On Mon, Aug 07, 2006 at 12:11:17PM -0400, Daniel Veillard wrote: > On Mon, Aug 07, 2006 at 11:55:00AM -0400, Jeremy Katz wrote: > > On Mon, 2006-08-07 at 11:34 -0400, Daniel Veillard wrote: > > > Okay, I nearly did that patch a couple of weeks ago, and then started > > > wondering if the port could not be extracted from some informations provided > > > by xend, but failed to finc anything in the xm --long output or on the > > > xenstore data, and didn't made the change. I really think xend should > > > provide the information, but agreed that's a first step toward sanity at > > > the application level, feel free to commit (unless someone knows how to > > > extract the port from xend !) > > > > It should be in xenstore now[1]... I saw the patch float by the end of > > last week. But we'll probably still want to fall back to this for > > versions of xend that didn't do so. > > > > Jeremy > > > > [1] Looks like under domainpath/console/vnc-port, commit is > > http://xenbits.xensource.com/xen-unstable.hg?cs=155385a02d0b > > Ha ha :-) > > that look way saner to me, except nobody but root can really get the > information (IIRC by default the xenstore _ro socket is not accessible). If the proxy has support for getting the XML it shouldn't matter ? > Now should we still hardcode 5900 + id if not available ? Yeah, I think we need to be able to guarentee that whenever there is a '<graphics type=vnc>' tag, it will always have a port number attribute. If the port number is missing some of the time, people won't be inlined to make use of it & will go back to hardcoding stuff in the application code instead. > The XML dump should go though the proxy too for non-root that's something > I need to add too. Yeah, I just noticed that that's going to be needed too. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|