Re: [gtkmm] 2 mainloops in 2 different threads?

[Paul Davis <pbd op net>]

>Thanks for the reply!
>
>Also thanks for the pointer to "Model-View-Controller". I will have a look
>at it and perhaps I will then understand why I should/need not have two
>main loops. Please forgive my ignorance which is due to me not being a
>computer scientist but a material physicist.
>
>I am pretty sure, that I do not actually need two main loops, but I face
>the problem, that the library can be incorporated into a a GUI programm or
>or simple "script like programm", that has no main loop in the first
>thread. I nevertheless need to service the database link. So either the
>programm has two main loops (one doing whatever the user wants it to do
>(Gui, ...) and the second connected to the database) or just one mainloop
>(connected to the database). The library should be the same for both
>szenarios. And thats why I came to this solution. I believe (I do stress the
>word believe - I do not know for sure), that possibly having two mainloops
>is not the worst solution to this problem. I am always willing to have a look
>at other possible solutions - e.g. "Model-View-Controller" and I am very
>grateful, that you have pointed out this concept.

ok, now i understand why you have two main loops. whether you really
need them or not depends mostly on whether there are any real-time
aspects to the database operation. by real-time, i mean "time
constrained". 

if not (and i suspect that this is the case), then you should be able
to just hook the database code into the GUI main loop by providing an
API for the library that:

    1) exposes 1 or more file descriptors that can be polled for
         "read possible" or "write possible"
    2) callbacks to be executed when either of these conditions
         occurs.

--p