awsproviderlint: New Check: Prefer (*awserr.Error) instead of (*awserr.Error).Code() / (*awserr.Error).Message() in fmt.Errorf()

[breathingdust]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Description

The AWS Go SDK provides a good String() implementation of errors including the ErrorCode, ErrorMessage, and request ID for troubleshooting with AWS Support. We should avoid trimming out useful bits from these error messages by referencing Code(), Message(), etc. and instead simplify to just referencing the error object itself.

Flagged Code

fmt.Errorf("AWS error oh no: %s", err.Code())
fmt.Errorf("AWS error oh no: %s: %s", err.Code(), err.Message())

Passing Code

fmt.Errorf("AWS error oh no: %w", err)

[bflad]

Inexact (would need to do type checking with go/analysis) but covers this case well, a potential semgrep rule:

  - id: fmt-Errorf-awserr-Error-Code
    languages: [go]
    message: Prefer `err` with `%w` format verb instead of `err.Code()` or `err.Message()`
    paths:
      exclude:
        - vendor/
      include:
        - aws/
    patterns:
      - pattern-either:
        - pattern: fmt.Errorf(..., $ERR.Code(), ...)
        - pattern: fmt.Errorf(..., $ERR.Message(), ...)
    severity: WARNING

[github-actions]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.