Encoding issue with Connect-HTTP
We are posting a message body to an external system through Connect-HTTP and the body contains Latin letters like öä.
Seems to be those special letters are encoded with cp1252 which is the default encoding (for example:- Löän is encoded as L[0xf6][0xe4]n)
> Is it possible to encode them using UTF-8 without updating the default encoding at the JVM level? ( like Löän gets encoded as L[0xc3][0xb6][0xc3][0xa4]n )
if we pass the same request(contains Latin letters) through connect-soap then it gets encoded in UTF-8 as expected even though the default encoding in JVM is cp1252.
Also having plenty of SOAP & REST calls in the system, no charset issue with them as a message is being encoded in UTF-8 as expected but only issue with CONNECT-HTTP call.
I would like to understand how encoding happens when connecting to an external system despite connect-REST/connect-soap/connect-http.
Cognitonic Systems Australia
AU
@BRAHMESH@ Following links might be useful:
https://collaborate.pega.com/question/connect-rest-encoding-special-characters-utf-8-not-working
Updated: 9 Nov 2021 5:18 EST
Skandiabanken
SE
@KotaC382 Thanks for the response. I had gone through those articles before.
I'm looking at how encoding happens for connectors. Does it always consider the JVM level encoding?
Cognitonic Systems Australia
AU
@BRAHMESH@Usually encoding wont be connector specific, whenever an Integration calls happens it will consider the encoding that is set at the application server level.
Updated: 15 Nov 2021 14:42 EST
Skandiabanken
SE
@KotaC382 Thanks for the reply . Where can we see the encoding configuration at the application server level? , Is it a JVM argument?
Also, Let's assume if Pega is communicating two different systems, one target system is expecting Window-1252 encoding & other one is expecting UTF-8. How do we handle this situation ?.
By default, the value of file encoding in Pega is Cp-1252. so if we pass the charset in the content-type header like charset=Cp-1252 then can app server of target system read the charset from the content-type header for decoding the message? , As per my observation, this is not happening.
Please correct me if I misunderstood this..
JP Morgan
US
@BRAHMESH@ it can be done at server level by passing jvm args
Can you try this as the jvm args . Also look for encoding.properties file if any on the server.
-Dclient.encoding.override=UTF-8