elementary-dev-community team mailing list archive

[Cameron Norman <camerontnorman@xxxxxxxxx>]

I guess what I do not understand is how the application gets the data 
by executing the command. What I see in that implementation is simply a 
return of a command to run with an existing file as an argument, 
instead of a way to retrieve a new file.

The difference is that the former says "I have a <type>, what can I do 
with it?" while the latter says "I want to do something with a <type>, 
how/where can I get one?".

El Tue, 18 de Mar 2014 a las 8:45 PM, Akshay Shekher 
<voldyman666@xxxxxxxxx> escribió:
> The earlier implementation IIRC worked in the following way.
>
> App -> contractor [give me a list if programs that handle file type x]
>
> <- [program a b ... ..]
>
> App asks the user to select one or selects one itself and
>
> App -> Contractor [program Id x, for file/uri y]
>
> <-  [command string]
>
> App executes this using execv or something similar.
>
> The '->' is for dbus function calls and '<-' for return.
> Dbus is just a protocol for doing interprocess communication so a 
> program/script can use contractor if it can use dbus.