hermes from convoyinc

Currently, we interpolate variables at each field that requests them - instead, we could do this once when building the dynamic edge map

Switch from `export type` blah to `export interface` blah in more places

Utilizing an interface gives the type a name, which makes it easier to debug (and also allows for augmentation)

Clean up terminology

Probably:

edge -> field
node.node -> node.value?
???

Cache query completeness per snapshot, and propagate it

Once we've found a query to be complete (either by reading it, or by writing it), we should be guaranteed that it will continue to be complete, unless nodes are explicitly removed (not garbage collected).

There's no need to walk results on every read.

An in-range update of typescript-eslint-parser is breaking the build 🚨

Version 8.0.1 of typescript-eslint-parser was just published.

Branch	Build failing 🚨
Dependency	typescript-eslint-parser
Current Version	8.0.0
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

typescript-eslint-parser is a devDependency of this project. It might not break your production code or affect downstream projects, but probably breaks your build or test tools, which may prevent deploying or publishing.

Status Details

❌ ci/circleci: build Your tests failed on CircleCI Details

Release Notes

v8.0.1

0401ffc Fix: Calculate typeArguments loc data correctly if empty (fixes #395) (#396) (James Henry)
a214f71 Chore: Add a way to test TSX specific issues (fixes #376) (#398) (James Henry)
9c71a62 Fix: add missing TSSymbolKeyword type (#385) (Ika)
e10aab8 Chore: Refactor alignment tests, now on by default (#387) (James Henry)
9e17d0b Chore: Minor cleanup, fix jQuery foundation copyright (#383) (James Henry)

Commits

The new version differs by 7 commits.

5e9e7fa 8.0.1
5a96293 Build: changelog update for 8.0.1
0401ffc Fix: Calculate typeArguments loc data correctly if empty (fixes #395) (#396)
a214f71 Chore: Add a way to test TSX specific issues (fixes #376) (#398)
9c71a62 Fix: add missing TSSymbolKeyword type (#385)
e10aab8 Chore: Refactor alignment tests, now on by default (#387)
9e17d0b Chore: Minor cleanup, fix jQuery foundation copyright (#383)

See the full diff

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

An in-range update of @types/node is breaking the build 🚨

Version 8.0.47 of @types/node was just published.

Branch	Build failing 🚨
Dependency	@types/node
Current Version	8.0.46
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

@types/node is a devDependency of this project. It might not break your production code or affect downstream projects, but probably breaks your build or test tools, which may prevent deploying or publishing.

Status Details

❌ ci/circleci: build Your tests failed on CircleCI Details

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

Rename query to operation

Let's get closer to actual GraphQL nomenclature, and is pretty damn confusing. The cache currently uses query interchangeably with operation in a bunch of places.

Support for field aliases

Gotta support 'em!

GC pass

There are several places where we are leaking, or are storing more than necessary, and will eventually need to perform some sort of GC:

Cached queries
Parameterized edges/entities
???

Are there any docs?

Hey guys, I'm really curious to see how your store performs compared with the InMemoryStore, especially when making several mutations per second and stacking a few hundred or a couple thousand optimistic updates before resolving them. However, there's very little documentation for this package and not even really a readme that tells me how to use it, and what to expect. For example, can I just plug it in place of the InMemoryStore with the same arguments (dataIdFromObject, addTypename being the most important ones) and expect everything to work, including readFragment and writeFragment? Or are there certain things that are different, things that haven't been implemented yet, etc.?

Btw, I looked into the source code to figure things out, but I still couldn't get it to work in Typescript because of missing exports and compiler issues (skipLibCheck didn't help). Are you actually using the package you published or are you using a local version?

Support for directives

Do we need to support them? Or can that be purely server-side
If we do, how do we do it?

Support for extract() and restore()

Needed for persistence!

Full support for union/interface types

Update: Much of this has been implemented. See this comment for remaining work.

~~We don't support ... on Type { queries currently.~~ But it shouldn't be too difficult to support, I think.

Rough outline of the work:

~~When parsing queries, we need to walk InlineFragment AST nodes. Most likely, they are added as another field node (but explicitly flagged as being a union field)~~
~~When writing fields, any union field present in the payload should be written normally~~ (except if not present, nulls shouldn't be inserted)
~~When reading queries there should be no work required for unions with only static fields~~

Potential challenges:

Parameterized fields contained by a union query may need special handling during (and __typename or schema introspection)

Test-only assertions

Consider peppering the code with assertions that are removed from the packaged output, but are active during tests

An in-range update of typescript is breaking the build 🚨

Version 2.6.1 of typescript was just published.

Branch	Build failing 🚨
Dependency	typescript
Current Version	2.5.3
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

typescript is a devDependency of this project. It might not break your production code or affect downstream projects, but probably breaks your build or test tools, which may prevent deploying or publishing.

Status Details

❌ ci/circleci: build Your tests failed on CircleCI Details

Release Notes

TypeScript 2.6

For release notes, check out the release announcement

For new features, check out the What's new in TypeScript 2.6.

For breaking changes, check out the Breaking changes in TypeScript 2.6 page.

For API breaking changes, check out the API Breaking changes in TypeScript 2.6 page.

For the complete list of fixed issues, check out the issues fixed in TypeScript 2.6 RC and after TypeScript 2.6 RC.

Download:

Special thanks to all contributors to this release:

Adrian Leonhard
Alex Chugaev
Alex Eagle
Anders Hejlsberg
Andrew Casey
Andy Hanson
Armando Aguirre
Arthur Ozga
Basarat Ali Syed
Benjamin Lichtman
Charles Pierce
Daniel Rosenwasser
@falsandtru
Filipe Silva
Francois Wouts
Henry Mercer
Herrington Darkholme
Ivan Enderlin
Joe Calzaretta
Klaus Meinhardt
Magnus Kulke
Matt Bierner
Matt Mitchell
Mohamed Hegazy
Nathan Shively-Sanders
Paul van Brenk
Ron Buckton
Ryan Cavanaugh
Sheetal Nandi
Stas Vilchik
Taras Mankovski
Thomas den Hollander
Vakhurin Sergey
Wesley Wigham
Zeeshan Ahmed
Zev Spitz

Commits

The new version differs by 1177 commits ahead by 1177, behind by 85.

477bdc7 Update LKG
6bc04a9 Use lowercased names for generated locale directories (#19581)
0f32f55 Make getCompletionEntryDetails and getCompletionEntrySymbol parameters non-optional (#19507) (#19592)
1359a93 Update LKG
a7e172b Support multiple completions with the same name but different source module (#19455) (#19496)
c35e90e Merge pull request #19484 from Microsoft/reloadScriptsWithoutProject
96c66a6 Add test case to verify reload works without open project
b16e6d4 Reload contents of file from disk irrespective of project presence and file already containing its own text
6489a75 Update LKG
1a1c672 Merge pull request #19445 from Microsoft/foundWhileSearchingNodeModules
4e56347 Set the files found during node_modules search correctly when reusing existing program structure
aaaa280 Add the test case that fails and reports error for files not in root directory in watch mode
3134111 Merge pull request #19434 from Microsoft/port-master-10-23
a362bbf Fix lint error
02cc07c For @types installing quickfix, only activate for implicit-any module (#19394) (#19429)

There are 250 commits in total.

See the full diff

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

GC orphaned array references

Fix the bugs that were introduced in the code removed by #127

Runtime type check for cache configuration

With TypeScript, type checker helps catching lots of mistake. However, when using using Hermes in JS codebase, I think it will be valuable to provide some runtime check for correct type for cache config.

One possible option to do so (https://github.com/rbuckton/reflect-metadata)

Update Apollo-specific interface

Rename apollo/Transaction -> DataProxy for consistency
Update method of ApolloTrasaction class to have similar API as DataProxy

Switch to asset caching in Circle

The workspaces are great, except they aren't preserved for reruns :(

Integration with Apollo Dev tools

We probably need to work with that project to figure out the right interface to expose to the dev tools.

Cache redirects

Need to support cache redirects.

Motivation is to support queries like:

query getAUser($id: ID!) {
  user(id: $id) { ... }
}

and return the entity with that id, no matter where it was retrieved from

An in-range update of apollo-utilities is breaking the build 🚨

Version 1.0.2 of apollo-utilities was just published.

Branch	Build failing 🚨
Dependency	apollo-utilities
Current Version	1.0.1
Type	dependency

This version is covered by your current version range and after updating it in your project the build failed.

apollo-utilities is a direct dependency of this project, and it is very likely causing it to break. If other packages depend on yours, this update is probably also breaking those in turn.

Status Details

✅ ci/circleci: build Your tests passed on CircleCI! Details
❌ ci/circleci: test:unit CircleCI is running your tests Details
❌ ci/circleci: test:style Your tests failed on CircleCI Details

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

Walk selection set when writing, rather than payload

Currently, we walk the payload when merges values into the cache. This leads to ambiguity and hacky situations where we can't know whether a value should be placed directly in the cache, or walked - particularly for fields that contain complex objects

An in-range update of @types/lodash is breaking the build 🚨

Version 4.14.79 of @types/lodash was just published.

Branch	Build failing 🚨
Dependency	@types/lodash
Current Version	4.14.78
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

@types/lodash is a devDependency of this project. It might not break your production code or affect downstream projects, but probably breaks your build or test tools, which may prevent deploying or publishing.

Status Details

❌ ci/circleci: build Your tests failed on CircleCI Details

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

Coverage for Apollo interfaces

Once the API solidifies, let's test the crap out of it.

[Infra] Pull test for both read/write so that each file contains a describe.

From doing rewrite work, I realize that it is quite hard to understand the test because in a file there can be multiple describe which make it a bit hard to parse and understand. I propose that we should have one describe per file. The downside is that running take will take a bit longer as there are more initialization process to do

operations/write
operations/read

Add heterogeneous type array once we have union types

An in-range update of eslint-plugin-jest is breaking the build 🚨

Version 21.3.0 of eslint-plugin-jest was just published.

Branch	Build failing 🚨
Dependency	eslint-plugin-jest
Current Version	21.2.0
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

eslint-plugin-jest is a devDependency of this project. It might not break your production code or affect downstream projects, but probably breaks your build or test tools, which may prevent deploying or publishing.

Status Details

✅ ci/circleci: build Your tests passed on CircleCI! Details
❌ ci/circleci: test:unit CircleCI is running your tests Details
❌ ci/circleci: test:style Your tests failed on CircleCI Details

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

Clean up test organization

They're getting pretty hard to categorize/find

An in-range update of @types/chai is breaking the build 🚨

Version 4.0.6 of @types/chai was just published.

Branch	Build failing 🚨
Dependency	@types/chai
Current Version	4.0.5
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

@types/chai is a devDependency of this project. It might not break your production code or affect downstream projects, but probably breaks your build or test tools, which may prevent deploying or publishing.

Status Details

✅ ci/circleci: build Your tests passed on CircleCI! Details
❌ ci/circleci: test:unit CircleCI is running your tests Details
❌ ci/circleci: test:style Your tests failed on CircleCI Details

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

An in-range update of eslint is breaking the build 🚨

Version 4.7.0 of eslint just got published.

Branch	Build failing 🚨
Dependency	eslint
Current Version	4.6.1
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

As eslint is “only” a devDependency of this project it might not break production or downstream projects, but “only” your build or test tools – preventing new deploys or publishes.

I recommend you give this issue a high priority. I’m sure you can resolve this 💪

Status Details

✅ ci/circleci: build Your tests passed on CircleCI! Details
❌ ci/circleci: test:unit CircleCI is running your tests Details
❌ ci/circleci: test:style Your tests failed on CircleCI Details

Release Notes

v4.7.0

787b78b Upgrade: Espree v3.5.1 (fixes #9153) (#9314) (Brandon Mills)
1488b51 Update: run rules after node.parent is already set (fixes #9122) (#9283) (Teddy Katz)
4431d68 Docs: fix wrong config in max-len example. (#9309) (薛定谔的猫)
7d24dde Docs: Fix code snippet to refer to the correct option (#9313) (Ruben Tytgat)
12388d4 �Chore: rewrite parseListConfig for a small perf gain. (#9300) (薛定谔的猫)
ce1f084 Update: fix MemberExpression handling in no-extra-parens (fixes #9156) (jackyho112)
0c720a3 Update: allow autofixing when using processors (fixes #7510) (#9090) (Teddy Katz)
838df76 Chore: upgrade deps. (#9289) (薛定谔的猫)
f12def6 Update: indent flatTernary option to handle return (fixes #9285) (#9296) (Teddy Katz)
e220687 Fix: remove autofix for var undef inits (fixes #9231) (#9288) (Victor Hom)
002e199 Docs: fix no-restricted-globals wrong config. (#9305) (薛定谔的猫)
fcfe91a Docs: fix wrong config in id-length example. (#9303) (薛定谔的猫)
2731f94 Update: make newline-per-chained-call fixable (#9149) (João Granado)
61f1093 Chore: avoid monkeypatching Linter instances in RuleTester (#9276) (Teddy Katz)
28929cb Chore: remove Linter#reset (refs #9161) (#9268) (Teddy Katz)
abc8634 Build: re-run browserify when generating site (#9275) (Teddy Katz)
7685fed Fix: IIFE and arrow functions in no-invalid-this (fixes #9126) (#9258) (Toru Nagashima)
2b1eba2 Chore: enable eslint-plugin/no-deprecated-context-methods (#9279) (Teddy Katz)
981f933 Fix: reuse the AST of source code object in verify (#9256) (Toru Nagashima)
cd698ba Docs: move RuleTester documentation to Node.js API page (#9273) (Teddy Katz)
4ae7ad3 Docs: fix inaccuracy in npm run perf description (#9274) (Teddy Katz)
cad45bd Docs: improve documentation for rule contexts (#9272) (Teddy Katz)
3b0c6fd Chore: remove extraneous linter properties (refs #9161) (#9267) (Teddy Katz)
c3231b3 Docs: Fix typo in array-bracket-newline.md (#9269) (宋文强)
51132d6 Fix: Formatters keep trailing '.' if preceded by a space (fixes #9154) (#9247) (i-ron-y)
88d5d4d Chore: remove undocumented Linter#markVariableAsUsed method (refs #9161) (#9266) (Teddy Katz)
09414cf Chore: remove internal Linter#getDeclaredVariables method (refs #9161) (#9264) (Teddy Katz)
f31f59d Chore: prefer smaller scope for variables in codebase (#9265) (Teddy Katz)
3693e4e Chore: remove undocumented Linter#getScope method (#9253) (Teddy Katz)
5d7eb81 Chore: refactor config hash caching in CLIEngine (#9260) (Teddy Katz)
1a76c4d Chore: remove SourceCode passthroughs from Linter.prototype (refs #9161) (#9263) (Teddy Katz)
40ae27b Chore: avoid relying on Linter#getScope/markVariableAsUsed in tests (#9252) (Teddy Katz)
b383d81 Chore: make executeOnFile a pure function in CLIEngine (#9262) (Teddy Katz)
5e0e579 Chore: avoid internal SourceCode methods in Linter tests (refs #9161) (#9223) (Teddy Katz)
adab827 Chore: remove unused eslint-disable comment (#9251) (Teddy Katz)
31e4ec8 Chore: use consistent names for apply-disable-directives in tests (#9246) (Teddy Katz)
7ba46e6 Fix: shebang error in eslint-disable-new-line; add tests (fixes #9238) (#9240) (i-ron-y)
8f6546c Chore: remove undocumented defaults() method (refs #9161) (#9237) (Teddy Katz)
82d8b73 Docs: Fix error in example code for sort-imports (fixes #8734) (#9245) (i-ron-y)
a32ec36 Update: refactor eslint-disable comment processing (#9216) (Teddy Katz)
583f0b8 Chore: avoid using globals in CLIEngine tests (#9242) (Teddy Katz)
c8bf687 Chore: upgrade [email protected] (#9234) (薛定谔的猫)
3c41a05 Chore: always normalize rules to new API in rules.js (#9236) (Teddy Katz)
c5f4227 Chore: move logic for handling missing rules to rules.js (#9235) (Teddy Katz)
bf1e344 Chore: create report translators lazily (#9221) (Teddy Katz)
2eedc1f Chore: remove currentFilename prop from Linter instances (refs #9161) (#9219) (Teddy Katz)
5566e94 Docs: Replace misleading CLA links (#9133) (#9232) (i-ron-y)
c991630 Chore: remove ConfigOps.normalize in favor of ConfigOps.getRuleSeverity (#9224) (Teddy Katz)
171962a Chore: remove internal Linter#getAncestors helper (refs #9161) (#9222) (Teddy Katz)
a567499 Chore: avoid storing list of problems on Linter instance (refs #9161) (#9214) (Teddy Katz)
ed6d088 Chore: avoid relying on undocumented Linter#getFilename API in tests (#9218) (Teddy Katz)

Commits

The new version differs by 55 commits.

439e8e6 4.7.0
2ec62f9 Build: changelog update for 4.7.0
787b78b Upgrade: Espree v3.5.1 (fixes #9153) (#9314)
1488b51 Update: run rules after node.parent is already set (fixes #9122) (#9283)
4431d68 Docs: fix wrong config in max-len example. (#9309)
9d1df92 Chore: Revert "avoid handling Rules instances in config-validator" (#9295)
7d24dde Docs: Fix code snippet to refer to the correct option (#9313)
12388d4 �Chore: rewrite parseListConfig for a small perf gain. (#9300)
ce1f084 Update: fix MemberExpression handling in no-extra-parens (fixes #9156)
0c720a3 Update: allow autofixing when using processors (fixes #7510) (#9090)
838df76 Chore: upgrade deps. (#9289)
f12def6 Update: indent flatTernary option to handle return (fixes #9285) (#9296)
e220687 Fix: remove autofix for var undef inits (fixes #9231) (#9288)
002e199 Docs: fix no-restricted-globals wrong config. (#9305)
fcfe91a Docs: fix wrong config in id-length example. (#9303)

There are 55 commits in total.

See the full diff

Not sure how things should work exactly?

There is a collection of frequently asked questions and of course you may always ask my humans.

Your Greenkeeper Bot 🌴

[Perf] To use ES6Map or not to use ES6Map that is the question

Summary: It is comment in the code. But we should investigating whether to just use native ES6 map instead of Object.create(null)

An in-range update of eslint is breaking the build 🚨

Version 4.12.0 of eslint was just published.

Branch	Build failing 🚨
Dependency	eslint
Current Version	4.11.0
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

eslint is a devDependency of this project. It might not break your production code or affect downstream projects, but probably breaks your build or test tools, which may prevent deploying or publishing.

Status Details

✅ ci/circleci: build Your tests passed on CircleCI! Details
❌ ci/circleci: test:unit CircleCI is running your tests Details
❌ ci/circleci: test:style Your tests failed on CircleCI Details

Release Notes

v4.12.0

76dab18 Upgrade: doctrine@^2.0.2 (#9656) (Kevin Partington)
28c9c8e New: add a Linter#defineParser function (#9321) (Ives van Hoorne)
5619910 Update: Add autofix for sort-vars (#9496) (Trevin Hofmann)
71eedbf Update: add beforeStatementContinuationChars to semi (fixes #9521) (#9594) (Toru Nagashima)
4118f14 New: Adds implicit-arrow-linebreak rule (refs #9510) (#9629) (Sharmila Jesupaul)
208fb0f Fix: Use XML 1.1 on XML formatters (fixes #9607) (#9608) (Daniel Reigada)
6e04f14 Upgrade: globals to 11.0.1 (fixes #9614) (#9632) (Toru Nagashima)
e13d439 Fix: space-in-parens crash (#9655) (Toru Nagashima)
92171cc Docs: Updating migration guide for single-line disable (#9385) (Justin Helmer)
f39ffe7 Docs: remove extra punctuation from readme (#9640) (Teddy Katz)
a015234 Fix: prefer-destructuring false positive on "super" (fixes #9625) (#9626) (Kei Ito)
0cf081e Update: add importNames option to no-restricted-imports (#9506) (Benjamin R Gibson)
332c214 Docs: Add @platinumazure to TSC (#9618) (Ilya Volodin)

Commits

The new version differs by 15 commits.

f4a65c6 4.12.0
1cd1627 Build: changelog update for 4.12.0
76dab18 Upgrade: doctrine@^2.0.2 (#9656)
28c9c8e New: add a Linter#defineParser function (#9321)
5619910 Update: Add autofix for sort-vars (#9496)
71eedbf Update: add beforeStatementContinuationChars to semi (fixes #9521) (#9594)
4118f14 New: Adds implicit-arrow-linebreak rule (refs #9510) (#9629)
208fb0f Fix: Use XML 1.1 on XML formatters (fixes #9607) (#9608)
6e04f14 Upgrade: globals to 11.0.1 (fixes #9614) (#9632)
e13d439 Fix: space-in-parens crash (#9655)
92171cc Docs: Updating migration guide for single-line disable (#9385)
f39ffe7 Docs: remove extra punctuation from readme (#9640)
a015234 Fix: prefer-destructuring false positive on "super" (fixes #9625) (#9626)
0cf081e Update: add importNames option to no-restricted-imports (#9506)
332c214 Docs: Add @platinumazure to TSC (#9618)

See the full diff

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

Come up with a better package name

Is apollo-cache-hermes (and Hermes in general) really the package name we want? Surely there's a better name to go with!

Rename ApolloTransaction to DataProxy

This will make us more consistent with Apollo's interfaces, and hopefully reduce some confusion.

Only change parameterized edges (and paths to them) if their values change

Today, when overlaying parameterized edges on top of a query, we generate new objects every time. They should be generated only when their (transitive) values change.

Can we pre-calculate them per query on write? It might require pre-processing the user's schema?

An in-range update of eslint-plugin-typescript is breaking the build 🚨

Version 0.8.1 of eslint-plugin-typescript was just published.

Branch	Build failing 🚨
Dependency	eslint-plugin-typescript
Current Version	0.8.0
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

eslint-plugin-typescript is a devDependency of this project. It might not break your production code or affect downstream projects, but probably breaks your build or test tools, which may prevent deploying or publishing.

Status Details

✅ ci/circleci: build Your tests passed on CircleCI! Details
❌ ci/circleci: test:unit CircleCI is running your tests Details
❌ ci/circleci: test:style Your tests failed on CircleCI Details

Commits

The new version differs by 7 commits.

12220cc Chore: Update prettier and pin the version (#94)
3284af4 Fix: Mark call expression type parameters as used (#89)
29bd53c Fix: Mark qualified name type annotations as used (#88)
5e3999d Fix: Destructured param type annotation not marked used (fixes #82) (#84)
2b197fd Fix: Handles type spacing on TSParenthesizedType expressions (fixes #79) (#80)
45cd23e Fix: no-triple-slash-reference (fixes #75)
485254b Chore: Publish version 0.8.0 to npm

See the full diff

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

Support for readFragment & writeFragment

I'm trying Apollo 2.0, and didn't see much performance improvements using the default in-memory cache, in some cases it's actually worse than the Redux based version. I'm testing with pages that have 1000+ objects (or nodes, in Hermes terminology), so I decided to try with Hermes, as it seems to be fastest that any GraphQL cache out there.

When testing 0.3.0-alpha (and apollo-client 2.0.0-beta.4) I get the following error:

TypeError: this.dataStore.getCache(...).transformDocument is not a function
    at QueryManager.fetchQuery
    at QueryManager.startQuery
    at ObservableQuery.setUpQuery
    at ObservableQuery.onSubscribe
    at new Subscription
    at ObservableQuery.Observable.subscribe
    at Expansion._this.subscribeToQuery
    at Expansion._this.handleQuery
    at Expansion._this.updateWrapperQuery

This is how I'm creating the client. Note I'm using dataIdFromObject and an auth link:

...
import { Cache as HermesCache } from 'apollo-cache-hermes';

export const createApolloClient = ({
  databaseId,
  previewKey,
  uri = `/api/v1/databases/${databaseId}/apollo?preview=${previewKey}`,
  ssrMode = false,
}) => {
  ...

  return new ApolloClient({
    link,
    cache: new HermesCache({
      dataIdFromObject: object => object.id
    }),
    ssrMode,
  });
};

Is there anything else I should do to make the Hermes cache work?

[Improvement] Split test into different features

Entity Transformers

Add an option for entity Transformers to support things like domain object classes

Fix install from git

See #49

[Infra] Consolidate common functions among test files

There are a lot of duplicate code across test files. We should consolidate them and write helper functions/classes etc. to do all test setup work.

This works will be much easier after #136

An in-range update of jest-junit is breaking the build 🚨

Version 3.3.0 of jest-junit was just published.

Branch	Build failing 🚨
Dependency	jest-junit
Current Version	3.2.1
Type	devDependency

This version is covered by your current version range and after updating it in your project the build failed.

jest-junit is a devDependency of this project. It might not break your production code or affect downstream projects, but probably breaks your build or test tools, which may prevent deploying or publishing.

Status Details

❌ ci/circleci: build Your tests failed on CircleCI Details

Commits

The new version differs by 2 commits.

c35c423 Add unit test for testSuite->tests attribute existence and correctness
e78b16f Fix tests attribute in testSuite and bump to 3.3.0

See the full diff

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

Optimize writes for pagination

Walking ever-expanding lists from fetchMore causes ever-worse performance. Can we be smarter about pagination? (without changing Apollo's interface? Or do we need to propose something else?)

Use extract and restore helper functions in the real cache API method

Better handle ambiguous array updates

Today, Apollo's in-memory cache generates an id from the relative path to it (containerId + path).
Hermes similarly merges values. However, that approach leads to ambiguous state in the cache specifically with arrays.

For example, imagine the following updates to the cache:

{
  stuff: [
    { foo: 1, bar: 1 },
    { foo: 2, bar: 2 }
  ]
}

And then say that the server's view resorted stuff, and added a new entry; and then we select only bar fields:

{
  stuff: [
    { bar: 2 },
    { bar: 1 },
    { bar: 3 }
  ]
}

At this point, the cache will have merged state into:

{
  stuff: [
    { foo: 1, bar: 2 },
    { foo: 2, bar: 1 },
    { bar: 3 }
  ]
}

Which is incorrect from the server's perspective.

An in-range update of apollo-cache is breaking the build 🚨

Version 1.0.1 of apollo-cache was just published.

Branch	Build failing 🚨
Dependency	apollo-cache
Current Version	1.0.0
Type	dependency

This version is covered by your current version range and after updating it in your project the build failed.

apollo-cache is a direct dependency of this project, and it is very likely causing it to break. If other packages depend on yours, this update is probably also breaking those in turn.

Status Details

✅ ci/circleci: build Your tests passed on CircleCI! Details
❌ ci/circleci: test:unit CircleCI is running your tests Details
❌ ci/circleci: test:style Your tests failed on CircleCI Details

FAQ and help

There is a collection of frequently asked questions. If those don’t help, you can always ask the humans behind Greenkeeper.

Your Greenkeeper Bot 🌴

What is the equivalent of the apollo-cache-inmemory IntrospectionFragmentMatcher?

This is how I am currently using apollo-client v2:

// @flow

import {
  ApolloClient
} from 'apollo-client';
import {
  HttpLink
} from 'apollo-link-http';
import InMemoryCache, {
  IntrospectionFragmentMatcher
} from 'apollo-cache-inmemory';

const fragmentMatcher = new IntrospectionFragmentMatcher({
  introspectionQueryResultData: {
    // eslint-disable-next-line id-match
    __schema: {
      types: [
        {
          kind: 'INTERFACE',
          name: 'Paragraph',
          possibleTypes: [
            {
              name: 'TextParagraph'
            },
            {
              name: 'ImageParagraph'
            }
          ]
        },
        {
          kind: 'INTERFACE',
          name: 'Markup',
          possibleTypes: [
            {
              name: 'DecorationMarkup'
            },
            {
              name: 'LinkMarkup'
            }
          ]
        }
      ]
    }
  }
});

const apolloClient = new ApolloClient({
  cache: new InMemoryCache({
    fragmentMatcher: fragmentMatcher.match
  }),
  link: new HttpLink({
    credentials: 'include',
    uri: '/api'
  })
});

export default apolloClient;

I have tried migrating to apollo-cache-hermes, but I could not find the equivalent of IntrospectionFragmentMatcher in the code base.

Blacklist ROOT_MUTATION nodes

Today we cache them just the same as ROOT_QUERY nodes; but there's no good reason for doing so. It also generally leaks large amounts of posted data (via serialized variables) into the cache

convoyinc / apollo-cache-hermes Goto Github PK

apollo-cache-hermes's People

Contributors

Stargazers

Watchers

Forkers

apollo-cache-hermes's Issues

Version 8.0.1 of typescript-eslint-parser was just published.

Version 8.0.47 of @types/node was just published.

Version 2.6.1 of typescript was just published.

Version 1.0.2 of apollo-utilities was just published.

Version 4.14.79 of @types/lodash was just published.

Version 21.3.0 of eslint-plugin-jest was just published.

Version 4.0.6 of @types/chai was just published.

Version 4.7.0 of eslint just got published.

Version 4.12.0 of eslint was just published.

Version 0.8.1 of eslint-plugin-typescript was just published.

Version 3.3.0 of jest-junit was just published.

Version 1.0.1 of apollo-cache was just published.

Recommend Projects

Recommend Topics

Recommend Org