Discovery reads only part of page

[Keith S]

I am testing OpenId implementation with Wordpress. The discovery
phase is failing. The HTML discovery appears to be reading only part
of the page. The log file says it reads 2395 bytes. I have a
separate program that reads that same address, and gets more than 60K
bytes. The HTML that it reads into the log file, and the HTML ends
just BEFORE the ocurance of the openid.server links that tell where
the server is. So, discovery fails.

I can't figure out why it is reading only 2395 bytes. As far as I can
tell, the maxsize setting is still set to 100K -- I have not changed
this. The response comes in less than 4 seconds so it should not be
timing. It simply is not reading enough of the file to do discovery.

We are using version 0.9.2. I know that is not the latest, but I found
incompatibility problems when I tried the latest. Is this a known
problem with that version? Or is there anything specific to www.wordpress.com?