Got a tricky one here.
I have written a SOAP::Lite service which happens to be a consumer of
a SOAP service hosted in Weblogic. The WLI service connects to a third
party. The problem is the 3rd party returns unicode characters, and
the WLI service calculates the content-length of the message in the
response before converting to UTF-8. Thus, by the time I get the
response, the content-length reported in the message is actually 1 or
more byts smaller than the actual content-length.
Is there a setting whereby I can make my service not sensitive to
content-length? Or is this really an issue with expat? I am working on
a new Solaris 2.8 box, my old 2.8 box this works fine, the only
difference is that we are using sun's crappy c compiler instead of gcc
on the new box. BTW I am still on SOAP::Lite 0.60.
Thanks for any help,