Golint: fix golint warnings in merkledag submodule #4665

hsanjuan · 2018-02-06T16:08:25Z

This fixes all golint warnings in merkledag (except in protobuf generated code).

Main change of note is that NewDAGService returns an ipld.DAGService and not a non-exported dagService.

License: MIT Signed-off-by: Hector Sanjuan <hector@protocol.ai>

whyrusleeping · 2018-02-06T18:33:12Z

merkledag/merkledag.go

 // NewDAGService constructs a new DAGService (using the default implementation).
-func NewDAGService(bs bserv.BlockService) *dagService {


Does golint really complain about this?

I actually prefer to return concrete types whenever possible. It helps to deal with multiple potentially fulfilled interfaces. For example, if *dagService implements X which is a superset of DAGService, I can pass NewDAGService to something that wants an X without extra type assertions.

Want me to export dagService instead?

whyrusleeping · 2018-02-06T18:34:04Z

merkledag/merkledag_test.go

@@ -241,9 +242,10 @@ func TestFetchGraph(t *testing.T) {
 	// create an offline dagstore and ensure all blocks were fetched
 	bs := bserv.New(bsis[1].Blockstore(), offline.Exchange(bsis[1].Blockstore()))

-	offline_ds := NewDAGService(bs)
+	// we know the default dagService implements LinkGetter
+	offlineDS := NewDAGService(bs).(ipld.LinkGetter)


This is exactly the case for having NewDAGService return the concrete type.

But it is a case only used in tests (and tests in the same module). From an external user's point of view, it's not possible to know what *dagService can do because it's unexported.

whyrusleeping · 2018-02-06T18:34:57Z

Thank you!!

Everything except the DAGService return type LGTM... If golint really complains about that it makes me a tad bit sad.

kevina · 2018-02-06T19:09:08Z

Everything except the DAGService return type LGTM... If golint really complains about that it makes me a tad bit sad.

I agree and I want to emphasis that we don't have to follow all off golints suggestions. If this will cause codeclimate to complain we should add exception.

Stebalien · 2018-02-06T19:32:12Z

Everything except the DAGService return type LGTM... If golint really complains about that it makes me a tad bit sad.

It does this for a (possibly debatable) reason. The lint disallows returning unexported types because they can't be named outside the package.

License: MIT Signed-off-by: Hector Sanjuan <hector@protocol.ai>

hsanjuan · 2018-02-06T23:41:26Z

Ok, I have reverted to what it was before and get the expected:

merkledag.go:33:43: exported func NewDAGService returns unexported type *merkledag.dagService, which can be annoying to use

If it is valuable that this is used as a LinkGetter (which is not except in tests), I would lean to exporting it though. The user won't have a clue that this returns a LinkGetter without looking at sources otherwise. It's a nuance anyway, I'm fine with whatever.

kevina · 2018-02-07T04:56:15Z

merkledag/errservice.go

@@ -14,28 +14,34 @@ type ErrorService struct {

 var _ ipld.DAGService = (*ErrorService)(nil)

+// Add returns an error.


I find this description confusing, rather than say "an error" something like "returns ErrorService.Err" or maybe just "returns the error" rather than "an error".

kevina · 2018-02-07T04:56:23Z

merkledag/errservice.go

 func (cs *ErrorService) Add(ctx context.Context, nd ipld.Node) error {
 	return cs.Err
 }

+// AddMany returns an error.


kevina · 2018-02-07T04:56:32Z

merkledag/errservice.go

 func (cs *ErrorService) AddMany(ctx context.Context, nds []ipld.Node) error {
 	return cs.Err
 }

+// Get returns an error.


kevina · 2018-02-07T04:58:37Z

merkledag/merkledag.go

@@ -23,7 +23,13 @@ func init() {
 	ipld.Register(cid.DagCBOR, ipldcbor.DecodeBlock)
 }

+// contextKey is a type to use as value for the ProgressTracker contexts.
+type contextKey string


I would export this type.

It has no use outside. If someone ever has a use, it can be exported then.

Yeah, your right. Sorry.

@kevina

Per @kevina's comments License: MIT Signed-off-by: Hector Sanjuan <hector@protocol.ai>

Stebalien · 2018-02-07T16:49:55Z

FYI,

If it is valuable that this is used as a LinkGetter (which is not except in tests), I would lean to exporting it though. The user won't have a clue that this returns a LinkGetter without looking at sources otherwise. It's a nuance anyway, I'm fine with whatever.

In general, users shouldn't care if something is a LinkGetter. They should just call ipld.GetLinks(ctx, dag, cid) which will, internally, try to cast dag to a LinkGetter if the passed dag implements that interface.

Stebalien

One import nit. Otherwise, LGTM. ❤️

Stebalien · 2018-02-07T17:04:13Z

merkledag/merkledag_test.go

@@ -13,6 +13,8 @@ import (
 	"testing"
 	"time"

+	blocks "gx/ipfs/Qmej7nf81hi2x2tvjRBF3mcp74sQyuDH4VMYDGd1YtXjb2/go-block-format"


This should be grouped with cid, u, etc. (we group by standard, in-repo, then out-of-repo).

License: MIT Signed-off-by: Hector Sanjuan <hector@protocol.ai>

hsanjuan · 2018-02-07T18:28:43Z

opps. fixed @Stebalien

ghost assigned hsanjuan Feb 6, 2018

ghost added the status/in-progress In progress label Feb 6, 2018

Golint: fix golint warnings in merkledag submodule

2d138b0

License: MIT Signed-off-by: Hector Sanjuan <hector@protocol.ai>

hsanjuan force-pushed the doc/golint-merkledag branch from f0b163e to 2d138b0 Compare February 6, 2018 16:45

whyrusleeping reviewed Feb 6, 2018

View reviewed changes

Golint: merkledag: let NewDagService return *dagService

2743dd2

License: MIT Signed-off-by: Hector Sanjuan <hector@protocol.ai>

kevina suggested changes Feb 7, 2018

View reviewed changes

Golint: improve comments in merkledag/errservice.go

d04de00

Per @kevina's comments License: MIT Signed-off-by: Hector Sanjuan <hector@protocol.ai>

kevina approved these changes Feb 7, 2018

View reviewed changes

Stebalien requested changes Feb 7, 2018

View reviewed changes

golint: merkledag_test: reorder imports

69497a6

License: MIT Signed-off-by: Hector Sanjuan <hector@protocol.ai>

Stebalien approved these changes Feb 7, 2018

View reviewed changes

whyrusleeping merged commit 12354cc into master Feb 7, 2018

ghost removed the status/in-progress In progress label Feb 7, 2018

whyrusleeping deleted the doc/golint-merkledag branch February 7, 2018 20:40

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Golint: fix golint warnings in merkledag submodule #4665

Golint: fix golint warnings in merkledag submodule #4665

hsanjuan commented Feb 6, 2018

whyrusleeping Feb 6, 2018

hsanjuan Feb 6, 2018

whyrusleeping Feb 6, 2018

hsanjuan Feb 6, 2018

whyrusleeping commented Feb 6, 2018

kevina commented Feb 6, 2018 •

edited

Loading

Stebalien commented Feb 6, 2018

hsanjuan commented Feb 6, 2018

kevina Feb 7, 2018

kevina Feb 7, 2018

kevina Feb 7, 2018

kevina Feb 7, 2018

hsanjuan Feb 7, 2018

kevina Feb 7, 2018

Stebalien commented Feb 7, 2018

Stebalien left a comment

Stebalien Feb 7, 2018

hsanjuan commented Feb 7, 2018

		// NewDAGService constructs a new DAGService (using the default implementation).
		func NewDAGService(bs bserv.BlockService) *dagService {

		@@ -14,28 +14,34 @@ type ErrorService struct {

		var _ ipld.DAGService = (*ErrorService)(nil)

		// Add returns an error.

Golint: fix golint warnings in merkledag submodule #4665

Golint: fix golint warnings in merkledag submodule #4665

Conversation

hsanjuan commented Feb 6, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

whyrusleeping commented Feb 6, 2018

kevina commented Feb 6, 2018 • edited Loading

Stebalien commented Feb 6, 2018

hsanjuan commented Feb 6, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Stebalien commented Feb 7, 2018

Stebalien left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

hsanjuan commented Feb 7, 2018

kevina commented Feb 6, 2018 •

edited

Loading