-frepo, search-mailing-list, precomp-hdrs, native-objects

[Mike Stump]

> Date: Fri, 12 Sep 1997 16:06:46 -0700
> From: Yotam Medini <yotamm@tmai.com>
> To: egcs@cygnus.com

> * I may have missed important announcements...
>   Could someone please clarify the current status of egcs
>   regarding template repository (-frepo ?),

Should work, report any bugs you find.

> future plans,

Talking over and over about future plans is not too useful.  Look at
what has been done in the past year, and you'll get an idea of
everything that will probably be worked on in the next year.  Though,
we get nice surprises from time to time.  If you read all the messages
in the archive, you can get a feel for what we will be working on.

> documentation.

Probably not much planned in this area.  Would you like to write some?

> Some suggestions an thoughts:

> * Let us have a web-<Search> function with the archives of the mailing lists.

Download the archive, and build a full text index...  :-)   A web search
button would be useful, though.

> * How about reviving the old idea of pre-compiled headers.

Ok, review the competition, lay out your plans for implementing them,
and we'll comment on them, after that, you can then implement them,
and we'll put them in.  Sound fair?  If this doesn't suit you, then
you will have to recruit someone to do them yourself.  If you fail to
do this, then the compiler probably won't do them.

> * The pain of C++ linkage.
>   Currently it is impossible to link C++ objects that are compiled
>   with different compilers. Reasons being: 
>      * name-mangling schemes
>      * virtual table implementations
>      * static initialization methods.
>      * ??
>   
>   Now, on platforms where the native compilers (Solaris' SunPro CC?)
>   has a published object implementation, is it possible for g++ to have
>   an option to compile an object that will be linkable with 'native objects'?

You must not have a in depth understanding of what it takes for the
ABI to be useful.  It is hard.  The Sun spec is incomplete.  It
doesn't define the entire repo mechanism for example.  It doesn't
mandate header file contents well enough to be useful.