Le 22/12/2014 20:36, sebb a écrit : > On 22 December 2014 at 19:17, Luc Maisonobe <l...@spaceroots.org> wrote: >> Le 22/12/2014 18:28, sebb a écrit : >>> On 22 December 2014 at 17:09, Luc Maisonobe <l...@spaceroots.org> wrote: >>>> Le 22/12/2014 17:33, sebb a écrit : >>>>> On 22 December 2014 at 16:19, Luc Maisonobe <l...@spaceroots.org> wrote: >>>>>> Le 22/12/2014 16:44, sebb a écrit : >>>>>>> -1 >>>>>>> >>>>>>> There are still problems in the NOTICE and/or LICENSE files. >>>>>>> >>>>>>> None of the additional attributions in the NOTICE file have >>>>>>> corresponding entries in the LICENSE file. >>>>>> >>>>>> There are 3 attributions in the Notice file. >>>>>> >>>>>> - one for the erf function. There were no license. We had to contact, >>>>>> the author directly and he allowed us to reuse the code under the >>>>>> Apache license. >>>>> >>>>> Where is that contact documented? >>>> >>>> It is explained in the MATH-948 JIRA issue. We get a Sofware grant from >>>> Mike Giles (Oxford University) that was sent to secretary on March 23th >>>> 2013. >>> >>> The JIRA does not reference the actual communication, so it's not >>> possible to trace this. >>> >>>>> It would be helpful to have it in the LICENSE file. >>> >>> Still true. >>> >>>>> >>>>>> - the second one was for Ben McCann work. This was distributed under >>>>>> the Apache License, which is in the license file >>>> >>>> For this one, we also received a grant on 2009-03-16. As the original >>>> files posted in MATH-246 did include an author tag. I asked Ben if I >>>> could remove it and put it in the NOTICE file. He was fine with this. >>> >>> Again, only part of the conversation is documented. >>> It is not clear if the author would have been happy with other forms >>> of attribution. >>> >>>>>> - the thirs one is for Orekit, which is also distributed under >>>>>> Apache 2 license and therefore is present in the License file. >>>>>> >>>>>> I do not think we need to copy the Apache license three times in one >>>>>> file. >>>>> >>>>> Of course not, but the LICENSE file must still state that AL 2.0 is >>>>> used for these items. >>>> >>>> So, should I put this in NOTICE or LICENSE. I am afraid I am lost there. >>> >>> The LICENSE file needs to mention all 3rd party licenses. >>> For AL 2.0, it just needs to mention which 3rd party items are >>> included under it. >>> For other licenses, the licenses themselves need to be included, >>> either in-line or via a link to a separate file that is included in >>> the distribution. >>> >>>> >>>>> >>>>>>> >>>>>>> That cannot be correct, regardless of whether the attributions are >>>>>>> required. >>>>>>> >>>>>>> AFAICT the attributions are not required *unless* they were removed >>>>>>> from source code. >>>>>> >>>>>> I strongly disagree here. It is completely forbidden to remove copyright >>>>>> from files unless allowed by the user. >>>>> >>>>> I agree (I wrote *unless* ... above), but is that the case here? >>>> >>>> At least for the Orekit case, the library does have a NOTICE file, so >>>> according to the Apache license we must include the relevant parts, >>>> and the parts that have been included are the ones for which the >>>> attribution is required (i.e. the other parts in Orekit NOTICE that were >>>> not included in Apache because we imported only a small part have been >>>> removed so Apache Commons Math NOTICE contains only the minimum required >>>> by the license, has stated in the commit message for 4edc00b). >>> >>> OK, but this does not explain why the Orekit references in N & L are >>> completely different. >>> >>>>> >>>>>> When a license states that >>>>>> attribution must be put "in the documentation and/or other materials >>>>>> provided with the distribution", it doesn't mean you are allowed to >>>>>> removed them for elsewhere. >>>>> >>>>> Attribution can also be satisified by the LICENSE file. >>>>> For example, this is true of BSD and MIT >>>>> >>>>> http://www.apache.org/dev/licensing-howto.html#mod-notice >>>>> >>>>>>> >>>>>>> The RC e-mail needs a link to the KEYS file (not a release blocker) >>>> >>>> Do you consider the above NOTICE problems to be release blockers? >>> >>> Yes. >>> >>> And there are LICENSE issues as well. >> >> So here is what I propose to do. >> >> Add this after the Apache license text in the license file: >> >> ---------------- >> The previous Apache license applies to the code specifically >> developed for Apache Commons Math by the committers, and to some >> code developed by third parties that was licensed to the Apache >> Software Foundation through Software Grants, or was already distributed >> under the terms of the Apache V2 license. This includes: > > I would put something like:
OK > > Apache Commons Math includes the following code provided to the ASF under the > Apache License 2.0: > >> - The inverse error function implementation in the Erf class is based >> on CUDA >> code developed by Mike Giles, Oxford-Man Institute of Quantitative >> Finance, >> and published in GPU Computing Gems, volume 2, 2010 (grant received on >> March 23th 2013) >> - The LinearConstraint, LinearObjectiveFunction, LinearOptimizer, >> RelationShip, SimplexSolver and SimplexTableau classes in package >> org.apache.commons.math3.optimization.linear include software >> developed by >> Benjamin McCann (http://www.benmccann.com) and distributed with >> the following copyright: Copyright 2009 Google Inc. (grant received on >> March 16th 2009) >> - The class >> "org.apache.commons.math3.exception.util.LocalizedFormatsTest" which >> is an adapted version of "OrekitMessagesTest" test class for the >> Orekit library >> - The "org.apache.commons.math3.analysis.interpolation.HermiteInterpolator" >> has been imported from the Orekit space flight dynamics library. >> --------------- >> >> Remove the attributions for Mike Giles and Ben McCann from NOTICE, as >> they would now be in the above text in LICENSE. >> >> Let only the Orekit very simple lines in NOTICE: >> >> ------------ >> This product includes software developed for Orekit by >> CS Systèmes d'Information (http://www.c-s.fr/) >> Copyright 2010-2012 CS Systèmes d'Information >> ------------ > > Why is the Orekit attribution needed? Because Orekit is licensed under Apache V2 license. Section 4d reads: If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file ... best regards, Luc > >> Would this solve the issues for you? >> >> best regards, >> Luc >> >>> >>>> best regards, >>>> Luc >>>> >>>>>>> It might help if the Git link pointed to the commit rather than the >>>>>>> tarball, for example >>>>>>> >>>>>>> https://git-wip-us.apache.org/repos/asf?p=commons-math.git;a=commit;h=fe08f0d3bd0309db15d27335757336a8ac6f32d0 >>>>>>> >>>>>>> This gives access to the tarball and the tree directly. >>>>>> >>>>>> I don't think access to the tree would help here since the KEYS file >>>>>> is shared with other components and is in the subversion part, >>>>>> not the git part. So the link must be something specific. I'll >>>>>> add it in the template. >>>>>> >>>>>> best regards, >>>>>> Luc >>>>>> >>>>>>> >>>>>>> >>>>>>> On 19 December 2014 at 15:54, Luc Maisonobe <l...@spaceroots.org> wrote: >>>>>>>> This is a [VOTE] for releasing Apache Commons Math 3.4 from release >>>>>>>> candidate 1. >>>>>>>> >>>>>>>> Tag name: >>>>>>>> MATH_3_4_RC2 (signature can be checked from git using 'git tag -v') >>>>>>>> >>>>>>>> Tag URL: >>>>>>>> >>>>>>>> <https://git-wip-us.apache.org/repos/asf?p=commons-math.git;a=snapshot;h=fe08f0d3bd0309db15d27335757336a8ac6f32d0;sf=tgz> >>>>>>>> >>>>>>>> Commit ID the tag points at: >>>>>>>> fe08f0d3bd0309db15d27335757336a8ac6f32d0 >>>>>>>> >>>>>>>> Site: >>>>>>>> <http://people.apache.org/~luc/commons-math-3.4-RC2-site> >>>>>>>> >>>>>>>> Distribution files: >>>>>>>> <https://dist.apache.org/repos/dist/dev/commons/math/> >>>>>>>> >>>>>>>> Distribution files hashes (SHA1): >>>>>>>> >>>>>>>> a0ef31377fa4a38be71ce6ddc28044e9c8d34ee8 commons-math3-3.4-bin.tar.gz >>>>>>>> a4ad5264c2016e893acb4e004438c370a927802c commons-math3-3.4-bin.zip >>>>>>>> 2c300e94f69761b34f1b4247133776440a37c801 commons-math3-3.4-src.tar.gz >>>>>>>> c49b4f25e227ab66dfe90627d794c1e1accccdb4 commons-math3-3.4-src.zip >>>>>>>> >>>>>>>> >>>>>>>> Maven artifacts: >>>>>>>> >>>>>>>> <https://repository.apache.org/content/repositories/orgapachecommons-1067/org/apache/commons/commons-math3/3.4/> >>>>>>>> >>>>>>>> [ ] +1 Release it. >>>>>>>> [ ] +0 Go ahead; I don't care. >>>>>>>> [ ] -0 There are a few minor glitches: ... >>>>>>>> [ ] -1 No, do not release it because ... >>>>>>>> >>>>>>>> This vote will close in 72 hours, at 2014-12-22T16:00:00Z (this is UTC >>>>>>>> time). >>>>>>>> >>>>>>>> --------------------------------------------------------------------- >>>>>>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>>>>>>> For additional commands, e-mail: dev-h...@commons.apache.org >>>>>>>> >>>>>>> >>>>>>> --------------------------------------------------------------------- >>>>>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>>>>>> For additional commands, e-mail: dev-h...@commons.apache.org >>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> --------------------------------------------------------------------- >>>>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>>>>> For additional commands, e-mail: dev-h...@commons.apache.org >>>>>> >>>>> >>>>> --------------------------------------------------------------------- >>>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>>>> For additional commands, e-mail: dev-h...@commons.apache.org >>>>> >>>>> >>>>> >>>> >>>> >>>> --------------------------------------------------------------------- >>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>>> For additional commands, e-mail: dev-h...@commons.apache.org >>>> >>> >>> --------------------------------------------------------------------- >>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>> For additional commands, e-mail: dev-h...@commons.apache.org >>> >>> >>> >> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >> For additional commands, e-mail: dev-h...@commons.apache.org >> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > > > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
