On Thu, 29 Jun 2006 17:39:20 -0400, [EMAIL PROTECTED] said:
> Great work Ben!

Thanks :)

> We may run into trouble with LL by automatically accepting the TOS.
> In the future someone may want to think about adding a callback that
> returns any new TOS and waits for a function to be called that either
> accepts or rejects the new TOS.  This would allow someone to display
> the new TOS in their own GUI and offer it up for acceptance.  It would
> keep LL's lawyers off libsecondlife's backs.

A couple things to mention that I left out in my haste:

* Re the gzip issue -- the reason why this wasn't seen before was that
presumably the server only bothers to compress its response if its above
a certain size -- Linden's TOS is almost 50k (7,827 words)!  Presumably,
the XmlRpc code won't set that accept=gzip flag, so this won't be a
problem.

* When it displays the TOS, it does so by returning (in XmlRpc format)
reason=tos, message=[big huge honkin' TOS], login=false.
  Presumably, the client will note that, and then ask for agreement, and
  then resend the login with agree_to_tos=true.  The server must then
  maintain that state, because the client only sends that once. 

With that in mind, the following patch is probably sufficient to log in.
 It's not technically correct to ALWAYS specify "agree_to_tos", even
when you haven't been asked to, but oh well.

--- NetworkManager.cs   (revision 38)
+++ NetworkManager.cs   (working copy)
@@ -673,6 +673,7 @@
                        values["build"] = build;
                        values["platform"] = platform;
                        values["mac"] = mac;
+                       values["agree_to_tos"] = "true";
                        values["viewer_digest"] = viewerDigest;
                        values["user-agent"] = userAgent + " (" +
                        Helpers.VERSION + ")";
                        values["author"] = author;

-b

_______________________________________________
libsecondlife-dev mailing list
[email protected]
https://mail.gna.org/listinfo/libsecondlife-dev

Reply via email to