wasmerio/wasmer · error

The page is non-empty

Error message

The page is non-empty

What it means

In the backend-api GraphQL query paging logic, after filtering out messages from a fetched page, the code takes the last element with .expect("The page is non-empty") to advance the pagination cursor. The library assumes that whenever it reaches this branch the page returned at least one message; if the page is empty at this point the invariant is violated and it panics.

Source

Thrown at lib/backend-api/src/query.rs:2366

                if page.is_empty() {
                    if watch {
                        /*
                         * [TODO]: The resolution here should be configurable.
                         */

                        #[cfg(all(target_arch = "wasm32", target_os = "unknown"))]
                        std::thread::sleep(Duration::from_secs(1));

                        #[cfg(not(all(target_arch = "wasm32", target_os = "unknown")))]
                        tokio::time::sleep(Duration::from_secs(1)).await;

                        continue;
                    }

                    break Ok(None);
                } else {
                    let last_message = page.last().expect("The page is non-empty");
                    let timestamp = last_message.timestamp;
                    // NOTE: adding 1 microsecond to the timestamp to avoid fetching
                    // the last message again.
                    let timestamp = OffsetDateTime::from_unix_timestamp_nanos(timestamp as i128)
                        .with_context(|| {
                            format!("Unable to interpret {timestamp} as a unix timestamp")
                        })?;

                    // FIXME: We need a better way to tell the backend "give me the
                    // next set of logs". Adding 1 nanosecond could theoretically
                    // mean we miss messages if multiple log messages arrived at
                    // the same nanosecond and the page ended midway.

                    let next_timestamp = timestamp + Duration::from_nanos(1_000);

                    break Ok(Some((page, next_timestamp)));
                }
            }

View on GitHub (pinned to 8c4b9ee9d3)

Solutions

1. Upgrade wasmer-backend / wasmer to a version with fixed pagination edge-case handling
2. Check the channel data via the API directly; remove or repair the malformed message(s) that are being skipped
3. Retry the query later if the API is returning transient inconsistent pages
4. Work around by paginating manually with smaller page sizes so the filtered page is never fully empty

Example fix

// before
let last_message = page.last().expect("The page is non-empty");
// after
let last_message = match page.last() {
    Some(m) => m,
    None => break Ok(None),
};

// The panic occurs inside library query code; catch at the call boundary
let result = std::panic::catch_unwind(|| query_last_message(connector))
    .map_err(|p| anyhow::anyhow!("last_message pagination panicked: {:?}", p))?;

When it happens

Trigger: Calling code that queries a chat/message channel's last_message via the paginated query when the API returns a page where every entry is filtered out by the preceding conditions (e.g. all messages fail validation or are skipped), leaving page empty at the .last() call.

Common situations: Using wasmer backend-api tooling against a channel with only malformed/skipped messages, or an API schema change that makes the filter drop all items in a page.

Related errors

• cron job '{cron_job}' not found
• invalid node type returned
• App deletion failed for an unknown reason
• No value found for secret with name '{}'
• not a `sys` backend!

AI-assisted analysis of wasmerio/wasmer@8c4b9ee9d3 (2026-09-01). Data as JSON: /api/errors/ab5c719c5d6edb9b. Report an issue: GitHub.