Robert Collins wrote: > On Tue, 2003-07-29 at 02:13, Max Bowsher wrote: > >> If we do add a class, it should probably be a thread class from which all of >> our threaded tasks can derive. Regardless, I don't see that any further >> cleanup will reverse these changes. I can see how it may involve changing >> the same lines to a third form, but that is not a reversal. > > Are you thinking of the (IIR the name C) the 'completion object' > pattern? That would work too.
I have no idea what this means. I'm just trying to convince you that applying the current patch as is in no way hinders future development. >> In any case, this is a tiny patch, and even *if* (partial) reversal is >> required, to do so will be trivial. We can benefit *now* from this cleanup, >> whereas classes may be months in the future. >> >> Requesting to commit, > > Denying. A trivial class will take you less time to create than this > series of email has. I'm not asking for the full McCoy, just a step in > the right direction rather than trading one problem for another. I really don't see where you really want to go with this class, so I don't understand what you want. The sole objective of this patch is to eliminate usage of a global variable. Further refactoring may be desirable, but unless I'm thoughroughly missing the point, is totally orthogonal to what I'm trying to do. You realise that *all* I am doing with this patch is to change how data is passed between a function and it's caller? I don't see the application for a class here. Sorry, you are either going to have to explain in more detail what design pattern you want, or let me commit the current trivial patch, or tell me to abandon attempts to remove the global next_dialog for the time being. Max.
