I like state machine usually, but the number and complexity of them in nanomsg was something I found to be too high to my taste. When I approached writing mangos, I took a different approach. There’s not much explicit FSM stuff in mangos, as a result, and the code is pretty straight-forward. - Garrett > On Oct 31, 2014, at 12:42 AM, Paul Colomiets <paul@xxxxxxxxxxxxxx> wrote: > > Hi Martin > > On Fri, Oct 31, 2014 at 6:33 AM, Martin Sustrik <sustrik@xxxxxxxxxx> wrote: >> One feature that I would like to specifically point out is that >> termination part of the state machine (which tends to be the most >> messy part) is generated automatically with little or no help from the >> programmer. >> > > The "generated automatically" thing sounds bold. Do you mean the > "cancel" clause helps? Or there are things, that I've not found by > skimming ocumentation? > >> Should we use that to generate nanomsg state machines? Thoughts? >> Ideas? Objections? >> > > Given that there are state machine bugs, that nobody fixed for more > than year, I'd say yes it would be nice. The downside is that it would > be almost complete rewrite of nanomsg. > > > > -- > Paul >