feat: plumb through context changes #268
Conversation
guseggert
force-pushed
the
feat/context
branch
from
6a25639 to
8e4ba02
Compare
We're pulling things up through every dependency of the go-ipfs and hydra-booster binaries. Any other consumers (e.g. lotus, estuary, indexer, ...) will have to fend for themselves.
Ah, thanks missed that one |
guseggert
force-pushed
the
feat/context
branch
5 times, most recently
from
8ad68a4 to
ecfc006
Compare
|
I've updated v2. And also the CLI, by plumbing through the first commit of this PR to the CLI in a second commit using pseudoversions...is that the preferred approach? |
v2/blockstore/readonly_test.go
Outdated
| @@ -18,13 +18,15 @@ import ( | |||
| "github.com/stretchr/testify/require" | |||
| ) | |||
|
|
|||
| var bg = context.Background() | |||
There was a problem hiding this comment.
please don't use a global context; I know it's a no-op and just used in tests, but I would rather avoid sharing any form of state between tests.
v2/blockstore/readwrite_test.go
Outdated
| @@ -32,6 +32,7 @@ var ( | |||
| rng = rand.New(rand.NewSource(1413)) | |||
| oneTestBlockWithCidV1 = merkledag.NewRawNode([]byte("fish")).Block | |||
| anotherTestBlockWithCidV0 = blocks.NewBlock([]byte("barreleye")) | |||
| bg = context.Background() | |||
v2/example_test.go
Outdated
| @@ -61,7 +62,7 @@ func ExampleWrapV1File() { | |||
| if err != nil { | |||
| panic(err) | |||
| } | |||
| fmt.Println(bs.Get(roots[0])) | |||
| fmt.Println(bs.Get(context.Background(), roots[0])) | |||
There was a problem hiding this comment.
for examples, I wonder if we should use context.TODO() instead, to highlight to the user that they should probably replace the context with something more useful - or consciously use a background context.
I guess it's not super important here, since the CAR blockstore doesn't yet use the context at all, but it might in the future.
guseggert
force-pushed
the
feat/context
branch
from
ecfc006 to
adf6d8a
Compare
|
This PR Is currently blocked b/c of a new change on master which adds a dependency on https://github.com/ipld/go-ipld-prime/tree/master/storage/bsadapter which uses the old datastore, causing the build to fail. We'll need to upgrade that and plumb it through first before this can be merged. |
|
FYI @willscott @warpfork |
|
I didnt see this PR, but i redid it here: #273 and it compiles and the tests pass and stuff |
|
It's not equivalent, we'll still want this one - it deals with the v2 module as well. |
|
v2 landed here #275 |
No description provided.