Danek Duvall wrote: > On Fri, Jan 23, 2009 at 01:44:57PM -0600, Kevin Monceaux wrote: > >> Dane, >> >> On Fri, 23 Jan 2009, Danek Duvall wrote: >> >>> I'd start by trussing the client (possibly with "-w all") to see exactly >>> what's getting sent across the wire. There may be something >>> enlightening in that. Or use snoop. You'll probably also want to see >>> what's getting returned, and "-r all" is going to be noisy (though you >>> could run truss once to find out what the socket fd is; it should be >>> predictable). >> I've never used truss before. I've tried the above and the problem might >> be staring me in the face without me knowing it. I've put the output of >> truss -w all and truss -r all runs at: >> >> http://www.RawFedDogs.net/pkgsend-truss-r-all.txt >> >> http://www.RawFedDogs.net/pkgsend-truss-w-all.txt >> >> if anyone would be willing to take a look at. > > Ah. You have an illegal version string: "9.06". We disallow leading zeros > in version components. You'll have to use "9.6". > > Now, why the depot's response is "400 Bad Request" along with an HTML page > that includes a body with > > <h2>400 Bad request</h2> > <p>Illegal FMRI '[email protected]': Bad Version: 9.06</p> > <pre id="traceback"></pre> > > I'm not sure. Perhaps we agreed that would be useful, but it's certainly > not clear to me now that it is. I guess if CherryPy allowed us to set the > response string the way the original server did, we would have seen the > problem much earlier. Though the HTML body still doesn't seem all that > useful.
Whenever you "raise cherrypy.HTTPError" it always outputs an HTTP body. I'm in the process of attempting to make this not suck for users of pkgsend. -- Shawn Walker _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
