Daniel Porteous (dport)
Daniel Porteous (dport)
## Description This PR moves the existing TS SDK into a `v0` directory and makes a new `v1` directory for the new SDK. The `v1` SDK is initially a copy...
## Summary This PR makes various improvements to metrics (per-endpoint logging), docs, spec generation (a new standalone binary), deduplication, and linting (to ensure spec changes are checked in). I can...
In the Poem based API (v1) there are few places where I use the `impl_poem_type` and `impl_poem_parameter` macros. These macros are hacks to deal with types that I can't apply...
There are certain types that are only ever used in requests, and others that are only ever used in responses. For those we can use write only and read only...
``` $ curl http://localhost:8080/v1/transactions?limit=1 | jq '.[] | length' 18 ``` This command output pretty much says it all. There is a bug somewhere in how we calculate limits (likely...
``` $ curl http://localhost:8080/v1/events/0x0000000000000000caa60eb4a01756955ab9b2d1caca52edcaa60eb4a0175aaaaaaaaaaaaaaaaaaa [] ``` Pretty self-explanatory, this not be a 200 with an empty response, it should be a 404 because the requested account doesn't exist. Note: This...
Same as https://github.com/aptos-labs/aptos-core/issues/2284 but for transactions. Example: ``` curl http://localhost:8080/v1/accounts/0x04343443/transactions ```
Currently there is a bit of code duplication and some nasty coupling between these different components, where events is something of an extension on top of accounts. As such, much...
Currently every endpoint handler has something like this as the first line: ``` fail_point_poem("endpoint_get_table_item")?; ``` This could realistically be middleware, all it's doing is just making the endpoint bail out...
Currently the spec files stored alongside the API (in `api/doc/v1`) are generated by hand, meaning they can fall out of sync with reality. We should generate these files automatically as...