On Nov 21 2012 3:05 PM, John Morris wrote:
> Hi EBo,
>
> On 11/21/2012 01:51 PM, EBo wrote:
>> True.  But one point that is in the compatibility list that you did 
>> not
>> cover is:
>>
>> if LCNCv3 is (L)GPL2+, then it conveys with anything that is 
>> (L)GPL3+
>
> Thanks for pointing this out.  I'm weak on this subject, only knowing
> what I do from packaging software for Fedora and RPMFusion, and not
> familiar with the 'conveys with' terminology.  My understanding is 
> that
> if LCNC (or any software) is GPLv2+ and it's linked against something
> GPLv3, it must then follow the GPLv3 rules.  That will then make
> anything GPLv2-only license-incompatible.  Did I get that right?

That jives with my understaning.

> If I did, then at least the portion of LCNC that links against the 
> RTAI
> and Xenomai kernels must follow the rules for GPLv2-only, and must 
> not
> link against anything GPLv3 or otherwise GPLv2 incompatible.

I'm not entirely clear what this would mean for us with using/linking 
OS level stuff with our code.  For that I would ask EFF, but there is 
probably something out there that explains it.

>> How much of readline do you really use?  Can it be easily replaced 
>> with
>> something else.
>
> I don't know what readline is used for in LCNC, but I imagine 
> readline
> v6 would be more or less easily replaced by readline v5.
>
> If the RTAPI/HAL portion and the UI portion only communicate through
> IPC, then there would be more flexibility on UI licensing.

probably, but IANAL...


>       John
>
>
>> What I am getting at is to start with *if* all developers are on 
>> board
>> with (L)GPL2+, then we look at the requirements.  Them make a list 
>> of
>> the potential libraries that can be broken out (HAL, ClassicLadder, 
>> ???)
>> with a complete list of each ones dependencies (like readline, 
>> boost,
>> ...)
>>
>> Once the above is done, we can either start experimenting with
>> breakouts, or finding replacements for functional components.
>>
>> Once we have the above, we can then do the same for the different 
>> UI's.
>>
>> That's my 2c...  I'll look at this again later tonight.
>>
>>     EBo --
>>
>>
>> On Nov 21 2012 12:30 PM, John Morris wrote:
>>> Hi list,
>>>
>>> On 11/21/2012 07:11 AM, Michael Haberler wrote:
>>>> Yes it is time, but we have not ticked off on a key sanitary
>>>> non-code
>>>> prerequisites
>>>>
>>>> The GPL2-only situation prevents bringing in outside code which
>>>> requires GPL2+ licenses.
>>>>
>>>> I table this issue again to get this finally resolved, I do not 
>>>> want
>>>> to 'slide into some L3 work' and that glaring issue somehow falls
>>>> under the table because of all the excitement about features and
>>>> grand ideas about what could be done. We had a push recently, and 
>>>> it
>>>> tapered off without tangible results.
>>>>
>>>> This licensing issue *must* be resolved before we go ahead, there 
>>>> is
>>>> no way around it.
>>>
>>> Licensing is a complex issue, not simply about the original 
>>> authors'
>>> wishes, but also what external software LCNC is built against, how
>>> LCNC
>>> is built against them (linked, code copied, or not), and how those
>>> external softwares are licensed.  Including those external 
>>> softwares
>>> introduces license compatibility constraints.
>>>
>>> An example of such a constraint:  On my system, LCNC is linked
>>> against
>>> the GNU readline library v6, licensed GPLv3+.  Therefore LCNC must
>>> either also be licensed GPLv3, or else must drop that library.  If
>>> the
>>> readline v6 library were dropped, LCNC could be licensed as GPLv2,
>>> according to my incomplete survey below.  One possible workaround
>>> would
>>> be to use the GPLv2+-licensed readline v5, available on fedora as
>>> 'compat-readline5'.  (Anyone know what readline library used in
>>> Debian?)
>>>
>>> An example of a non-constraint:  LCNC is compiled with gcc and uses
>>> source-highlight to process documentation.  Those tools are 
>>> licensed
>>> GPLv3, but since LCNC does not link to them or copy code from them,
>>> LCNC
>>> licensing isn't affected.
>>>
>>> An example of a problematic constraint:  The RTAI and 
>>> Xenomai-kernel
>>> versions link against both the GPLv3+ readline v6 library (on my
>>> system,
>>> anyway) and the GPLv2-only Linux kernel.  LCNC cannot legally be
>>> shipped
>>> like that.  Unless...
>>>
>>> I don't understand the architecture of LCNC well, but for example 
>>> if
>>> the
>>> RT piece linked to the kernel talks to the UI piece linked to
>>> readline
>>> only through shm, there may be an opportunity to ship the two 
>>> pieces
>>> separately under separate licenses.
>>>
>>> Here's an incomplete survey based on the requirements discovered
>>> while
>>> building the el6/fedora packages.  I don't know how some of the
>>> programs
>>> are combined with LCNC, and I don't know about compatibility of all
>>> of
>>> the licenses; hopefully others can help with that.  Otherwise, a
>>> quick
>>> compatibility chart for the GNU licenses can be found here:
>>>
>>> http://www.gnu.org/licenses/gpl-faq.html#AllCompatibility
>>>
>>>     John
>>>
>>>
>>> Linked:
>>> LGPLv2+:  gtk2-devel, libgnomeprintui22-devel
>>> MIT:  mesa-libGL-devel, mesa-libGLU-devel, libXaw-devel
>>> Boost:  boost-devel
>>> LGPLv2+:  pth-devel, libmodbus-devel
>>> GPLv3+:  readline-devel
>>> GPLv2:   Linux kernel (Xenomai-kernel, RTAI, RTLinux)
>>>
>>> Dunno:
>>> TCL:  tcl-devel, tk-devel, bwidget
>>> LGPLv3+:  python-mtTkinter
>>> MIT:  blt-devel
>>> GPLv3+ and LGPLv2+:  gettext
>>> python:  python-devel, python-lxml
>>>
>>> Not linked:
>>> GPLv3+:  gcc, gcc-c++
>>> GPLv2+:  lyx, dblatex
>>> GPLv3+:  source-highlight
>>> ImageMagick:  ImageMagick
>>> GPLv2+ and OFSFDL:  dvipng
>>> GPL+ and GPLv2+:  asciidoc >= 8.5
>>> GPLv2:   Linux kernel (Xenomai-user, RT_PREEMPT, Posix)
>>>
>>>
>>>
>>> 
>>> ------------------------------------------------------------------------------
>>> Monitor your physical, virtual and cloud infrastructure from a 
>>> single
>>> web console. Get in-depth insight into apps, servers, databases,
>>> vmware,
>>> SAP, cloud infrastructure, etc. Download 30-day Free Trial.
>>> Pricing starts from $795 for 25 servers or applications!
>>> http://p.sf.net/sfu/zoho_dev2dev_nov
>>> _______________________________________________
>>> Emc-developers mailing list
>>> Emc-developers@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/emc-developers
>>
>>
>> 
>> ------------------------------------------------------------------------------
>> Monitor your physical, virtual and cloud infrastructure from a 
>> single
>> web console. Get in-depth insight into apps, servers, databases, 
>> vmware,
>> SAP, cloud infrastructure, etc. Download 30-day Free Trial.
>> Pricing starts from $795 for 25 servers or applications!
>> http://p.sf.net/sfu/zoho_dev2dev_nov
>> _______________________________________________
>> Emc-developers mailing list
>> Emc-developers@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/emc-developers
>>
>
>
> 
> ------------------------------------------------------------------------------
> Monitor your physical, virtual and cloud infrastructure from a single
> web console. Get in-depth insight into apps, servers, databases, 
> vmware,
> SAP, cloud infrastructure, etc. Download 30-day Free Trial.
> Pricing starts from $795 for 25 servers or applications!
> http://p.sf.net/sfu/zoho_dev2dev_nov
> _______________________________________________
> Emc-developers mailing list
> Emc-developers@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-developers


------------------------------------------------------------------------------
Monitor your physical, virtual and cloud infrastructure from a single
web console. Get in-depth insight into apps, servers, databases, vmware,
SAP, cloud infrastructure, etc. Download 30-day Free Trial.
Pricing starts from $795 for 25 servers or applications!
http://p.sf.net/sfu/zoho_dev2dev_nov
_______________________________________________
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to