On Fri, 9 Nov 2001, Sven wrote:

> On Thu, Nov 08, 2001 at 11:30:56PM -0500, Patrick M Doane wrote:
>
> No, the worse that happens is that you must distribute the .cmo, .cmi, .cmx,
> and possibly .o, hopefully with a working makefile, but this last one is not
> demanded.

As mentioned before - I don't think this is true. The LGPL is very clear
that the user must be able to modify the work (i.e. the application and
not the library). This is not possible to do with .cm[iox] files.

Also, I still must permit users to reverse engineer my application.

It is standard practice to strip an executable of all symbols to prevent
users from snooping around in the code.  Even if all I had to do was
include object files, the names of identifiers would still be intact.

> The reason for that, in the C context (you need the .o only), is so if you
> link with a buggy version of the LGPLed library, your client, or whatever, can
> correct the LGPLed library, or grab a fixed version from the net, and rebuild
> your app without the bug.

This is a good goal - but it's not what the LPGL says.

> Is this still unaceptable, or do you think this clarification will be ok with
> you ?

It may be acceptable to me, but probably not for my employeer or clients
who work in vertical markets that are very sensitive about proprietary
information.

Patrick

-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr