chore: upgrade deps with new typedefs #3550

achingbrain · 2021-02-15T11:25:18Z

Upgrades to new version with types

Uses default aegir ts config
Fixes all ts errors
Fully types core-api in ipfs-core-types package
Makes ipfs-core implement types from ipfs-core-types package
Removes duplicate types, ipfs-core-types as single source of type truth
Reduces use of external APIs by internal components in ipfs-core

Depends on:

To do:

packages/ipfs-core/src/components/files/utils/add-link.js

packages/ipfs-core/src/components/files/utils/hamt-utils.js

rvagg · 2021-03-12T06:25:06Z

packages/ipfs-core/src/components/files/chmod.js

              yield async function () {
-                const cid = await persist(entry.content.serialize(), block, opts)
+                /** @type {DAGNode} */
+                // @ts-ignore


I think you get to remove this ignore now

rvagg · 2021-03-12T06:28:35Z

packages/ipfs-core/src/components/files/stat.js

-      size: size,
-      cumulativeSize: cumulativeSize,
-      blocks: blocks,
+      size: file.unixfs.fileSize(),


do we have a guarantee now that there is a file.unixfs?

rvagg · 2021-03-12T06:30:37Z

packages/ipfs-core/src/components/files/stat.js

+      mode: file.unixfs.mode
    }

-    if (file.unixfs) {


the things I think are missing from this deleted block now are - stat.type setting (it's always 'file' now), stat.size set to 0 for isDirectory(), should these be added back?

... or maybe this is now safely disambiguating the file vs directory case before we even get here ... I guess that makes them unnecessary here!

rvagg · 2021-03-12T08:20:12Z

packages/ipfs-core/src/components/files/utils/hamt-utils.js

      }

-      return (rootBucket || bucket).put(link.Name.substring(2), {
+      return rootBucket.put(link.Name.substring(2), {


I'm not seeing the guarantee in the changes here that there will be a rootBucket at this point. It's added in recreateHamtLevel if there's not a parentBucket, but are we sure that there's always a rootBucket if there's a parentBucket? A quick git grep suggests that recreateHamtLevel either has only the first arg or all of the args, but I don't know whether those args always have a non-null value.

rvagg · 2021-03-12T08:22:02Z

packages/ipfs-core/src/components/files/utils/remove-link.js

  log(`Updating shard ${prefix} with name ${newName}`)

-  const size = DAGNode.isDAGNode(result.node) ? result.node.size : result.node.Tsize
+  const size = result.node instanceof DAGNode ? result.node.size : result.node.Tsize


I always get told off by browser people for using instanceof, isn't this something we're supposed to avoid by duck typing at a minimum?

ipfs/js-ipfs#3550 switches away from default exports for `ipfs-http-client` and `ipfs-client` to enable an easier transition to es6 modules. All `ipfs` modules now export a `.create` factory function which returns instances of the client module. BREAKING CHANGE: expects `ipfs-http-client` and `ipfs-client` to export a `.create` function

hugomrdias

partial review
some comments inline, im a little worried that our inputs are forcing the users to pull in CID, PeerId, multiaddr, ipld-block etc etc
i know we talked about this a million times and i understand why its like this but i dunno if its the right way.
So if we force the input to be the instance of those classes we should do it everywhere, and put an @see 'url to package' in the js docs.
Or we can accept the low level version of those like string for CID or UInt8array for Block etc.

packages/ipfs-cli/src/commands/add.js

packages/ipfs-cli/src/commands/bitswap/stat.js

packages/ipfs-cli/src/commands/bitswap/unwant.js

packages/ipfs-cli/src/commands/bitswap/wantlist.js

packages/ipfs-cli/src/commands/block/put.js

hugomrdias · 2021-03-22T12:45:05Z

packages/ipfs-core-types/src/basic.d.ts

@@ -0,0 +1,136 @@
+import CID from 'cids'
+import { AwaitIterable } from './basic'


is this importing itself ?

Oops, looks like it.

hugomrdias · 2021-03-22T12:46:20Z

packages/ipfs-core-types/src/basic.d.ts

+ */
+export type ToContent =
+  | string
+  | InstanceType<typeof String>


this can't be just string ?

If we want to support 'string' and new String('string') then no as they are different.

Who does that though?

lol right, this is just a nit so feel free to ignore

packages/ipfs-core-types/src/block/index.d.ts

packages/ipfs-core-types/src/files/index.d.ts

ipfs/js-ipfs#3550 switches away from default exports for `ipfs-http-client` and `ipfs-client` to enable an easier transition to es6 modules. All `ipfs` modules now export a `.create` factory function which returns instances of the client module. BREAKING CHANGE: expects `ipfs-http-client` and `ipfs-client` to export a `.create` function

- Uses default aegir ts config - Fixes all ts errors - Fully types core-api in ipfs-core-types package - Makes ipfs-core implement types from ipfs-core-types package - Removes duplicate types, ipfs-core-types as single source of type truth - Reduces use of external APIs by internal components in ipfs-core

Upgrades to new version with types - Uses default aegir ts config - Fixes all ts errors - Fully types core-api in ipfs-core-types package - Makes ipfs-core implement types from ipfs-core-types package - Removes duplicate types, ipfs-core-types as single source of type truth - Reduces use of external APIs by internal components in ipfs-core - Switches to named exports BREAKING CHANGE: all core api methods now have types, some method signatures have changed, named exports are now used by the http, grpc and ipfs client modules

achingbrain requested a review from hugomrdias February 17, 2021 17:34

This was referenced Feb 17, 2021

feat: add types ipld/js-ipld-dag-pb#189

Merged

cids should be a dependency, not a devDependency ipfs/js-ipfs-unixfs#118

Closed

hugomrdias suggested changes Feb 23, 2021

View reviewed changes

packages/ipfs-core/src/components/files/utils/add-link.js Show resolved Hide resolved

packages/ipfs-core/src/components/files/utils/hamt-utils.js Outdated Show resolved Hide resolved

achingbrain changed the title ~~chore: upgrade unixfs deps~~ chore: upgrade deps with new typedefs Mar 5, 2021

rvagg reviewed Mar 12, 2021

View reviewed changes

achingbrain force-pushed the chore/update-unixfs branch from 8d63c7c to e443160 Compare March 13, 2021 14:53

achingbrain mentioned this pull request Mar 15, 2021

Track: Typescript types support #2945

Closed

BigLep added the status/in-progress In progress label Mar 15, 2021

achingbrain mentioned this pull request Mar 21, 2021

chore: use .create function for http client and ipfs client ipfs/js-ipfsd-ctl#616

Merged

achingbrain linked an issue Mar 21, 2021 that may be closed by this pull request

Will not build on Windows 10 64-bit #3581

Closed

hugomrdias suggested changes Mar 22, 2021

View reviewed changes

hugomrdias approved these changes Mar 25, 2021

View reviewed changes

achingbrain mentioned this pull request Mar 25, 2021

fix: allow use in bundlers rvagg/cborg#10

Closed

achingbrain mentioned this pull request Mar 26, 2021

chore: update ipfsd-ctl to use http client create method ipfs/interop#257

Merged

achingbrain force-pushed the chore/update-unixfs branch from 3f94436 to df7e158 Compare March 27, 2021 15:10

achingbrain added 7 commits March 29, 2021 11:40

chore: remove peer-id from core-ipfs-types

a8e5e7c

chore: fix tests

15af2ae

chore: rename basic.d.ts to utils.d.ts

c69f996

chore: increase timeout for webworkers

583fb85

chore: revert previous commit

83c54b1

chore: disable ff webworker tests

e57c82c

chore: disable ff webworker tests

d247aa0

achingbrain added 4 commits March 30, 2021 16:41

chore: replace bignumber with bigint

044bd61

chore: fix stat output

923d52b

chore: fix tests

706d795

chore: linting

ab3be7a

achingbrain merged commit a418a52 into master Mar 31, 2021

achingbrain deleted the chore/update-unixfs branch March 31, 2021 13:49

achingbrain removed the status/in-progress In progress label Apr 1, 2021

		@@ -0,0 +1,136 @@
		import CID from 'cids'
		import { AwaitIterable } from './basic'

chore: upgrade deps with new typedefs #3550

chore: upgrade deps with new typedefs #3550

Uh oh!

Conversation

achingbrain commented Feb 15, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

hugomrdias left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

achingbrain commented Feb 15, 2021 •

edited

Loading