celix-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Pepijn Noltes <pepijnnol...@gmail.com>
Subject Re: Working towards a next release
Date Sun, 18 Nov 2012 06:56:40 GMT
On Sat, Nov 17, 2012 at 9:17 PM, Alexander Broekhuis

> > I agree with all the topics mentioned. Personally I think we can gain
> much
> > by improving remote services and updating the documentation so that it is
> > easier to get started with Apache Celix. There is nice page for a hello
> > world example in Apache Celix, but a page how get a CMake build
> environment
> > working separated from Apache Celix is missing.
> >
> Good point, I agree. Is it a good idea to have some sort of "building
> framework" which can be used?

Yeah, I think that would be great. It should be to hard, we can reuse most
of the CMake files.

> >
> > I also think is a good idea to get the naming and typedefs straight. For
> > example we now have:
> > typedef struct listener_hook *listener_hook_t;
> > and
> > typedef struct bundle * BUNDLE;
> Agree with this one as well, I prefer the lower_case_t version, and have
> been using that one for newer code as well. What do you think?

I also prefer the lower_case_t version, mainly because its closer the the
APR convention.

That being said, I'm curious if there is a reason why we don't complete
follow APR example (so instead of typedef with a pointer to a struct, just
a typedef of a struct and the output argument as first argument instead of
Apache Celix is build on top of APR and uses a lot of the same conventions
which are used in APR (e.g. apr_status_t <-> celix_status_t), but with
some exceptions. IMO its worth discussing if this is "right" way to do this
and if its not, if its worth the time to correct this and if we want to
change it if this should be in the next release.

> --
> Met vriendelijke groet,
> Alexander Broekhuis

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message