Skip to content

BREAKING.md

Latest commit

 

History

History
119 lines (89 loc) · 3.7 KB

BREAKING.md

File metadata and controls

119 lines (89 loc) · 3.7 KB

Breaking Changes

Version 1.6

libdns 1.0 Compatibility

Version 1.6 requires libdns v1.0 or later. The libdns v1.0 release introduced typed record structs that replace the generic libdns.Record type. This is a fundamental change to the libdns API.

What Changed in libdns 1.0

  • Typed Records: Instead of using generic libdns.Record structs, libdns v1.0 introduced typed record implementations like libdns.Address, libdns.TXT, libdns.SRV, etc.
  • Parse() Method: The new Record interface includes a Parse() method that returns typed structs
  • RR() Method: All record types implement RR() to get the underlying resource record data

Migration for libdns 1.0

See the libdns documentation for complete details on migrating to typed records.

Example of the new API:

// Old (libdns <1.0)
records := []libdns.Record{
    {
        Type:  "A",
        Name:  "www",
        Value: "1.2.3.4",
        TTL:   300 * time.Second,
    },
}

// New (libdns >=1.0)
records := []libdns.Record{
    libdns.Address{
        Name:  "www",
        Value: netip.MustParseAddr("1.2.3.4"),
        TTL:   300 * time.Second,
    },
}

Field Renames

Two provider configuration fields have been renamed for clarity:

1. MaxWaitDur → Route53MaxWait

Old (pre-v1.6):

provider := &route53.Provider{
    MaxWaitDur: 60,  // Was treated as seconds (multiplied by time.Second internally)
}

New (v1.6+):

provider := &route53.Provider{
    Route53MaxWait: 60 * time.Second,  // Use proper time.Duration
}

Important: In versions before v1.6, MaxWaitDur was silently multiplied by time.Second in the provider's init function. This was non-idiomatic Go and has been fixed. You must now provide a proper time.Duration value (like 60 * time.Second or 2 * time.Minute), as is standard in Go.

Failure to multiply by time.Second will result in a 60-nanosecond timeout instead of 60 seconds!

Rationale: The new name clearly indicates this is a Route53-specific timeout for AWS internal propagation, not general DNS propagation.

2. WaitForPropagation → WaitForRoute53Sync

Old (pre-v1.6):

provider := &route53.Provider{
    WaitForPropagation: true,
}

New (v1.6+):

provider := &route53.Provider{
    WaitForRoute53Sync: true,
}

Rationale: The new name clearly indicates this waits for Route53's internal synchronization, not worldwide DNS propagation (which can take hours depending on TTL values).

Removed Deprecated Fields

Two deprecated fields have been removed in v1.6:

  • AWSProfile → Use Profile instead
  • Token → Use SessionToken instead

These fields were deprecated several versions ago and have identical functionality to their replacements.

// Old (removed in v1.6)
provider := &route53.Provider{
    AWSProfile: "my-profile",
    Token:      "my-session-token",
}

// New (v1.6+)
provider := &route53.Provider{
    Profile:      "my-profile",
    SessionToken: "my-session-token",
}

JSON Configuration: If using JSON config, update field names: aws_profile → profile, token → session_token

Migration Checklist

  • Update to libdns v1.0+ (see libdns documentation for typed records)
  • Rename MaxWaitDur to Route53MaxWait in your code
  • Change from plain integer (e.g., 60) to proper time.Duration (e.g., 60 * time.Second)
  • Rename WaitForPropagation to WaitForRoute53Sync in your code
  • Replace AWSProfile with Profile (if using)
  • Replace Token with SessionToken (if using)
  • Update JSON/YAML configuration files with new field names
  • Test your code thoroughly after migration