Jeremy Evans
Jeremy Evans
Raised `Invalid request parameters` errors using `application/x-www-form-urlencoded` as content type
> This also seems to break mailgun email ingestion with attachments over 4MB. Then mailgun needs to increase the bytesize limit (shown earlier in thread), assuming they want to accept...
Raised `Invalid request parameters` errors using `application/x-www-form-urlencoded` as content type
The middleware isn't thread safe. It's also inefficient (it would be faster to create appropriate query parsers up front, instead of creating one per request). I assume a better approach...
Raised `Invalid request parameters` errors using `application/x-www-form-urlencoded` as content type
> These limits are being triggered for us in unexpected ways. Users attempting to upload files have hit limits on the number of parameters when attempting to upload a file....
Raised `Invalid request parameters` errors using `application/x-www-form-urlencoded` as content type
You can do this using a route-specific `Rack::QueryParser` instance, but you need to override `Rack::Request#query_parser`. Maybe we should support `Rack::Request#query_parser=` to make this easier. What are your thoughts on that,...
Raised `Invalid request parameters` errors using `application/x-www-form-urlencoded` as content type
I submitted #2349 to add `Rack::Request#query_parser=`, which should make it easier to support higher limits on a per-Request basis. Note that `env.request.query_parser = QUERY_PARSER` won't work. `env` is a hash....
I'm guessing this was used so that the body would not be closed until after `body.each` was called (e.g. after the server had sent the empty response body). I'm not...
@ioquatix I think we should revert #2149. If you can come up with an alternative approach that is backwards compatible, we can consider that. Regarding failures showing up as successes,...
I like `called`, because it better describes the method used `call` (`invoke` seems useful for Rake tasks, as that is the method used for those). That being said, phrases such...
I believe it is intentional that Rack no longer URI encodes the `%` in the cookie in Rack 3.1+. @ioquatix is that correct?
I'm not sure it is worth changing this. If you are going to raise anyway (a very rare event that only happens if someone is trying to exploit an issue),...