244717 - SOAP requests should state the charset of their payloads

Status: RESOLVED FIXED

(Core Graveyard :: Web Services, defect)

Description

[Cédric Chantepie]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6) Gecko/20040113
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6) Gecko/20040113

It's seems I've some encoding trouble with webservice. In the js I've a string
with value "Nouveau répertoire", this string is passed to a webservice and in
the webservice I get the string value "Nouveau répertoire". In the js the
string value is get from a message bundle encoded in UTF-8, and the SOAP
response is too in UTF-8 (Apache SOAP)

Reproducible: Always
Steps to Reproduce:
1. Make string with accent in JS, display it for debug
2. Send the string param to a webservice operation
3. Display received string param in webservice

Actual Results:  
JS displays "Nouveau répertoire"
Webservice displays "Nouveau répertoire"

Expected Results:  
Webservice would display the unchanged string "Nouveau répertoire"

Comment 1

[Boris Zbarsky [:bzbarsky]]

Um... that string your webservice displays is precisely what you get if you take
the UTF-8 encoding of your original string and decode it as ISO-8859-1.  So the
webservice is somehow screwing up its character encodings, sounds like.

Comment 2

[Harshal Pradhan]

Attachment: Possible fix

The idea is to explicitly state that we are sending utf8. I use this since it
fixed some similar issues with .Net webservices for me. I have no idea if it
will help Apache SOAP services.

Comment 3

[Harshal Pradhan]

Cédric,it would be great if you could build with the above patch and test and
let me know if it helps. If it doesn't, then we probably need a publicly
accessible testcase or something to help us figure out what is going on. 

Comment 4

[Harshal Pradhan]

I have no way of knowing for certain what the problem with Apache SOAP is, but I
can confirm that something similar happens with .Net webservices. Confirming and
changing summary to the known part of the problem. Lets fix that first and then
see if anything else remains to be done.

Comment 5

[Harshal Pradhan]

Attachment: Updated fix

In principle this is the same as attachment 149370 [details] [diff] [review]. But I couldn't resist some
cleanup and factoring out of duplicated code.

Comment 6

[Harshal Pradhan]

Attachment: diff -w version of the above

Comment 7

[Harshal Pradhan]

Comment on attachment 149640 [details] [diff] [review]
diff -w version of the above

jst, would you please review.

The basic idea is to send the charset with the contenttype.

This fixes some issues I have seen with sending strings to .Net webservices.
And it also is in line with what a lot of SOAP clients seem to do (based on
wire dumps I could find amongst the  SOAPbuilders interop results at
http://www.whitemesa.com/interop.htm ).

The rest of the patch is just refactoring some duplicated code.

Comment 8

[Johnny Stenback (:jst)]

Comment on attachment 149640 [details] [diff] [review]
diff -w version of the above

Looks good, r+sr=jst

Comment 9

[Johnny Stenback (:jst)]

Harshal, would you be able to attach a patch w/o the cleanup of the duplicated
code too? We could try to get that in for 1.7 since it seems like this would be
a big step forward for our webservices support.

Comment 10

[Harshal Pradhan]

Patch checked in on trunk. As far as branch goes, I guess my earlier version
should do.

Comment 11

[Harshal Pradhan]

Comment on attachment 149370 [details] [diff] [review]
Possible fix

requesting review for branch version.

Comment 12

[Johnny Stenback (:jst)]

Comment on attachment 149370 [details] [diff] [review]
Possible fix

Yeah, this would do it for the branch. r+sr=jst

Comment 13

[Cédric Chantepie]

Attachment: testcase

Patch seems to work according this testcase.
Read first the 'README' file in the tarball plz.

Comment 14

[Asa Dotzler [:asa]]

Comment on attachment 149370 [details] [diff] [review]
Possible fix

a=asa (on behalf of drivers) for checkin to Mozilla 1.7. We're wrapping things
up and so this will need to land today if it's going to make the release.

Comment 15

[Johnny Stenback (:jst)]

Attachment 149370 [details] [diff] checked in on the 1.7 branch.

Comment 16

[Harshal Pradhan]

Thanks jst.