CODEWITHSUNDEEP
asp.netweb-servicesencodingcharacter-encodingasmx
pettys
pettys

Reputation: 2468

ASMX Web Service using wrong encoding on incoming request

My .NET ASMX webservice is accepting requests from a client I don't have direct control over. It's sending a request that looks like this:

POST /Service.asmx HTTP/1.1
Connection: Keep-Alive
Pragma: no-cache
Content-Length: 1382
Content-Type: text/xml
Accept: text/xml
Host: localhost
User-Agent: Borland SOAP 1.1
SOAPAction: "http://domain.com/InsertRecords"

<?xml version="1.0"?>
<SOAP-ENV:Envelope... <v>ÄLMÅ BÄCK</v></SOAP-ENV:Envelope>

In my WebMethod, the string ÄLMÅ BÄCK gets munged to ??LM?? B??CK -- typical encoding mess-up.

In my testing I've found that if I simply tweak the content-type header, all is well:

Content-Type: text/xml; charset=utf-8

Why is .NET choosing an encoding other than utf-8 when it's unspecified, and is there any way I can coerce this ASMX to use UTF-8 encoding?

Upvotes: 6

Views: 7133

Answers (2)

JeffK
JeffK

Reputation: 3039

IIS 7.5 has configuration options to help with this. (I don't know if earlier versions support this.) I had a similar issue where I have a web application that receives requests from a system that uses the upper 5 characters of the extended ASCII character set as meaningful delimiters. These were getting munged by the decoding that IIS was applying to incoming requests. I found some IIS config options to fix this issue.

First, in IIS Manager, select your website and open the .NET Globalization settings:

IIS .NET Globalization Settings

There are settings for the expected encodings for File, Requests, Response Headers, and Responses. There are dozens of encoding options to choose from:

Setting encodings in IIS

This worked perfectly for my scenario because this particular site only received requests from a program that I wrote, so I controlled both ends. With an unknown audience, you're just hoping your users encode their requests appropriately. (But then, that's true even if you use the default encodings...)

Upvotes: 1

pettys
pettys

Reputation: 2468

The following code run before the web service handler is invoked resulted in the HTTP Request correctly decoded:

if (HttpContext.Current.Request.ContentType == "text/xml") {
    HttpContext.Current.Request.ContentType = "text/xml; charset=UTF-8";
}

This feels a bit hacky, but I believe it'll work well for my circumstances. I'm still very interested in some background information on why this was an issue at all, and if there's a better way to pull this off (besides getting the client to be more explicit about the encoding).

Upvotes: 4

Related Questions