Re: [libxml++] 2.8.0 needed

["Murray Cumming" <murrayc murrayc com>]

> Murray Cumming wrote:
>
>>>BTW, I already updated the website with the new versions, the licence
>>>clarification (although I did remove the static linking part because I
>>>have doubts about it, I'll try to get more information).
>>>
>>>
>>
>>What do you doubt?
>>
> Here is what I read in the LGPL text

Only lawyers actually (pretend to) understand licenses, so I'll only talk
about my understanding of the general consensus.

> (version 2.1 shipped in libxml++
> package) about static/dynamic linking. There is only two places the word
> shared or statically is used :
>
> In the preamble :
>
>      73   When a program is linked with a library, whether statically or
> using
>      74 a shared library, the combination of the two is legally speaking a
>      75 combined work, a derivative of the original library.  The ordinary
>      76 General Public License therefore permits such linking only if the
>      77 entire combination fits its criteria of freedom.  The Lesser
> General
>      78 Public License permits more lax criteria for linking other code
> with
>      79 the library.
>
>  From this I understand that there is no big difference between linking
> statically of dynamically.

No, I don't think that section says any such thing. It just says that the
LGPL allows something that the GPL does not. It does not yet say what it
allows.

> Then in Section 6:
>
>     278   You must give prominent notice with each copy of the work that
> the
>     279 Library is used in it and that the Library and its use are
> covered by
>     280 this License.  You must supply a copy of this License.  If the
> work
>     281 during execution displays copyright notices, you must include the
>     282 copyright notice for the Library among them, as well as a
> reference
>     283 directing the user to the copy of this License.  Also, you must
> do *one*
>     284 of these things:
>     285
>     286     a) Accompany the work with the complete corresponding
>     287     machine-readable source code for the Library including
> whatever
>     288     changes were used in the work (which must be distributed under
> [...]
>     297
>     298     b) Use a suitable shared library mechanism for linking with
> the
>     299     Library.  A suitable mechanism is one that (1) uses at run
> time a
>     300     copy of the library already present on the user's computer
> system,
>     301     rather than copying library functions into the executable,
> and (2)
>     302     will operate properly with a modified version of the library,
> if
>     303     the user installs one, as long as the modified version is
>     304     interface-compatible with the version that the work was made
> with.
>     305
> [then points c, d and e]
>
>
> What I understand from this is that the use of a shared linking is one
> of the options proprietary projects have, but it's not an obligation if
> one of the other conditions is respected (ie a, c, d or e).

Yes, and all those other options involve providing the source code to
people who have received the object code.

> None of the words 'shared', 'static' ant 'dynamic' can be found anywhere
> else in the License.

I don't think they need to be mentioned. It clearly says that one of your
options is to use a shared library. A shared library is a
dynamically-linked library. Statically linked libraries are not shared
libraries.

> This is why I have doubts. If you think I misunderstood the license, I
> welcome any clear explanation. Although the simpliest way to close the
> discussion is to ask the question to the fsf.

Of course. However, note that the FSF will try to push you towards the use
of the GPL. They don't like the LGPL much because it does not force
software to be proprietary. When asking, try to ask about a specific
situation.

Actually, it probably is worth asking, because I found this:
http://www.openoffice.org/FAQs/faq-licensing.html#13
although I think that still means that you need to provide source code
(via debug symbols).

Murray Cumming
murrayc murrayc com
www.murrayc.com
www.openismus.com