sgl-project/sglang · error · ValueError

base64 URI header is too large

Error message

base64 URI header is too large

What it means

For 'base64://' scheme URIs, an optional media-type segment may precede a ','; that segment must stay within MINIMAX_H3_BASE64_HEADER_MAX_CHARS. The error fires when the comma-delimited header portion of a base64:// URI exceeds the cap.

Solutions

1. Ensure the optional header segment is short and followed by ',' before the payload
2. If no media type is needed, emit 'base64://<payload>' with no comma at all
3. Validate comma position (if present, must be within the header cap) at ingestion

Example fix

// before
uri = "base64://image/png" + payload  # missing comma
// after
uri = "base64://image/png," + payload  # or simply "base64://" + payload

if uri.startswith('base64://'):
    sep = uri.find(',', len('base64://'))
    if sep >= 0 and sep - len('base64://') > MINIMAX_H3_BASE64_HEADER_MAX_CHARS:
        raise ValueError('base64 URI header too large')

def base64_uri_header_ok(uri: str, cap: int) -> bool:
    sep = uri.find(',', len('base64://'))
    return sep < 0 or sep - len('base64://') <= cap

When it happens

Trigger: A base64:// URI whose text before the first ',' is longer than the header cap — typically because the comma separator was lost and the base64 payload itself is being counted as header.

Common situations: Malformed base64:// URIs where the payload was appended without the ',' separator, or a media-type segment stuffed with junk.

Related errors

• data URI header is too large
• tar material URI encoded header is too large
• data URI must contain a comma separator
• data URI must use ;base64 encoding
• material URI base64 payload must be ASCII

AI-assisted analysis of sgl-project/sglang@0132848349 (2026-08-28). Data as JSON: /api/errors/567d2bc8b76ffd11. Report an issue: GitHub.

Appendix: source

Thrown at python/sglang/multimodal_gen/runtime/pipelines_core/stages/model_specific_stages/minimax_h3/material_io.py:110

def _base64_uri_payload_start(uri: str) -> tuple[int, str | None]:
    media_type = None
    if uri.startswith("data:"):
        separator = uri.find(",")
        if separator < 0:
            raise ValueError("data URI must contain a comma separator")
        if separator > MINIMAX_H3_BASE64_HEADER_MAX_CHARS:
            raise ValueError("data URI header is too large")
        header = uri[:separator]
        if ";base64" not in header:
            raise ValueError("data URI must use ;base64 encoding")
        media_type = header[5:].split(";", 1)[0].lower() or None
        payload_start = separator + 1
    elif uri.startswith("base64://"):
        payload_start = len("base64://")
        separator = uri.find(",", payload_start)
        if separator >= 0:
            if separator - payload_start > MINIMAX_H3_BASE64_HEADER_MAX_CHARS:
                raise ValueError("base64 URI header is too large")
            header = uri[payload_start:separator]
            media_type = header.split(";", 1)[0].lower() or None
            payload_start = separator + 1
    else:  # pragma: no cover - guarded by the caller
        raise ValueError("not a base64 material URI")
    return payload_start, media_type


def _iter_base64_payload_bytes(uri: str, payload_start: int):
    """Yield validated, unquoted base64 bytes without copying the payload."""

    index = payload_start
    while index < len(uri):
        character = uri[index]
        if character == "%":
            if index + 2 >= len(uri):
                raise ValueError("material URI has an invalid percent escape")
            try:

View on GitHub (pinned to 0132848349)