fix(astrbot): name the wrong-endpoint case instead of calling it a bad image
一些检测仍在等待运行
ci / Portable quality checks (push) Waiting to run
ci / Windows Tauri build (push) Waiting to run

The capture settled why an attachment rendered as a broken image:
HTTP 200, Content-Type text/html, 201 bytes. AstrBot answers an unknown
/api/v1 path with its dashboard shell rather than a 404, so a wrong
download endpoint arrives looking like a successful download of an
undecodable file. The status check passed it and the type sniffer
correctly refused to call it an image, which left the user reading
"not a decodable image format" for what is really a wrong URL.

Reject an HTML body outright and say so. GET /api/v1/files/{id}/content
is a guess — upload posts to /api/v1/file, singular — and the protocol
review records the file endpoints as never tested.

Also trace the shape of the reference the media frame carries, which
the client currently discards in favour of the id from
attachment_saved. Scheme, whether it looks like a path, and the file
extension are recorded; the reference itself is not.

Refs: N-6.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
这个提交存在于:
YokiiroBW 2026-07-27 02:37:52 +08:00
父节点 ca65c465b9
当前提交 2ba07bda9d

查看文件

@ -529,6 +529,16 @@ impl AstrBotClient {
bytes = bytes.len(),
"astrbot attachment download"
);
// A short HTML body served with 200 is a web page, not a file. AstrBot
// answers an unknown `/api/v1/...` path with its dashboard shell rather
// than a 404, so a wrong endpoint arrives here looking like success.
if mime_type.starts_with("text/html") {
return Err(AstrBotClientError::ApiStatus(
"服务器在附件地址上返回了网页而不是文件,下载端点可能不正确。".into(),
));
}
Ok((mime_type, bytes))
}
@ -792,6 +802,39 @@ thread_local! {
}
impl FrameTrace {
/// Structural description of an attachment reference.
///
/// The media frame carries a string the client currently discards, and the
/// id taken from `attachment_saved` resolves to a web page. Whether that
/// string is a URL, a path, or a bare name decides how the attachment
/// should be fetched. Records the scheme, whether it looks like a path, and
/// the file extension — a format marker — but not the reference itself.
fn reference_shape(text: &str) -> String {
let scheme = if text.starts_with("http://") {
"http"
} else if text.starts_with("https://") {
"https"
} else if text.starts_with("file://") {
"file"
} else if text.starts_with('/') {
"absolute-path"
} else if text.starts_with('[') {
"bracketed"
} else {
"bare"
};
let extension = text
.rsplit_once('.')
.map(|(_, tail)| tail)
.filter(|tail| tail.len() <= 5 && tail.chars().all(|c| c.is_ascii_alphanumeric()))
.unwrap_or("<none>");
format!(
"scheme={scheme} has_separator={} extension={extension} len={}",
text.contains('/'),
text.chars().count()
)
}
fn observe(&mut self, line: &str) {
if !tracing::enabled!(target: FRAME_TRACE_TARGET, tracing::Level::DEBUG) {
return;
@ -814,6 +857,14 @@ impl FrameTrace {
self.blocks.push(text.to_owned());
}
(Some("complete"), Some(final_text)) => self.report(final_text),
(Some("image" | "record" | "file" | "video"), Some(reference)) => {
tracing::debug!(
target: FRAME_TRACE_TARGET,
media_type = frame_type.unwrap_or("<absent>"),
reference = Self::reference_shape(reference).as_str(),
"astrbot media reference"
);
}
_ => {}
}
trace_frame_fields(&value);