Re: [Gnu-arch-users] Nit

[Joshua Haberman]

On Sun, 2003-10-19 at 13:07, Tom Lord wrote:
>     > From: Joshua Haberman <address@hidden>
> 
>     > On Sun, 2003-10-19 at 11:27, Charles Duffy wrote:
>     > > On Sun, 2003-10-19 at 13:11, Joshua Haberman wrote:
>     > > > 2) deserializers have to be written for every target language
> 
>     > > Why's that?
> 
>     > > Why couldn't someone interested in having a library interface (*ahem*)
>     > > write just *one* deserializer exposed as a C library which could then 
> be
>     > > used by any language-specific interface?
> 
>     > Funny, this is exactly the argument for making libarch wrappable.  The
>     > answer I've gotten is equivalent to:
> 
>     > "Because then the deserializer will take up memory and bugs in it could
>     > make the language-specific interface crash.  The deserializer needs to
>     > be in its own process."
> 
> I think you're just confused.   The deserializer cduffy is talking
> about goes in the same process as the higher level application.   The
> serializer is in a separate process.

I'm pretty sure I understand.  I still think it's a close analogy.  

cduffy is proposing wrapping a shared implementation of tla
deserializing (call it "libtladeserialize") with bindings to different
languages, putting libtladeserialize in the same process as the
application.

This is very similar to the proposal of wrapping a shared implementation
of manipulating arch archives/trees (call it "libarch") with bindings to
different languages, putting libarch in the same process as the
application.

The implementation of libtladeserialize is less complex than the
implementation of libarch, which may be the differentiating factor for
you, but otherwise arguments to keep libarch in a separate process from
high-level applications would also apply to libtladeserialize.

Josh