Submodule conversion #89
Conversation
|
grod220
force-pushed
the
submodule-conversion
branch
from
c9bf54c to
a1b9999
Compare
grod220
force-pushed
the
submodule-conversion
branch
from
a1b9999 to
a452ede
Compare
.github/workflows/rust-checks.yml
Outdated
| execute_and_check: | ||
| name: Generate registry and Check for Diffs | ||
| runs-on: ubuntu-latest | ||
| steps: | ||
| - uses: actions/checkout@v3 | ||
| with: | ||
| submodules: 'true' | ||
| - uses: actions-rs/toolchain@v1 | ||
| with: | ||
| profile: minimal | ||
| toolchain: stable | ||
| override: true | ||
| - name: Generate registry | ||
| run: cargo run | ||
| env: | ||
| RUST_LOG: info | ||
| working-directory: tools/compiler | ||
| - name: Check for Uncommitted Changes | ||
| run: git diff --exit-code | ||
| working-directory: tools/compiler |
There was a problem hiding this comment.
Allows us now to ensure folks are properly running the script and not doing manual edits
npm/package.json
Outdated
There was a problem hiding this comment.
Cutting a new release for consumers using bundled registries
| @@ -1,4 +1,4 @@ | |||
| import * as Deimos8 from '../../registry/chains/penumbra-testnet-deimos-8.json'; | |||
| import * as Deimos8 from '../../registry/chains/penumbra-testnet-deimos-8-x6de97e39.json'; | |||
There was a problem hiding this comment.
It looks like the penumbra-testnet-deimos-8.json has been removed from the inputs. To pass the test, these need to be updated.
There was a problem hiding this comment.
Related to #73, last time L and I were debugging. IIRC we were surprised to see that the relevant codepaths were dropping the -x6de97e39 and trying to import the suffix-less JSON file, which is not what I wanted. I suspect this is still an issue, but we can follow up and file if testing surfaces it again.
There was a problem hiding this comment.
Discussed offline: made an update to the chain id swapping logic to only do so if the exact match fails
registry/chains/penumbra-1.json
Outdated
There was a problem hiding this comment.
Diff due to pulling from the latest commit from the cosmos registry
There was a problem hiding this comment.
I don't think anyone is depending on this (?). But the input that generates this seems to have been deleted a while ago.
| futures = "0.3.30" | ||
| serde = { version = "1.0.204", features = ["derive"] } | ||
| serde_json = "1.0.120" | ||
| regress = "0.10.0" | ||
| reqwest = { version = "0.12.5", features = ["json"] } | ||
| tempdir = "0.3.7" | ||
| thiserror = "1.0.63" | ||
| tokio = { version = "1.39.1", features = ["full"] } |
There was a problem hiding this comment.
No longer requiring async libs
grod220
force-pushed
the
submodule-conversion
branch
from
e3e2265 to
000b796
Compare
Applies yamllint reccos and updates to latest version of the 'checkout' action.
Moves out of the Rust "src/" dir because they're not source code files. Updates references accordingly.
conorsch
force-pushed
the
submodule-conversion
branch
from
6f8efd9 to
eadfd75
Compare
|
Tacked on a few nice-to-haves: yaml linting, more logging, and 🚨 moved the submodule from |
Closes #5
Closes #85
The initial motivation to use http requests has proven to be more finicky than hoped for. It introduces failure possibility during runtime when the submodule strategy pulls those issues in only after the cosmos registry dep is updated. This will allow us to have deterministic builds for the current commit and worry less about the
cargo runcommand failing when we are trying to add another chain id. The downside is that it requires 150mb download when pulling down for the first time.