Comments on: 11. The End http://tromey.com/blog/?p=552&utm_source=rss&utm_medium=rss&utm_campaign=11-the-end Wed, 15 May 2013 18:03:31 +0000 hourly 1 http://wordpress.org/?v=3.5.1 By: Mark Seaborn http://tromey.com/blog/?p=552&cpage=1#comment-156460 Mark Seaborn Tue, 16 Dec 2008 12:40:01 +0000 http://tromey.com/blog/?p=552#comment-156460 If you want to follow RPC calls, you might want to look at the Causeway debugger for ideas:
http://www.erights.org/elang/tools/causeway/index.html
http://www.erights.org/elang/tools/causeway/causeway-paper.pdf
“Causeway, an open source distributed debugger written in E, lets you browse the causal graph of events in a distributed computation.”

]]> By: tom http://tromey.com/blog/?p=552&cpage=1#comment-156446 tom Fri, 12 Dec 2008 23:55:44 +0000 http://tromey.com/blog/?p=552#comment-156446 Nice idea.

I think there are some subtleties… RPC means that a simple backtrace may cross multiple processes. So, you definitely would want a way to parameterize this — a way to ask for the “logical” trace or the “real” one. Perhaps we can add options to our existing filtering backtrace to pass down to the filters…

What is funny about this idea is that, aside from figuring out how to associate the caller and callee, it might not actually be very hard.

]]> By: Dave Malcolm http://tromey.com/blog/?p=552&cpage=1#comment-156445 Dave Malcolm Fri, 12 Dec 2008 21:34:22 +0000 http://tromey.com/blog/?p=552#comment-156445 One idea that just occurred re multi-process support: follow an RPC call. When I was working on evolution, it would have been very useful to trap a Bonobo call, so you step from the caller through to the ORB process and then into the callee process. Could do similar for DBus calls, or for reading from pipes, etc.

Would needs plugins for each type of IPC, I guess, but I suspect the python framework could deal with that. Not that I’m volunteering to implement it :-)

]]>