On Wed, Oct 29, 2008 at 08:48:42PM +0100, Rutger Hofman wrote:
> [I CC: this to the eCos list, because the licensing issue is frequently  
> raised there, too. --R]
>
> Charles Manning wrote:
>> On Sunday 26 October 2008 01:36:26 Emmanuel Blot wrote:
>>>> Sounds great, thanks! (Or if you have no time to clean it up, I'll be
>>>> glad to see it anyway.)
>>> Here it is:
>>> http://moaningmarmot.blogspot.com/2008/10/yaffs2-for-ecos.html
>>
>>
>> Hello all. 
>>
>> I am very pleased to see an eCos port, but I would like to raise the 
>> licensing issue.
>>
>> eCos is released under the eCos license which has some similarities to 
>> GPL, but has an important exception, that you may also link in  
>> proprietary "application code".
> [snip]
>> YAFFS GPL licensing does not provide this exception
>>
>> YAFFS is released under GPL or is available under alternate licensing 
>> from Aleph One.
> [snip]
>> If I have misunderstood something, then please help to clarify.
>
> Of course I can only speak for myself, but I wanted to be license-aware  
> in my approach. My intention is the following:
>
> 1) make a wrapper layer between the eCos file system package and YAFFS
> 2) make a wrapper layer between YAFFS and the eCos NAND Flash package
>
> both based solely on published API information by YAFFS.
>
> As far as I understand, these two wrapper layers would probably not be  
> GPL'd, in the sense that they do not copy any YAFFS code. When I am done  
> writing, these two wrapper layers are fed back to eCos. They can be GPL  
> + eCos exception, if my reasoning above is OK.

Yes.

> If users want to use YAFFS on eCos, they separately download the YAFFS  
> code from Aleph One, and compile the direct/ stuff to be included in  
> their eCos application. It is most certainly not my intention to feed  
> any GPL'd YAFFS code to eCos -- the eCos people won't hear of it, for  
> starters.

Correct.

> If the users' project is GPL, fine. If their project is not GPL, then  
> they will need to approach Aleph One for a commercial licence. In short,  
> they will follow the YAFFS model. (For the record: our project is GPL.)
>
> Is this approach satisfactory?

Yep.

        Andrew