Create coded errors to cover all TransactionErrors from the RPC
#2213
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This was referenced Feb 28, 2024
steveluscher
force-pushed
the
02-28-Create_coded_errors_to_cover_all_TransactionErrors_from_the_RPC
branch
from
3585002
to
2f394a0
Compare
steveluscher
commented
Feb 28, 2024
| export const SOLANA_ERROR__TRANSACTION_ERROR_INVALID_ACCOUNT_FOR_FEE = 7050006 as const; | ||
| export const SOLANA_ERROR__TRANSACTION_ERROR_ALREADY_PROCESSED = 7050007 as const; | ||
| export const SOLANA_ERROR__TRANSACTION_ERROR_BLOCKHASH_NOT_FOUND = 7050008 as const; | ||
| // `InstructionError` intentionally omitted |
There was a problem hiding this comment.
I'll break these off into a separate special case in the next PR.
Comment on lines
+89
to
+101
| } else if (codeOffset === 29 /* DuplicateInstruction */) { | ||
| errorContext = { | ||
| index: transactionErrorContext as number, | ||
| }; | ||
| } else if ( | ||
| codeOffset === 30 /* InsufficientFundsForRent */ || | ||
| codeOffset === 34 /* ProgramExecutionTemporarilyRestricted */ | ||
| ) { | ||
| errorContext = { | ||
| accountIndex: (transactionErrorContext as { account_index: number }).account_index, | ||
| }; | ||
| } |
There was a problem hiding this comment.
These are all special cases, because the RPC spits them out as { ErrorName: { some: "struct" } } instead of "ErrorName".
steveluscher
force-pushed
the
02-28-Create_coded_errors_to_cover_all_TransactionErrors_from_the_RPC
branch
from
2f394a0
to
a737c14
Compare
buffalojoec
reviewed
Feb 29, 2024
There was a problem hiding this comment.
This lgtm. Shame about the map, but it makes sense. This utility is very useful and I believe was asked for at some point.
I think we should iron out the concept of maybe grouping error codes by discriminator, as you do here, before this goes in.
See #2205 (review)
steveluscher
force-pushed
the
02-28-Create_coded_errors_to_cover_all_TransactionErrors_from_the_RPC
branch
from
a737c14
to
8fdf687
Compare
This was referenced Feb 29, 2024
steveluscher
changed the base branch from
02-24-refactor_experimental_errors_codecs-core_package
to
02-29-Prevent_a_runtime_fatal_in_the_error_message_formatter_when_the_context_contains_null_or_undefined
steveluscher
force-pushed
the
02-28-Create_coded_errors_to_cover_all_TransactionErrors_from_the_RPC
branch
from
8fdf687
to
42eea0b
Compare
Merge activity
|
Base automatically changed from
02-29-Prevent_a_runtime_fatal_in_the_error_message_formatter_when_the_context_contains_null_or_undefined
to
master
February 29, 2024 17:52
steveluscher
force-pushed
the
02-28-Create_coded_errors_to_cover_all_TransactionErrors_from_the_RPC
branch
from
42eea0b
to
a8f8adf
Compare
steveluscher
deleted the
02-28-Create_coded_errors_to_cover_all_TransactionErrors_from_the_RPC
branch
steveluscher
added a commit
that referenced
this pull request
Feb 29, 2024
# Summary In this PR, we build atop the work in #2213 and introduce coded exceptions for each `InstructionError` returned from the RPC's `sendTransaction` method. > [!NOTE] > Because the RPC doesn't return structured errors or error codes, we had to break our own rules in this PR and hardcode a map between the error names and the code numbers. My [first crack](https://gist.github.com/steveluscher/aaa7cbbb5433b1197983908a40860c47#file-fml-ts-L12) at this employed a source code compression scheme that I later deemed too risky for the 250 gzipped bytes it saved. We might consider such a scheme in the future, especially since the next PR will add `InstructionError` to the mix. # Test Plan ```shell cd packages/errors pnpm test:unit:browser pnpm test:unit:node ``` Addresses #2118.
|
🎉 This PR is included in version 1.90.1 🎉 The release is available on: Your semantic-release bot 📦🚀 |
steveluscher
added a commit
that referenced
this pull request
Mar 12, 2024
# Summary So wow. I completely missed these as part of #2118, and it turns out they're a really big deal and required a ton of changes. There are: * Errors that have enough data to format a message * Errors that have no data, but also no context in the message * Errors that have no data, but really should, because you can't format a message without it * Preflight errors in which is nested a `TransactionError` (see #2213) In this PR we create a helper that takes in the `RpcSimulateTransactionResult` from the RPC and reformats it as a coded `SolanaError`. As always, everything you need to know is in the `packages/errors/src/__tests__/json-rpc-error-test.ts`. # Test Plan ``` pnpm turbo test:unit:browser pnpm turbo test:unit:node ```
|
Because there has been no activity on this PR for 14 days since it was merged, it has been automatically locked. Please open a new issue if it requires a follow up. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
Wouldn't it be nice if you could
catchparticular transaction errors in your application, likeBlockhashNotFound, and choose the correct mitigation based on the type of transaction error?In this PR, we introduce coded exceptions for each
TransactionErrorreturned from the RPC'ssendTransactionmethod.Note
Because the RPC doesn't return structured errors or error codes, we had to break our own rules in this PR and hardcode a map between the error names and the code numbers. My first crack at this employed a source code compression scheme that I later deemed too risky for the 250 gzipped bytes it saved. We might consider such a scheme in the future, especially since the next PR will add
InstructionErrorto the mix.Test Plan
cd packages/errors pnpm test:unit:browser pnpm test:unit:nodeAddresses #2118.