Jonas Thiem via llvm-dev
2017-Nov-25 10:08 UTC
[llvm-dev] Runtime library components licensing / MIT License / credit requirements
Hi, I am writing here because the LLVM Developer Policy on licensing as found here http://llvm.org/docs/DeveloperPolicy.html#license says that questions about the licensing should be sent to this list. The LLVM Developer Policy on licensing says the following about runtime components compiled into/linked to programs compiled with clang: "In addition to the UIUC license, the runtime library components of LLVM (compiler_rt, libc++, and libclc) are also licensed under the MIT License, which does not contain the binary redistribution clause. [...] We feel that this is important for runtime libraries, because they are implicitly linked into applications and therefore should not subject those applications to the binary redistribution clause." This sounds like the idea of picking the MIT license instead of clang's regular Illinois licensing is that there shall be no need to have a condition similar to the referenced "binary redistribution clause" which reads as follows: "Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimers in the documentation and/or other materials provided with the distribution." However, the MIT license says the following: "The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software." which doesn't define what a substantial portion of the software even is, and notably does NOT exclude redistributions in binary form at all. I am also pretty certain I've read interpretations on the web which understand the license implications of MIT similarly that binary redistributions must also adhere to the copyright & conditions notice requirement such that it doesn't actually provide any weaker requirements here than the University of Illinois/NCSA license. Therefore, how is the MIT even better here? It seems like it doesn't actually achieve the intended goal with any certainty. Wouldn't it be a much better idea to pick the zlib/libpng license or another license with a similarly clear text like this: "This notice may not be removed or altered from any source distribution." .. which very clearly excludes include binary redistribution from this requirement? Something again the documentation appears to suggest for the MIT license but which the MIT license actually doesn't seem to fulfill. Disclaimer: I am not a lawyer so all of what I am writing here might be utterly wrong Regards, Jonas Thiem
Jonas Thiem via llvm-dev
2017-Nov-28 17:30 UTC
[llvm-dev] Runtime library components licensing / MIT License / credit requirements
To explain a bit my motivation behind this e-mail due to the lack of reactions: What I am hoping to achieve with this request? I asked mainly because I would love to hear either more in-depth explanations why the current MIT licensing of the clang runtime parts has no such legal implications, or acknowledgment of the issue with some sort of basic considerations of a potential refinement of the clang licensing to improve this. Why do I care in the first place? I am considering using a clang-compiled executable as a bytecode runtime for something which other developers use (and which ships with their self-written programs), and since my own project is zlib/libpng licensed I currently have zero such license notice requirements. If clang introduces them, it would complicate things a bit: It wouldn't be an absolute showstopper of course, but I would feel compelled to inform those people through various means and write FAQs about it and all, and it might lead to 1. making the licensing compliance for the clang-compiled bytecode runtime I am writing more complicated (compared to "nothing required") 2. a bit of a potential grey area for people who miss out on this despite FAQ and all since people are usually aware of lib deps & compliance with that but not hidden core runtime deps put in by the compiler. I hope this helps making my questions more interesting and relatable. Regards, Jonas Thiem On 11/25/2017 11:08 AM, Jonas Thiem via llvm-dev wrote:> Hi, > > I am writing here because the LLVM Developer Policy on licensing as > found here http://llvm.org/docs/DeveloperPolicy.html#license says that > questions about the licensing should be sent to this list. > > The LLVM Developer Policy on licensing says the following about runtime > components compiled into/linked to programs compiled with clang: > > "In addition to the UIUC license, the runtime library components of LLVM > (compiler_rt, libc++, and libclc) are also licensed under the MIT > License, which does not contain the binary redistribution clause. [...] > We feel that this is important for runtime libraries, because they are > implicitly linked into applications and therefore should not subject > those applications to the binary redistribution clause." > > This sounds like the idea of picking the MIT license instead of clang's > regular Illinois licensing is that there shall be no need to have a > condition similar to the referenced "binary redistribution clause" which > reads as follows: "Redistributions in binary form must reproduce the > above copyright notice, this list of conditions and the following > disclaimers in the documentation and/or other materials provided with > the distribution." > > However, the MIT license says the following: "The above copyright notice > and this permission notice shall be included in all copies or > substantial portions of the Software." which doesn't define what a > substantial portion of the software even is, and notably does NOT > exclude redistributions in binary form at all. I am also pretty certain > I've read interpretations on the web which understand the license > implications of MIT similarly that binary redistributions must also > adhere to the copyright & conditions notice requirement such that it > doesn't actually provide any weaker requirements here than the > University of Illinois/NCSA license. > > Therefore, how is the MIT even better here? It seems like it doesn't > actually achieve the intended goal with any certainty. Wouldn't it be a > much better idea to pick the zlib/libpng license or another license with > a similarly clear text like this: "This notice may not be removed or > altered from any source distribution." .. which very clearly excludes > include binary redistribution from this requirement? Something again the > documentation appears to suggest for the MIT license but which the MIT > license actually doesn't seem to fulfill. > > Disclaimer: I am not a lawyer so all of what I am writing here might be > utterly wrong > > Regards, > Jonas Thiem > _______________________________________________ > LLVM Developers mailing list > llvm-dev at lists.llvm.org > http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev >