On 07/06/2011 07:10 PM, Chris Lattner wrote:> On Jul 6, 2011, at 1:55 AM, Tor Gunnar Houeland wrote: > >>> There is no need to include any notices in the binaries of an application built with clang, or some with some other application that links to the LLVM runtime libraries that are dual licensed. >> Thanks for your response. Is this ability to distribute binaries without notices based on your personal assertion that the MIT license used does not require them? > Yes. I am not a lawyer and am not giving legal advice. This is based on my understanding of the MIT license.OK, thanks. That's contradictory to the actual conditions written in the license, and as such the dual-licensing does not serve that purpose. Could http://llvm.org/docs/DeveloperPolicy.html#license be updated to reflect that? (Or do you strongly disagree that "all copies of the software" includes "binary"? When changing the license, you mentioned an example of Mozilla building with Clang. As can be seen in about:license, they consider MIT licenses to require reproduction of the license text. What's your reasoning for your understanding?) I really agree with the intentions, it would be nice for user-compiled programs not to unwittingly include code that imposes licensing conditions. It would be great if it were feasible to also provide such a different license, or additional permissions similar to what GNU libstdc++ does. - Tor Gunnar
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 06/07/11 20:47, Tor Gunnar Houeland wrote: [...]> I really agree with the intentions, it would be nice for user-compiled > programs not to unwittingly include code that imposes licensing > conditions.It's pretty narsty, but including the text of the license in a data segment that gets automatically linked into the final program will satisfy the letter of the license text --- after all, it doesn't say it has to be *readable*. My reading of the MIT license is that it requires binary attribution, too --- it doesn't distinguish between source and binary and therefore applies to both. But I'd very much like to be wrong. - -- ┌─── ｄｇ＠ｃｏｗｌａｒｋ．ｃｏｍ ───── http://www.cowlark.com ───── │ │ "I have a mind like a steel trap. It's rusty and full of dead mice." │ --- Anonymous, on rasfc -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iD8DBQFOFOwLf9E0noFvlzgRAkEMAJ9jgb9uMnd94A24x3AaEJgwLMKRxACfX99P 8zACWzwm3xO8bS1F762GMVI=yqTO -----END PGP SIGNATURE-----
On Jul 6, 2011, at 12:47 PM, Tor Gunnar Houeland wrote:> On 07/06/2011 07:10 PM, Chris Lattner wrote: >> On Jul 6, 2011, at 1:55 AM, Tor Gunnar Houeland wrote: >> >>>> There is no need to include any notices in the binaries of an application built with clang, or some with some other application that links to the LLVM runtime libraries that are dual licensed. >>> Thanks for your response. Is this ability to distribute binaries without notices based on your personal assertion that the MIT license used does not require them? >> Yes. I am not a lawyer and am not giving legal advice. This is based on my understanding of the MIT license. > > OK, thanks. That's contradictory to the actual conditions written in the license, and as such the dual-licensing does not serve that purpose. Could http://llvm.org/docs/DeveloperPolicy.html#license be updated to reflect that? > > (Or do you strongly disagree that "all copies of the software" includes "binary"? When changing the license, you mentioned an example of Mozilla building with Clang. As can be seen in about:license, they consider MIT licenses to require reproduction of the license text. What's your reasoning for your understanding?)I, and many other reasonable people, consider the phrase: "The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software." ... to be talking about *copies of the software*. A binary is not a copy of the software, it is a lump of bits derived from it. I am not a lawyer, and do not offer this as legal advise. However, lawyers that I respect agree with this interpretation. You can choose to interpret it however you would like. -Chris
On 07/07/2011 02:25 AM, Chris Lattner wrote:> I, and many other reasonable people, consider the phrase: > > "The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software." > > ... to be talking about *copies of the software*. A binary is not a copy of the software, it is a lump of bits derived from it. > > I am not a lawyer, and do not offer this as legal advise. However, lawyers that I respect agree with this interpretation. You can choose to interpret it however you would like.Thanks, that looks quite short and clear, though it doesn't seem to be obvious (and I would not personally wish to rely on it). Would it be possible to add something like that to http://llvm.org/docs/DeveloperPolicy.html#license so it's clear how it was interpreted? Do you happen to know if all the copyright holders also interpret it that way? - Tor Gunnar