[RFC] speech dispatcher client/server data types

[William Hubbs]

Hello Hynek,

can you please take a look at this thread and provide comments?

Andrei is proposing that we put all of the typedefs used by the server
and API into spd_types.h, remove them from libspeechd.h, and have
libspeechd.h #include spd_types.h.

I can understand Andrei's argument for doing this, which is consistency,
and no duplication.

I'm fine with the first part of this proposal, putting the types from
libspeechd.h into spd_types.h and reworking the server to use those
types.

The second and third parts, however, would result in spd_types.h being
shared between the server and the c API.  This is the part of the
proposal I am concerned about, and this is why.

I remember a statement you made a while back saying that you believe the
API's should do their own thing, separately, based on the documentation,
and I believe having a shared header file goes against your design.

Also, I am concerned about having a shared header file, because, any
time you change this header file you have to bump the revision number of
the c API.  There could be situations in which there would have to be
unnecessary revision bumps.

What do you think?  are you ok with a shared header file?

Thanks,

William

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: 
<http://lists.freebsoft.org/pipermail/speechd/attachments/20101021/1f43559b/attachment.pgp>