robustness #271
robustness #271
Conversation
Signed-off-by: pi-pi-miao <[email protected]>
Signed-off-by: pi-pi-miao <[email protected]>
There was a problem hiding this comment.
Hi @pi-pi-miao, the CI should be fixed now. Sorry for the delay!
As for your changes, may I ask what motivated them?
You should never see those unknown statuses from the Proxy-Wasm compliant proxies. See: #282.
As for UTF-8 String failures, those are indeed a problem, but using either _bytes() variants (which were introduced as a solution to that) or a HeaderValue type (#287) is a better solution than returning errors and/or ignoring non-UTF-8 values.
I believe all other changes are a viral effect of propagating errors from the 2 issues above, but please correct me if I've missed something.
Thanks!
| error!("[get_map_value] Failed to convert to UTF-8 string: {}", e); | ||
| Err(Status::SerializationFailure) // Return an appropriate error | ||
| } | ||
| } |
There was a problem hiding this comment.
While this is better than existing behavior, this isn't correct. See #287.
(same for all other instances of failed String::from_utf8() conversions.
| status => { | ||
| error!("[log] unexpected status: {}", status as u32); | ||
| Err(status) | ||
| }, |
There was a problem hiding this comment.
This is done by design, since those errors are illegal are not actionable. See #282.
| }); | ||
|
|
||
| // 将成功转换的键值对推入 map | ||
| map.push((key_string, value_string)); |
There was a problem hiding this comment.
Map is serialized by a trusted host environment, so it's not untrusted/malicious input. See: #282.
No description provided.