[prev in list] [next in list] [prev in thread] [next in thread] 

List:       freedesktop-dbus
Subject:    Re: IDL language
From:       David Zeuthen <david () fubar ! dk>
Date:       2009-05-12 13:19:03
Message-ID: 1242134343.11439.7.camel () localhost ! localdomain
[Download RAW message or body]

On Tue, 2009-05-12 at 14:08 +0100, Simon McVittie wrote:
> At the risk of repeating myself: you are designing the IDL, now, so what is
> valid or invalid IDL is your choice. If you want the definition of the IDL to
> include "problematic names MUST have a NamingHint" (which I would advocate)
> then now is the time to say that (I assume there's a description of the IDL
> other than "what this tool parses"?), 

Sure, as I've already said, I think that's sensible to have on the IDL
level. And I would even propose that we should enforce it at the typelib
level (e.g. XML), e.g. a typelib is not valid if problematic names does
not having a naming hint - but on the typelib level you may find people
disagreeing with you....

Anyway, all this is a bit hypothetical until we actually get to a point
where there's working code spitting out typelibs from some kind of IDL -
which is something I'm planning to finish working on in a few weeks
(unfortunately the real world caught up with me and I have to do other
work this and next week). Then we can have a long discussion about the
annotations we want to add, what makes up a typelib; and all that good
stuff. And then someone can write a dbus-typelib-validate(1) program and
so on.

> and it'll be other-tool users' problem if 
> their other-tool generates (something that you'll define as) invalid IDL.

I don't expect people will writing programs spitting out IDL; it's more
likely they'll write programs spitting out typelibs.

     David


_______________________________________________
dbus mailing list
dbus@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dbus
[prev in list] [next in list] [prev in thread] [next in thread]