[RUM-7478] Encode non-encodable attribute values as their string description

[mariedm]

What and why?

When a non-encodable value is passed as an attribute, the entire event was previously dropped. This fixes the gap for AnyEncodable by falling back to String(describing:) as a last resort, consistent with Android SDK behavior.

This is particularly relevant for cross-platform SDKs (e.g. Kotlin Multiplatform) where objects bridged from other languages aren't Swift-encodable but have a meaningful toString() / description.

How?

In AnyEncodable's default encoding case (previously a hard throw), the value's string description is now encoded instead. A debug log is emitted to inform developers when this fallback is triggered.

Review checklist

• Feature or bugfix MUST have appropriate tests (unit, integration)
• Make sure each commit and the PR mention the Issue number or JIRA reference
• Add CHANGELOG entry for user facing changes
• Add Objective-C interface for public APIs - see our guidelines (internal)
• Run make api-surface when adding new APIs

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 8d17abd1b0

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Truncate fallback descriptions before encoding

When a Log user attribute wraps a non-Encodable value whose description is longer than the 25,600 UTF-16-unit attribute limit, this fallback creates and encodes the string after LogEventSanitizer has already run. The sanitizer only truncates values it can already decode as String (DatadogLogs/Sources/Log/LogEventSanitizer.swift, AttributesSanitizer.sanitizeValues), so these new fallback strings bypass the documented backend hard limit and can produce over-limit log payloads instead of the same truncated string behavior as normal string attributes.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Avoid double-quoting fallback span attributes

For Trace spans, non-String tag/user/account attributes are converted to strings by JSON-encoding AnyEncodable in SpanEventBuilder.castValuesToString; with this fallback, a non-encodable value whose description is foo first encodes as the JSON string "foo", and that raw JSON literal is then emitted as the meta.* value. This makes affected span attributes contain extra quote characters (unlike normal String tags, which are emitted as foo), so ObjC/manual AnyEncodable span tags no longer carry the promised string description.

Useful? React with 👍 / 👎.