on Mon Jun 02 2008, "troy d. straszheim" <troy-rPoGFrA5WPqukZHgTAicrQ-AT-public.gmane.org> wrote:
> David Abrahams wrote: > [snip] >> >> Yes, that's what bjam is doing today. It uses popen to invoke all the >> commands and capture their output. To do the same with CMake you may >> need to request/implement some patches (?) >> > > Turns it out wasn't necessary. :) > >>>> Maybe we should pursue both tracks in parallel until we discover which >>>> one will be easiest? >>> Let's me get this ctest-rfc out and the traash demo up, let's discuss that, >>> then decide. > > If the new xml-generation stuff in cmake looks good to people (comments? > Is it workable on windows?), then things have changed here a bit. > We control both ends of the protocol, and both ends are python. > Python has a built in xmlrpc client and trac has an xmlrpc plugin > for the server side: > > http://trac-hacks.org/wiki/XmlRpcPlugin Yep, I know about that thing. > which could *vastly* simplify the code. One wouldn't even have to touch XML. > On the client, you just marshal python datastructures to a log. At POST time, > you demarshal them, send them through an xmlrpc call, and they appear, > unpacked, > in the arguments to a function call inside your trac plugin. Voila, > bye-bye tangly dart-log-parsing code. Going to play with this this evening. Very nice! -- Dave Abrahams BoostPro Computing http://www.boostpro.com _______________________________________________ Boost-cmake mailing list Boost-cmake@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-cmake
